View 3 - Waterfall View for All Capabilities v1.2

Autoregressive/Wastewater-based Infectious Disease Forecasting

Budget Allocation

Agent-Based Infectious Disease Modeling

Consumer Archetyping for DTC Marketing

Treatment Initiation Drivers Analysis

Patient Diagnosis/Prognosis Drivers Analysis

High Risk Patient Identification

Patient Journey Analysis

Treatment Selection Drivers Analysis

Driver Temporal Evolution Analysis

Treatment Switching Drivers Analysis

Patient Need Subnational Region Archetyping

Channel Affinity Segmentation

Healthcare Readiness Subnational Region Archetyping

Treatment Compliance Drivers Analysis

Driver-to-KPI Impact Simulation Analysis

Prescriptive Analytics

Predictive Analytics

Diagnostic Analytics

Descriptive Analytics

HCP Value Segmentation Analysis

HCP Headroom Drivers Analysis

Launch Excellence

Post Launch

Pre-Launch

Promotional Impact & mROI

HCP Brand Loyalty/Churn Analysis

Next Best Action Engine

HCP Adoption Segmentation Analysis

Omnichannel Orchestration

Tactical Planning

Performance Measurement

Strategic Planning

Market Assessment & Estimation

AIDA BioPharma Analytics Capabilities

Autoregressive/Wastewater-based Infectious Disease Forecasting

Budget Allocation

Agent-Based Infectious Disease Modeling

Consumer Archetyping for DTC Marketing

Treatment Initiation Drivers Analysis

Patient Diagnosis/Prognosis Drivers Analysis

High Risk Patient Identification

Patient Journey Analysis

Treatment Selection Drivers Analysis

Driver Temporal Evolution Analysis

Treatment Switching Drivers Analysis

Patient Need Subnational Region Archetyping

Channel Affinity Segmentation

Healthcare Readiness Subnational Region Archetyping

Treatment Compliance Drivers Analysis

Driver-to-KPI Impact Simulation Analysis

HCP Value Segmentation Analysis

HCP Headroom Drivers Analysis

Launch Excellence

Post Launch

Pre-Launch

Promotional Impact & mROI

HCP Brand Loyalty/Churn Analysis

Next Best Action Engine

HCP Adoption Segmentation Analysis

Omnichannel Orchestration

Tactical Planning

Performance Measurement

Strategic Planning

Market Assessment & Estimation

AIDA BioPharma Analytics Capabilities

Prescriptive Analytics

Predictive Analytics

Diagnostic Analytics

Descriptive Analytics

Autoregressive/Wastewater-based Infectious Disease Forecasting

Budget Allocation

Agent-Based Infectious Disease Modeling

Consumer Archetyping for DTC Marketing

Patient Journey Analysis

Treatment Selection Drivers Analysis

Treatment Switching Drivers Analysis

Treatment Initiation Drivers Analysis

HCP Value Segmentation Analysis

Patient Diagnosis/Prognosis Drivers Analysis

High Risk Patient Identification

Treatment Compliance Drivers Analysis

Driver Temporal Evolution Analysis

Patient Need Subnational Region Archetyping

HCP Headroom Drivers Analysis

Launch Excellence

Post Launch

Pre-Launch

Healthcare Readiness Subnational Region Archetyping

Promotional Impact & mROI

Driver-to-KPI Impact Simulation Analysis

Channel Affinity Segmentation

HCP Brand Loyalty/Churn Analysis

Next Best Action Engine

HCP Adoption Segmentation Analysis

.............................

.............................

.............................

...........................................................

.............................

.............................

............................................

............................................

............................................

............................................

............................................

..........................................................................

...............

.............................

.............................

..............

.............................

..............................

HCP

GEOGRAPHY

PATIENT

Amplifying Message

Promo Deployment & Orchestration

GTM Budget Optimization & Impact (ROI)

Untapped HCP Opportunity

Brand Choice Evaluation

GTM Planning Resource Allocation

Segmentation Targeting

Epidemiology Assessment

Opportunity Assessment

Omnichannel Orchestration

Tactical Planning

Performance Measurement

Strategic Planning

Market Assessment & Estimation

AIDA BioPharma Analytics Capabilities

.............................

.............................

..............

..............

........................................................................

• How do the unmet needs vary by geography?

Key Business Questions

• Correlation Matrix against KPI for Feature Selection
• Cluster Analysis (K-means, Hierachical)

Advance Analytics Pipeline

As the local BU lead:

• Understand which subnational regions showcase a higher Patient Need for a specific Disease Area, to appropriately allocate budget and activities to maximize the resulting impact.

Developed on

Data Assets Epidemiological, Demographic, Disease Prevalence, MDM & C360/CEM Sales

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Patient Need Subnational Region Archetyping

Capability OverviewThis Patient Need Subnational Region Archetyping Analysis, allows us to capture those location-based metrics that are most associated with increased Patient Needs in a specific Disease Area, such as Demographics, Comorbidities, Infectious Disease Dynamics, Socio-economic factors, Vaccinations, etc., and the naturally forming subnational region archetypes. This analysis can be used for the following:

• Identify areas of increased Patient Need for the specific TA, to focus MR and CFC operations where they could have the greatest potential impact.
• Tailor resource allocation and DTC regional marketing campaigns appropriately to those areas where there will be greater impact of activity.

• How do the unmet needs vary by geography?

Key Business Questions

• Aggregate data to national and patient archetype level
• Calculate longitudinal or regional profile
• Calculate total cases according to the infection rate
• Calculate cases per wastewater concentration unit
• Utilize wastewater concentration shape to calculate the cases profile
• Use cases/concentration unit to estimate true cases

Advance Analytics Pipeline

As the local BU Lead:

• Understand what are the future projections of the infecious disease in question to organize future activities appropriately

As the local FF orchestration lead:

• Understand which areas are projected to have a more sever viral load and what is the seasonal peaks and troughs of the projections to coordinate effort allocation and timing more appropriately to maximize the reached patients.

Developed on

Data Assets Longitudinal epidemiological Data, Wastewater concentration data

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Autoregressive/Wastewater-based Infectious Disease Forecasting

Capability OverviewAutoregressive/Wastewater-based Infectious Disease Forecasting captures the seasonal evolution of infectious disease by analyzing its historical datapoints (when undereporting is non time-variant and stable) or by leveraging viral weight in wastewater concentration datapoints, and project in the future. This can be used to:

• Identify the seasonal peaks and troughs of an infectious disease to better coordinate FF activities' timing and volume to maximize patient reach out and treatment/awareness.
• Calculate the effect that well documented events have on the underlying level, trend and seasonality of the disease to estimate the expected impact of an infectious disease overall.
• Estimate the attribution of national effect to subnational regions where disease metrics (reported cases, hospitalizations, deaths from disease) or wastewater metrics are available

Agent-Based Modelling Disease Forecasting

Audience/User: Marketer, Leadership, Sales

Purpose: Used to estimate new variant of breakout behavior and amplitude conditional on disease characteristics and mobility information. Can provide a simulated expected outcome of the severity of new infectious diseases and to understand relative impact of diseases to coordinate FF activities, replenishment and even government outreach. Name of AIDA Analytic Capability: Agent-Based Modelling Disease Forecasting Platform Developed on: VAW

Treatment Adherence & Persistence Analysis

Audience/User: Marketer, Leadership, Sales

Purpose: Used to understand what drives treatment Adherence and/or Peristency. Could be used for HCP targetting, content tailoring or to understand emerging trends and tailoring actions to increase the LTV of patients. Predominantly useful on chronic or long-term treatment. Name of AIDA Analytic Capability: Treatment Adherence & Persistence Analysis Platform Developed on: VAW

Data Assets

• What are unmet opportunities in the market?
• How important/large are these needs/opportunities
• What are the drivers of Brand Choice?

Key Business Questions

• Modelling Feature Clustering
• Multicollinearity Elimination
• Patient Clustering & Cluster Definition
• Over and/or Under Sampling for balancing
• Multi-stage Modelling of variable categories
• LR Regression/Decision Trees

Advance Analytics Pipeline

As a Local Medical Affairs Lead:

• Understand which clinical markers indicate a person is high-risk for having a disease/their relevant sequence as well as the HCPs along the patient's journey to be able to generate appropriate informational content and coordinate MR activities to increase awareness and diagnosis rates.

As a FF coordinator:

• Find appropriate HCPs to target that have a large number of high-risk patients currently undiagnosed at an earlier time in their patient journey to raise awareness and reduce the time-to-diagnosis.

Developed on

EHR data Claims Data

Brands

Audience/User Marketer, Leadership, Sales

User Stories

High Risk Patient Identification

Capability Overview The High Risk Patient Identification Analysis allows us to create models that estimate the probability an individual, given their clinical records, might have an undiagnosed disease. This model can be used to:

• Find Patients earlier in their clinical journey and track back to the HCPs to coordinate awareness reach-out on the disease by the FF, lead to earlier diagnosis and thus improved outcomes, better diagnosis rates and greater LTV of the patient.
• Make a sizing estimation of the market to understand estimated disease prevalance and not reported ones.
• Identify HCPs that have significant number of undiagnosed high risk patients to coordinate FF reach out.
• Identify key clinical markers and their sequence that would indicate the existence of the disease.

Data Assets

• What are unmet opportunities in the market?
• How important/large are these needs/opportunities?
• What are the drivers of Brand Choice?

Key Business Questions

• Modelling Feature Clustering
• Multicollinearity Elimination
• Patient Clustering and Cluster Definition
• Over and/or Under Sampling for balancing
• Multi-stage Modelling of variable categories
• LR Regression/Decision Trees

As the local Marketing Lead:

• Understand which are major Patient Archetypes that are being non-compliant to tailor appropriate material for DTC marketing campaigns focuses on those archetypes or to tailor content delivered to HCPs that service those archetypes to increase Patient LTV.

As the Medical Affairs local lead:

• Understand which major archetypes with respect to Treatment Compliance to tailor MR content and efforts orchestration to increase Treatment Compliance and positive Treatment Outcomes.

Advance Analytics Pipeline

Developed on

EHR data Claims Data SPP Data Experian Data

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Treatment Compliance Drivers Analysis

Capability OverviewTreatment Compliance Drivers Analysis helps understand which factors affect whether a patient wil remain adherent and persistent to a particular treatment pathway or whether they will skip doses or discontinue treatment and not complete treatment in a compliant way. Those factors could be clinical, SDOH, HCP, Socio-Economic in aggregate, Insurance related or any that have been identified by the BU. This analysis can be used to:

• Identify the most prevalent Patient Archetypes affecting compliance, whether that is adherence or persistence.
• Map those archetypes back to HCPs or Subnational Regions to understand distribution of those archetypes for each HCP or location.
• Tailor content appropriately to the identified archetypes for improving compliance either via DTC marketing campaigns (Location anchored or not) or to HCPs that overindex non-compliant patient archetypes.

Data Assets

• What are unmet opportunities in the market?
• How important/large are these needs/opportunities?
• What are the drivers of Brand Choice?

Key Business Questions

• Modelling Feature Clustering
• Multicollinearity Elimination

• Over and/or Under Sampling for balancing
• Multi-stage Modelling of variable categories
• LR Regression/Decision Trees
• Survival Analysis/Time-to-Event Analysis

Advance Analytics Pipeline

N/A

As a Local Medical Affairs Lead:

• Understand which factors affect time-to-diagnosis to tailor content and MR reach out efforts to reduce diagnosis lead time and improve health outcomes.

Developed on

EHR Data, Claims Data, Experian Data, Census Data

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Patient Diagnosis/Prognosis Drivers Analysis

Capability Overview The Patient Diagnosis/Prognosis Drivers Analysis allows us to create models that capture what affects time-to-diagnosis and progression to a more severe stage of the disease. These models can be used for the following:

• Understand what clinical, SDOH, exogenous or HCP characteristics seem to accelerate or delay time-to-diagnosis.
• Understand what clinical, observational or treatment and treatment compliance related characteristics might lead to a more severe stage of the disease area.
• Find which HCPs seem to have Patients at risk of progression to a more severe disease stage to coordinate awareness efforts.

Data Assets

• Which HCPs could address the unmet needs?
• What are the drivers of HCP prescribing behavior?

Key Business Questions

• TRx/NRx/Patient Number deciling and classification
• Unsupervised Learning Models (K-Means/ Hierarchical Clustering)

Advance Analytics Pipeline

As the local FF activities Orchestration Lead:

• Understand current Value HCPs driving Sales/ Market Share to understand which HCPs FF needs to prioritize in maintaining a relationship with

As the local FF activities Orchestration Lead:

• Understand potential value behaviors/ segments such as HCPs of higher TA Patient Share, HCPs operating in the treatment class who are splitting prescriptions or HCPs that operate in a Treatment Class of high relevance to the Treatment Class Pfizer operates in and are referencing high volume of patients to inform the approach and targetting of FF.

Developed on

Claims Data MDM (HCP/HCO) Onekey Data C360 / CEM Data

Brands

Audience/User Marketer, Sales

User Stories

HCP Value Segmentation Analysis

Capability OverviewHCP Value Segmentation Analysis allows us to understand/capture important Segments of HCPs with respect to their prescribing behavior in terms of rendering volume, pre-rending referencing, brand prescription share within the treatment class/prior treatment class than where Pfizer operates in. This analysis can be used to:

• Identify which HCPs are splitting within the treatment class where we operate.
• Identify which HCPs belong in higher deciles with respect to the volume share of patients within the TA.
• Identify which HCPs operate in Treatment Classes preceding the one where we operate and have a high reference out volume for subsequent Treatment Classes.
• Tailor prioritity and method/content of approach to those HCPs based on the above.

EHR dataClaims Data

• What are unmet opportunities in the market?
• How important/large are these needs/opportunities?
• What are the drivers of Brand Choice?

Key Business Questions

• N/A

Advance Analytics Pipeline

As a Local BU Lead:

• Understand which Patient Journey end-points exist in the disease area in question, which of those are servicable by our products/which are servicable by competitor in-line or prelaunch to better coordinate activities to defend or grow our product's share

As a FF coordinator:

• Understand which Referral Transition nodes and the associated HCPs to target to ensure that more patients can get treatment on time with our product

Developed on

Data Assets

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Patient Journey Analysis

Capability Overview The Patient Journey Analysis shows a complete view of the Patient's Journey in a dendrical structure, with end nodes representing various endpoints of their journey and intermediary nodes their transitions to them. End-points and transitions allow us to understand untapped potential/threats in juxtaposition with our product's and the competitor's coverage of end-points and where along the journey our efforts could have the most impact to commercial and medical outcomes/KPIs. Examples of endpoints and transitions: Patients with delayed Treatment, Patients without Treatment, Patient Reference prior to Diagnosis, Patient Reference prior to Treatment, Non-Adherant/Persistent Patients, Patients where Primary Treatment sufficed, Patients that moved from Primary Treatment to Secondary Treatment in X months, Patients that moved from X Treatment Class to Y Treatment Class.

Data Assets

• What are unmet opportunities in the market?
• How important/large are these needs/opportunities?
• What are the drivers of Brand Choice?

Key Business Questions

• Modelling Feature Clustering
• Multicollinearity Elimination
• Patient Clustering and Cluster Definition
• Over and/or Under Sampling for balancing
• Multi-stage Modelling of variable categories
• LR Regression/Decision Trees

As the local Marketing Lead:

• Understand which are major Patient Archetypes that are being non-compliant to tailor appropriate material for DTC marketing campaigns focuses on those archetypes or to tailor content delivered to HCPs that service those archetypes to increase Patient LTV.

As the Medical Affairs local lead:

• Understand which major archetypes with respect to Treatment Compliance to tailor MR content and efforts orchestration to increase Treatment Compliance and positive Treatment Outcomes.

Advance Analytics Pipeline

Developed on

EHR data Claims Data SPP Data Experian Data

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Treatment Compliance Drivers Analysis

Capability OverviewTreatment Compliance Drivers Analysis helps understand which factors affect whether a patient wil remain adherent and persistent to a particular treatment pathway or whether they will skip doses or discontinue treatment and not complete treatment in a compliant way. Those factors could be clinical, SDOH, HCP, Socio-Economic in aggregate, Insurance related or any that have been identified by the BU. This analysis can be used to:

• Identify the most prevalent Patient Archetypes affecting compliance, whether that is adherence or persistence.
• Map those archetypes back to HCPs or Subnational Regions to understand distribution of those archetypes for each HCP or location.
• Tailor content appropriately to the identified archetypes for improving compliance either via DTC marketing campaigns (Location anchored or not) or to HCPs that overindex non-compliant patient archetypes.

Data Assets

• Which HCPs could address the unmet needs?
• What are the drivers of HCP prescribing behavior?

Key Business Questions

Developed on

Claims, MDM/ Onekey, C360/CEM, HCP Channel Affinities / Adoption Segments

• Modelling Feature Clustering
• Multi-collinearity Elimination
• Over and/or Under Sampling for balancing (SMOTE)
• Logistic Regression based modelling and quantification of effect of independent variable to dependent variable

Advance Analytics Pipeline

As the local FF activities Orchestration Lead:

• Understand which HCP Segments and specific HCPs within them have a high probability on transitioning to a higher or lower value Adoption Stage and orchestrate content and FF reach out to address the emerging Opportunities or Threats.

Brands

Audience/User Marketer, Sales

User Stories

HCP Brand Loyalty/Churn Analysis

Capability OverviewThis HCP Brand Loyalty/Churn Drivers Analysis, allows us to understand/capture the main differentiators between HCP of associated Adoption Stages (Aware to Trial, Trial to Usage while active in Class, Tria to Usage while active in prior Class of Treatment, etc.). Those differentiators may be FF engagement, exogenous events, HCP Specialty or Patient Archetypes evolution, delivered content among a few. This analysis can be used for the following:

• Identify which HCPs are likely to switch to or away from a particular Adoption Stage of interest and appropriate orchestrate FF activities.
• Identify tactically which factors can be affected to facilitate or prevent Churn to a particular Adoption Stage of interest to orchistrate FF activity and improve KPIs related to Adoption.

• How do the unmet needs vary by geography?

Key Business Questions

• Aggregate data to national and patient archetype level
• Calculate longitudinal or regional profile
• Calculate total cases according to the infection rate
• Calculate cases per wastewater concentration unit
• Utilize wastewater concentration shape to calculate the cases profile
• Use cases/concentration unit to estimate true cases

Advance Analytics Pipeline

As the local BU Lead:

• Understand what are the future projections of the infecious disease in question to organize future activities appropriately

As the local FF orchestration lead:

• Understand which areas are projected to have a more sever viral load and what is the seasonal peaks and troughs of the projections to coordinate effort allocation and timing more appropriately to maximize the reached patients.

Developed on

Data Assets Longitudinal epidemiological Data, Wastewater concentration data

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Autoregressive/Wastewater-based Infectious Disease Forecasting

Capability OverviewAutoregressive/Wastewater-based Infectious Disease Forecasting captures the seasonal evolution of infectious disease by analyzing its historical datapoints (when undereporting is non time-variant and stable) or by leveraging viral weight in wastewater concentration datapoints, and project in the future. This can be used to:

• Identify the seasonal peaks and troughs of an infectious disease to better coordinate FF activities' timing and volume to maximize patient reach out and treatment/awareness.
• Calculate the effect that well documented events have on the underlying level, trend and seasonality of the disease to estimate the expected impact of an infectious disease overall.
• Estimate the attribution of national effect to subnational regions where disease metrics (reported cases, hospitalizations, deaths from disease) or wastewater metrics are available

• N/A

Across all Markets

• How do the unmet needs vary by geogr aphy?

Key Business Questions

Advance Analytics Pipeline

As a Regional President

• Optimize allocation of resources across brands / countries to improve overall sales / profitability

As a Country President

• Optimize allocation of resources across brands / therapy areas, to improve overall sales / profitability

As a TA / BU Lead

• Optimize allocation of resources across brands in my TA to improve overall sales / profitability

Developed on

Data Assets Finance, Forecasts, margins, C360, Other country planning inputs

Brands

Audience/User Regional President / Country President / BU or TA Lead

User Stories

Budget Allocation (Portfolio / Cross-brand / Cross-market)

Capability Overview Budget Allocation exercises involve using financial data such as revenues, margins, growth / evolution, market shares & benchmark-based response curves / channel-level impact estimates to directionally estimate the optimal allocation of spend across brands & markets. They are usually done at a regional-level across markets or combinations of brands / markets in order to prioritze where to reallocate budgets from & to.

Across Brands in Europe, LATAM, Gulf regions along with Netherlands & Saudi Arabia

Data Assets

• Which HCPs are interested in F2F promotion and have engaged well with all Pfizer channels?

• How can I define a target list to optimize HQ-level & FF-driven promotion

Key Business Questions

Developed on

C360 MDM

• Cluster Analysis (K-means, Hierachical, business-rules driven)

Advance Analytics Pipeline

As a Sales Lead / Marketer

• Understand the channel affinities of HCPs to tailor promotional activity to their channels of preference, thereby optimizing customer experience & driving overall sales

Brands

Audience/User Marketer, Sales

User Stories

Channel Affinity Segmentation

Capability OverviewChannel Affinity Segmentation looks at historical activity done with an HCP for a specific brand or overall, looks at Depth (Interaction & Engagement) & Consistency over a 6-,12- or 24-month period & generates a channel score for each HCP by summing normalized depth & consistency. Different clustering models are applied to create segments of customers who have similar affinity to one or more channels & can then be manually combined / tweaked further & renamed.

Treatment Adherence & Persistence Analysis

Audience/User: Marketer, Leadership, Sales

Purpose: Used to understand what drives treatment Adherence and/or Peristency. Could be used for HCP targetting, content tailoring or to understand emerging trends and tailoring actions to increase the LTV of patients. Predominantly useful on chronic or long-term treatment. Name of AIDA Analytic Capability: Treatment Adherence & Persistence Analysis Platform Developed on: VAW

EHR dataClaims Data

• What are unmet opportunities in the market?
• How important/large are these needs/opportunities?
• What are the drivers of Brand Choice?

Key Business Questions

• N/A

Advance Analytics Pipeline

As a Local BU Lead:

• Understand which Patient Journey end-points exist in the disease area in question, which of those are servicable by our products/which are servicable by competitor in-line or prelaunch to better coordinate activities to defend or grow our product's share

As a FF coordinator:

• Understand which Referral Transition nodes and the associated HCPs to target to ensure that more patients can get treatment on time with our product

Developed on

Data Assets

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Patient Journey Analysis

Capability Overview The Patient Journey Analysis shows a complete view of the Patient's Journey in a dendrical structure, with end nodes representing various endpoints of their journey and intermediary nodes their transitions to them. End-points and transitions allow us to understand untapped potential/threats in juxtaposition with our product's and the competitor's coverage of end-points and where along the journey our efforts could have the most impact to commercial and medical outcomes/KPIs. Examples of endpoints and transitions: Patients with delayed Treatment, Patients without Treatment, Patient Reference prior to Diagnosis, Patient Reference prior to Treatment, Non-Adherant/Persistent Patients, Patients where Primary Treatment sufficed, Patients that moved from Primary Treatment to Secondary Treatment in X months, Patients that moved from X Treatment Class to Y Treatment Class.

Census DataMobility Data Demographic Data

• How do the unmet needs vary by geography?

Key Business Questions

• Agent-Based Simulation Models

Advance Analytics Pipeline

As the local Medical Affairs Lead

• Understand what disease impact a new variant of a disease or a brand new disease will have on patients and the countries healthcare system to organize MR activities and disease awareness to address the impact more effectively.

Developed on

Data Assets

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Agent-Based Infectious Disease Modelling

Capability OverviewThe Agent-based Infectious Disease modelling analysis, allows us to simulate the effects of an infectious disease when that disease is a significantly different derivation/evolution of a prior disease (a variant that has differnt R0 or skips prior immunity) or when historical information is non-available due to a completely new disease. The analysis can be used for the following:

• Identify areas that due their characteristics (population density, demographics, population mobility to and from) will showcase a significantly different behavior than other areas either in terms of lagging effect of effect amplitude to orchestrate and allocate efforts appropriately.
• Estimate the effect an emerging variant or disease will have on the underlying population to orchestrate and allocate efforts and resources appropriately.

Census DataMobility Data Demographic Data

• How do the unmet needs vary by geography?

Key Business Questions

• Agent-Based Simulation Models

Advance Analytics Pipeline

As the local Medical Affairs Lead

• Understand what disease impact a new variant of a disease or a brand new disease will have on patients and the countries healthcare system to organize MR activities and disease awareness to address the impact more effectively.

Developed on

Data Assets

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Agent-Based Infectious Disease Modelling

Capability OverviewThe Agent-based Infectious Disease modelling analysis, allows us to simulate the effects of an infectious disease when that disease is a significantly different derivation/evolution of a prior disease (a variant that has differnt R0 or skips prior immunity) or when historical information is non-available due to a completely new disease. The analysis can be used for the following:

• Identify areas that due their characteristics (population density, demographics, population mobility to and from) will showcase a significantly different behavior than other areas either in terms of lagging effect of effect amplitude to orchestrate and allocate efforts appropriately.
• Estimate the effect an emerging variant or disease will have on the underlying population to orchestrate and allocate efforts and resources appropriately.

Major Brands acrossUS & EU region along with Japan

C360 CDM & NBA Consumption / Application Layer

• Customer Value Index (xGBoost with normalization)
• Verso OOTB CNN & GA models for optimal sequence generation / recommendation

Data Assets

• What recommendations/insights can I provide to CFC colleagues to drive engagement with HCPs?
• What is the optimal sequence of promotional activity to drive sales?

Key Business Questions

Developed on

Advance Analytics Pipeline

As a CFC:

• Insights & suggestions to use the right channel with the right content to reach out to Target HCP / Account at the right time in order to drive an optimal customer experience & maximize the impact

Brands

Audience/User CFCs

User Stories

Next Best Action Engine

Capability Overview The Next Best Action Engine is:

• Built on AI models that listen, assess & generate targeted recommendations
• Supplemented by business-rule-driven recommendations
• Integrated in their Veeva user interface

... to drive excellence in execution to meet the needs of HCPs & optimize outcomes. It comprises capturing brand/market-level requirements, curating a fit-for-purpose AI-engine & deploying a change-management program to provide relevant recommendations optimized to drive adoption and better outcomes.

HCP Adoption Ladder Analysis

Audience/User: Marketer, Sales

Purpose: Used to identify the historical Adoption Staging behavior, its evolution and drivers of the latter. Name of AIDA Analytic Capability: HCP Adoption Ladder Analysis Platform Developed on: VAW

Data Assets

• Which HCPs could address the unmet needs?
• What are the drivers of HCP prescribing behavior?

Key Business Questions

Developed on

Claims Data MDM (HCP/HCO) Onekey Data C360 / CEM Data

• Modelling Feature Clustering
• Variance Inflation Factor (VIF)
• Over and/or Under Sampling for balancing (SMOTE)
• Supervised Learning
• LR Regression/Decision Trees

Advance Analytics Pipeline

As the local FF activities Orchestration Lead:

• Understand which HCP Segments have a high probability of moving to Higher Value Segments and through which levels that I control such as FF volume of approach, channel of approach or content, to orchetrate the FF activities appropriately.

Brands

Audience/User Marketer, Sales

User Stories

HCP Headroom Driver Analysis

Capability OverviewThe HCP Headroom Drivers Analysis, allows us to understand/capture the main differentiators between HCPs of comparable/compatible HCP Value segments to understand what differentiates them and which of those HCPs can be moved to a more valuable Value Segment via interaction activity and content. Those differentiators can be their own characteristics and current activity within the TA, their Channel Affinities, our pior engagements on other brands, our prior engagements within this TA, etc. This analysis can be used for the following:

• Identify which HCPs are more likely to move to a higher value segment within the TA given an increase in FF activity or given specific content delivery to inform FF prioritization and approach to increase market share.
• Estimate potential growth opportunities within the TA.

• How do the unmet needs vary by geography?

Key Business Questions

• Correlation Matrix against KPI for Feature Selection
• Cluster Analysis (K-means, Hierachical)

Advance Analytics Pipeline

As the local BU lead:

• Understand which subnational regions showcase a higher Patient Need for a specific Disease Area, to appropriately allocate budget and activities to maximize the resulting impact.

Developed on

Data Assets Epidemiological, Demographic, Disease Prevalence, MDM & C360/CEM Sales

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Patient Need Subnational Region Archetyping

Capability OverviewThis Patient Need Subnational Region Archetyping Analysis, allows us to capture those location-based metrics that are most associated with increased Patient Needs in a specific Disease Area, such as Demographics, Comorbidities, Infectious Disease Dynamics, Socio-economic factors, Vaccinations, etc., and the naturally forming subnational region archetypes. This analysis can be used for the following:

• Identify areas of increased Patient Need for the specific TA, to focus MR and CFC operations where they could have the greatest potential impact.
• Tailor resource allocation and DTC regional marketing campaigns appropriately to those areas where there will be greater impact of activity.

HCP Adoption Ladder Analysis

Audience/User: Marketer, Sales

Purpose: Used to identify the historical Adoption Staging behavior, its evolution and drivers of the latter. Name of AIDA Analytic Capability: HCP Adoption Ladder Analysis Platform Developed on: VAW

• k-value analysis
• GLM / Baseline Models
• Feature Transformation
• SHAP & Univariate GLM
• Bayesian Models

Across Major brands inUS, IDM & EM Markets

• What are the ROIs/Impact of promotion channels?
• What is the responsiveness of different channels & the optimal frequency to optimize revenue / profitability?

Key Business Questions

Advance Analytics Pipeline

As the local BU Lead/ Sales Lead / Marketing Lead

• Understand the impact of promotion across channels to better allocate brand budgets & resources to maximize overall sales / profitability

Developed on

Data Assets C360 Promotional costs / budgets

Brands

Audience/User BU Lead, Sales, Marketing

User Stories

Promotional Impact / Mix Modeling & mROI

Capability Overview Promotional Impact Assessment involves evaluating the relationship between promotional activity at brand-level & sales, to:

• Understand the level of overall sales impacted by promotion
• Attribution of those impactable sales to channels / segments
• mROI at channel-level

... in order to further optimize promotional spend / mix in the future.

• How do the unmet needs vary by geography?

Key Business Questions

• Trained Logistic Regression Models

Advance Analytics Pipeline

As the local BU Lead:

• Understand which activities could have a greater affect on TA or Business related KPIs to understand where Medical or Commercial needs to focus more on to lead to the greatest impact.

Developed on

Data Assets Trained Driver Models for patent journey, HCP behavior Analytics

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Driver-to-KPI Impact Simulation Analysis

Capability OverviewThe Driver-to-KPI Impact Simulation is a meta-analysis appropriate for all the driver analytics, allowing us to capture which factors affecting an outcome (whether that is Diagnosis, Treatment Initiation, Selection or Compliance, HCP Adoption or Value), could have the greatest potential effect to the associated KPIs via simulation.

Name of AIDA Analytic Capability:Treatment Selection Analysis Platform Developed on: VAW

Treatment Selection Analysis

Audience/User: Marketer, Leadership, Sales

Purpose:Identifying what drives Treatment Selection (whether a patient will be treated with product A vs product B), considering Demographic, Clinical, Insurance, Patient Journey, Exogenous and Provider factors. Useful for inferring product preference by certain factors and perceived product appropriateness.Could be used for content tailoring towards Customers or for identifying population microsegments for targetting via DTC

Data Assets

• What are unmet opportunities in the market?
• How important/large are these needs/opportunities?
• What are the drivers of Brand Choice?

Key Business Questions

• Modelling Feature Clustering
• Multicollinearity Elimination
• Patient Clustering and Cluster Definition
• Over and/or Under Sampling for balancing
• Multi-stage Modelling of variable categories
• LR Regression/Decision Trees

Advance Analytics Pipeline

As the local Marketing Lead:

• Understand which Patient Archetypes are switching to/from Treatment Class or Pathway PFE operates to tailor material for DTC marketing campaigns focused on archetypes or to tailor content delivered to HCPs in those archetypes

As the Medical Affairs local lead:

• Understand which major archetypes affect Treatment Switching intra/inter-class, to ascertain validity of switch from a medical perspective and tailor awareness campaigns' content and targetting activities of MRs to those HCPs.

Developed on

EHR Data Claims Data Census Data

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Treatment Switching Drivers Analysis

Capability OverviewTreatment Switching/Churn Drivers Analysis allows us to understand which factors affect whether a patient will remain on a specific treatment pathway or whether they will move to a different pathway within class or to another treatment class. This analysis can be used for the following:

• Identify Patient Archetypes affecting switching treatment to/from our preferred product/treatment class
• Map archetypes back to HCPs/Subnational Regions to understand distribution for each HCP and location
• Tailor content appropriately to grow share of product in those HCPs overindexing in archetypes that switch to our product or treatment class or to defend share of product in those HCPs overindexing in archetypes that switch away from our product or treatment class.
• Tailor DTC marketing campaigns' content for underperforming patient archetypes.

Data Assets

• What are unmet opportunities in the market?
• How important/large are these needs/opportunities?
• What are the drivers of Brand Choice?

Key Business Questions

• Modelling Feature Clustering
• Multicollinearity Elimination
• Patient Clustering and Cluster Definition
• Over and/or Under Sampling for balancing
• Multi-stage Modelling of variable categories
• LR Regression/Decision Trees

Advance Analytics Pipeline

As the local Marketing Lead:

• Understand which Patient Archetypes are switching to/from Treatment Class or Pathway PFE operates to tailor material for DTC marketing campaigns focused on archetypes or to tailor content delivered to HCPs in those archetypes

As the Medical Affairs local lead:

• Understand which major archetypes affect Treatment Switching intra/inter-class, to ascertain validity of switch from a medical perspective and tailor awareness campaigns' content and targetting activities of MRs to those HCPs.

Developed on

EHR Data Claims Data Census Data

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Treatment Switching Drivers Analysis

Capability OverviewTreatment Switching/Churn Drivers Analysis allows us to understand which factors affect whether a patient will remain on a specific treatment pathway or whether they will move to a different pathway within class or to another treatment class. This analysis can be used for the following:

• Identify Patient Archetypes affecting switching treatment to/from our preferred product/treatment class
• Map archetypes back to HCPs/Subnational Regions to understand distribution for each HCP and location
• Tailor content appropriately to grow share of product in those HCPs overindexing in archetypes that switch to our product or treatment class or to defend share of product in those HCPs overindexing in archetypes that switch away from our product or treatment class.
• Tailor DTC marketing campaigns' content for underperforming patient archetypes.

Data Assets

• What are unmet opportunities in the market?
• How important/large are these needs/opportunities
• What are the drivers of Brand Choice?

Key Business Questions

• Modelling Feature Clustering
• Multicollinearity Elimination
• Patient Clustering & Cluster Definition
• Over and/or Under Sampling for balancing
• Multi-stage Modelling of variable categories
• LR Regression/Decision Trees

Advance Analytics Pipeline

As a Local Medical Affairs Lead:

• Understand which clinical markers indicate a person is high-risk for having a disease/their relevant sequence as well as the HCPs along the patient's journey to be able to generate appropriate informational content and coordinate MR activities to increase awareness and diagnosis rates.

As a FF coordinator:

• Find appropriate HCPs to target that have a large number of high-risk patients currently undiagnosed at an earlier time in their patient journey to raise awareness and reduce the time-to-diagnosis.

Developed on

EHR data Claims Data

Brands

Audience/User Marketer, Leadership, Sales

User Stories

High Risk Patient Identification

Capability Overview The High Risk Patient Identification Analysis allows us to create models that estimate the probability an individual, given their clinical records, might have an undiagnosed disease. This model can be used to:

• Find Patients earlier in their clinical journey and track back to the HCPs to coordinate awareness reach-out on the disease by the FF, lead to earlier diagnosis and thus improved outcomes, better diagnosis rates and greater LTV of the patient.
• Make a sizing estimation of the market to understand estimated disease prevalance and not reported ones.
• Identify HCPs that have significant number of undiagnosed high risk patients to coordinate FF reach out.
• Identify key clinical markers and their sequence that would indicate the existence of the disease.

• How do the unmet needs vary by geography?

Key Business Questions

• Correlation Matrix against KPI for Feature Selection
• Cluster Analysis (K-means, Hierachical)

Advance Analytics Pipeline

As the local BU lead:

• Understand which subnational regions showcase a reduced Healthcare System readiness for a specific Disease Area, to appropriately orchestrate efforts to imrpove patient outcomes.

Developed on

Data Assets Epidemiological, Demographic, Disease Prevalence, MDM & C360/CEM Sales

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Healthcare Readiness Subnational Region Archetyping

Capability OverviewThe Healthcare Readiness Subnational Region Archetyping Analysis, allows us to capture those location-based metrics that are most associated with improved healthcare access and healthcare quality, such as ICU capacity, HCP-to-Patient ratio, ICU-to-Patient availbility etc., and the naturally forming Subnational Region Archetypes. This analysis can be used for the following:

• Identify areas that are better or less able to handle emerging or seasonal trends of certain disease areas to tailor tactics, content and weight of efforts of MR and CFC activities.

Agent-Based Modelling Disease Forecasting

Audience/User: Marketer, Leadership, Sales

Purpose: Used to estimate new variant of breakout behavior and amplitude conditional on disease characteristics and mobility information. Can provide a simulated expected outcome of the severity of new infectious diseases and to understand relative impact of diseases to coordinate FF activities, replenishment and even government outreach. Name of AIDA Analytic Capability: Agent-Based Modelling Disease Forecasting Platform Developed on: VAW

Name of AIDA Analytic Capability:Treatment Selection Analysis Platform Developed on: VAW

Treatment Selection Analysis

Audience/User: Marketer, Leadership, Sales

Purpose:Identifying what drives Treatment Selection (whether a patient will be treated with product A vs product B), considering Demographic, Clinical, Insurance, Patient Journey, Exogenous and Provider factors. Useful for inferring product preference by certain factors and perceived product appropriateness.Could be used for content tailoring towards Customers or for identifying population microsegments for targetting via DTC

Patient-Need Subnational Archetyping

Audience/User: Marketer, Leadership, Sales

Purpose: Name of AIDA Analytic Capability: Patient-Need Subnational Archetyping Platform Developed on: VAW

EHR DataClaims Data Experian Data Census Data

• What are unmet opportunities in the market?
• How important/large are these needs/opportunities?
• What are the drivers of Brand Choice?

Key Business Questions

• Modelling Feature Clustering
• Multicollinearity Elimination
• Patient Clustering and Cluster Definition
• Over and/or Under Sampling for balancing
• Multi-stage Modelling of variable categories
• LR Regression/Decision Trees

Advance Analytics Pipeline

As the local Marketing Lead:

• Understand which major Patient Archetypes are underperforming with respect to Treatment Rate to either tailor material for DTC marketing campaigns focused on those archetypes or to tailor content delivered to HCPs that service those archetypes.

As the Medical Affairs Local Lead:

• Understand which HCPs service Treatment Rate underperforming patient archetypes to coordinate MR reach out with appropriate content related to their patients.

Developed on

Data Assets

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Treatment Initiation Analysis

Capability Overview Treatment Initiation Drivers Analysis helps understand factors such as clinical characteristics, HCP attributes, SDOH or exogenous factors that affect whether a patient will receive treatment or not. Analysis can be used to:

• Identify Patient Archetypes/Associated Treatment Rates for underperforming groups and thus MUN.
• Map Patient Archetypes back to HCPs or Subnational regions to understand the distribution of those archetypes for each HCP and location.
• Tailor content appropriately for the underperforming archetypes and adjust FF activities to HCPs that overindex on those archetypes that also seem to have low treatment rates.
• Tailor DTC marketing campaigns' content for underperforming patient archetypes.

Treatment Adherence & Persistence Analysis

Audience/User: Marketer, Leadership, Sales

Purpose: Used to understand what drives treatment Adherence and/or Peristency. Could be used for HCP targetting, content tailoring or to understand emerging trends and tailoring actions to increase the LTV of patients. Predominantly useful on chronic or long-term treatment. Name of AIDA Analytic Capability: Treatment Adherence & Persistence Analysis Platform Developed on: VAW

Agent-Based Modelling Disease Forecasting

Audience/User: Marketer, Leadership, Sales

Purpose: Used to estimate new variant of breakout behavior and amplitude conditional on disease characteristics and mobility information. Can provide a simulated expected outcome of the severity of new infectious diseases and to understand relative impact of diseases to coordinate FF activities, replenishment and even government outreach. Name of AIDA Analytic Capability: Agent-Based Modelling Disease Forecasting Platform Developed on: VAW

Data Assets

• What are unmet opportunities in the market?
• How important/large are these needs/opportunities?
• What are the drivers of Brand Choice?

Key Business Questions

• Modelling Feature Clustering
• Multicollinearity Elimination
• Patient Clustering and Cluster Definition
• Over and/or Under Sampling for balancing
• Multi-stage Modelling of variable categories
• LR Regression/Decision Trees

Advance Analytics Pipeline

As the local Marketing Lead:

• Understand which are major Patient Archetypes that are underperforming with respect to Product Market Share to either tailor material for DTC marketing campaigns or to tailor content delivered to HCPs that service those archetypes

As the Medical Affairs local lead:

• Understand which major archetypes exist affecting Treatment Class or intra-class selection, to ascertain whether they receive the appropriate treatment class and tailor awareness campaigns content and HCP targeting of MRs.

Developed on

EHR Data Claims Data Experian Data Census Data

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Treatment Selection Drivers Analysis

Capability Overview Treatment Selection Drivers Analysis helps understand factors that affect if a patient will receive treatment with a specific product/class vs treatment with another product/class. It captures decision factors (clinical characteristics, HCP attributes, SDOH or exogenous) for inter/cross class treatment selection. This analysis can be used to:

• Identify prevalent Patient Archetypes affecting Treatment Class or Intra-Class Treatment selection
• Map archetypes to HCPs or Subnational Regions to understand distribution for each HCP and Location
• Tailor content appropriately to target underperforming archetypes in terms of market share and adjust FF activities to HCPs that overindex on those archetypes that also seem to have low market share
• Tailor DTC marketing campaigns' content for underperforming patient archetypes

Data Assets

• What are unmet opportunities in the market?
• How important/large are these needs/opportunities?
• What are the drivers of Brand Choice?

Key Business Questions

• Modelling Feature Clustering
• Multicollinearity Elimination

• Over and/or Under Sampling for balancing
• Multi-stage Modelling of variable categories
• LR Regression/Decision Trees
• Survival Analysis/Time-to-Event Analysis

Advance Analytics Pipeline

N/A

As a Local Medical Affairs Lead:

• Understand which factors affect time-to-diagnosis to tailor content and MR reach out efforts to reduce diagnosis lead time and improve health outcomes.

Developed on

EHR Data, Claims Data, Experian Data, Census Data

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Patient Diagnosis/Prognosis Drivers Analysis

Capability Overview The Patient Diagnosis/Prognosis Drivers Analysis allows us to create models that capture what affects time-to-diagnosis and progression to a more severe stage of the disease. These models can be used for the following:

• Understand what clinical, SDOH, exogenous or HCP characteristics seem to accelerate or delay time-to-diagnosis.
• Understand what clinical, observational or treatment and treatment compliance related characteristics might lead to a more severe stage of the disease area.
• Find which HCPs seem to have Patients at risk of progression to a more severe disease stage to coordinate awareness efforts.

Name of AIDA Analytic Capability:Treatment Selection Analysis Platform Developed on: VAW

Treatment Selection Analysis

Audience/User: Marketer, Leadership, Sales

Purpose:Identifying what drives Treatment Selection (whether a patient will be treated with product A vs product B), considering Demographic, Clinical, Insurance, Patient Journey, Exogenous and Provider factors. Useful for inferring product preference by certain factors and perceived product appropriateness.Could be used for content tailoring towards Customers or for identifying population microsegments for targetting via DTC

• How do the unmet needs vary by geography?

Key Business Questions

• Trained Logistic Regression Models

Advance Analytics Pipeline

As the local BU Lead:

• Understand which activities could have a greater affect on TA or Business related KPIs to understand where Medical or Commercial needs to focus more on to lead to the greatest impact.

Developed on

Data Assets Trained Driver Models for patent journey, HCP behavior Analytics

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Driver-to-KPI Impact Simulation Analysis

Capability OverviewThe Driver-to-KPI Impact Simulation is a meta-analysis appropriate for all the driver analytics, allowing us to capture which factors affecting an outcome (whether that is Diagnosis, Treatment Initiation, Selection or Compliance, HCP Adoption or Value), could have the greatest potential effect to the associated KPIs via simulation.

Census DataMobility Data Demographic Data

• How do the unmet needs vary by geography?

Key Business Questions

• Agent-Based Simulation Models

Advance Analytics Pipeline

As the local Medical Affairs Lead

• Understand what disease impact a new variant of a disease or a brand new disease will have on patients and the countries healthcare system to organize MR activities and disease awareness to address the impact more effectively.

Developed on

Data Assets

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Agent-Based Infectious Disease Modelling

Capability OverviewThe Agent-based Infectious Disease modelling analysis, allows us to simulate the effects of an infectious disease when that disease is a significantly different derivation/evolution of a prior disease (a variant that has differnt R0 or skips prior immunity) or when historical information is non-available due to a completely new disease. The analysis can be used for the following:

• Identify areas that due their characteristics (population density, demographics, population mobility to and from) will showcase a significantly different behavior than other areas either in terms of lagging effect of effect amplitude to orchestrate and allocate efforts appropriately.
• Estimate the effect an emerging variant or disease will have on the underlying population to orchestrate and allocate efforts and resources appropriately.

Patient-Need Subnational Archetyping

Audience/User: Marketer, Leadership, Sales

Purpose: Name of AIDA Analytic Capability: Patient-Need Subnational Archetyping Platform Developed on: VAW

Data Assets

• What are unmet opportunities in the market?
• How important/large are these needs/opportunities?
• What are the drivers of Brand Choice?

Key Business Questions

• Modelling Feature Clustering
• Multicollinearity Elimination
• Patient Clustering and Cluster Definition
• Over and/or Under Sampling for balancing
• Multi-stage Modelling of variable categories
• LR Regression/Decision Trees

Advance Analytics Pipeline

As the local Marketing Lead:

• Understand which are major Patient Archetypes that are underperforming with respect to Product Market Share to either tailor material for DTC marketing campaigns or to tailor content delivered to HCPs that service those archetypes

As the Medical Affairs local lead:

• Understand which major archetypes exist affecting Treatment Class or intra-class selection, to ascertain whether they receive the appropriate treatment class and tailor awareness campaigns content and HCP targeting of MRs.

Developed on

EHR Data Claims Data Experian Data Census Data

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Treatment Selection Drivers Analysis

Capability Overview Treatment Selection Drivers Analysis helps understand factors that affect if a patient will receive treatment with a specific product/class vs treatment with another product/class. It captures decision factors (clinical characteristics, HCP attributes, SDOH or exogenous) for inter/cross class treatment selection. This analysis can be used to:

• Identify prevalent Patient Archetypes affecting Treatment Class or Intra-Class Treatment selection
• Map archetypes to HCPs or Subnational Regions to understand distribution for each HCP and Location
• Tailor content appropriately to target underperforming archetypes in terms of market share and adjust FF activities to HCPs that overindex on those archetypes that also seem to have low market share
• Tailor DTC marketing campaigns' content for underperforming patient archetypes

• How do the unmet needs vary by geography?

Key Business Questions

• Correlation Matrix against KPI for Feature Selection
• Cluster Analysis (K-means, Hierachical)

Advance Analytics Pipeline

As the local BU lead:

• Understand which subnational regions showcase a higher Patient Need for a specific Disease Area, to appropriately allocate budget and activities to maximize the resulting impact.

Developed on

Data Assets Epidemiological, Demographic, Disease Prevalence, MDM & C360/CEM Sales

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Patient Need Subnational Region Archetyping

Capability OverviewThis Patient Need Subnational Region Archetyping Analysis, allows us to capture those location-based metrics that are most associated with increased Patient Needs in a specific Disease Area, such as Demographics, Comorbidities, Infectious Disease Dynamics, Socio-economic factors, Vaccinations, etc., and the naturally forming subnational region archetypes. This analysis can be used for the following:

• Identify areas of increased Patient Need for the specific TA, to focus MR and CFC operations where they could have the greatest potential impact.
• Tailor resource allocation and DTC regional marketing campaigns appropriately to those areas where there will be greater impact of activity.

• How do the unmet needs vary by geography?

Key Business Questions

• Aggregate data to national and patient archetype level
• Calculate longitudinal or regional profile
• Calculate total cases according to the infection rate
• Calculate cases per wastewater concentration unit
• Utilize wastewater concentration shape to calculate the cases profile
• Use cases/concentration unit to estimate true cases

Advance Analytics Pipeline

As the local BU Lead:

• Understand what are the future projections of the infecious disease in question to organize future activities appropriately

As the local FF orchestration lead:

• Understand which areas are projected to have a more sever viral load and what is the seasonal peaks and troughs of the projections to coordinate effort allocation and timing more appropriately to maximize the reached patients.

Developed on

Data Assets Longitudinal epidemiological Data, Wastewater concentration data

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Autoregressive/Wastewater-based Infectious Disease Forecasting

Capability OverviewAutoregressive/Wastewater-based Infectious Disease Forecasting captures the seasonal evolution of infectious disease by analyzing its historical datapoints (when undereporting is non time-variant and stable) or by leveraging viral weight in wastewater concentration datapoints, and project in the future. This can be used to:

• Identify the seasonal peaks and troughs of an infectious disease to better coordinate FF activities' timing and volume to maximize patient reach out and treatment/awareness.
• Calculate the effect that well documented events have on the underlying level, trend and seasonality of the disease to estimate the expected impact of an infectious disease overall.
• Estimate the attribution of national effect to subnational regions where disease metrics (reported cases, hospitalizations, deaths from disease) or wastewater metrics are available

Patient-Need Subnational Archetyping

Audience/User: Marketer, Leadership, Sales

Purpose: Name of AIDA Analytic Capability: Patient-Need Subnational Archetyping Platform Developed on: VAW

• How do the unmet needs vary by geography?

Key Business Questions

• Trained Logistic Regression Models

Advance Analytics Pipeline

As the local BU Lead:

• Understand which activities could have a greater affect on TA or Business related KPIs to understand where Medical or Commercial needs to focus more on to lead to the greatest impact.

Developed on

Data Assets Trained Driver Models for patent journey, HCP behavior Analytics

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Driver-to-KPI Impact Simulation Analysis

Capability OverviewThe Driver-to-KPI Impact Simulation is a meta-analysis appropriate for all the driver analytics, allowing us to capture which factors affecting an outcome (whether that is Diagnosis, Treatment Initiation, Selection or Compliance, HCP Adoption or Value), could have the greatest potential effect to the associated KPIs via simulation.

Major Brands acrossUS & EU region along with Japan

C360 CDM & NBA Consumption / Application Layer

• Customer Value Index (xGBoost with normalization)
• Verso OOTB CNN & GA models for optimal sequence generation / recommendation

Data Assets

• What recommendations/insights can I provide to CFC colleagues to drive engagement with HCPs?
• What is the optimal sequence of promotional activity to drive sales?

Key Business Questions

Developed on

Advance Analytics Pipeline

As a CFC:

• Insights & suggestions to use the right channel with the right content to reach out to Target HCP / Account at the right time in order to drive an optimal customer experience & maximize the impact

Brands

Audience/User CFCs

User Stories

Next Best Action Engine

Capability Overview The Next Best Action Engine is:

• Built on AI models that listen, assess & generate targeted recommendations
• Supplemented by business-rule-driven recommendations
• Integrated in their Veeva user interface

... to drive excellence in execution to meet the needs of HCPs & optimize outcomes. It comprises capturing brand/market-level requirements, curating a fit-for-purpose AI-engine & deploying a change-management program to provide relevant recommendations optimized to drive adoption and better outcomes.

Data Assets

• What are unmet opportunities in the market?
• How important/large are these needs/opportunities?
• What are the drivers of Brand Choice?

Key Business Questions

• Modelling Feature Clustering
• Multicollinearity Elimination

• Over and/or Under Sampling for balancing
• Multi-stage Modelling of variable categories
• LR Regression/Decision Trees
• Survival Analysis/Time-to-Event Analysis

Advance Analytics Pipeline

N/A

As a Local Medical Affairs Lead:

• Understand which factors affect time-to-diagnosis to tailor content and MR reach out efforts to reduce diagnosis lead time and improve health outcomes.

Developed on

EHR Data, Claims Data, Experian Data, Census Data

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Patient Diagnosis/Prognosis Drivers Analysis

Capability Overview The Patient Diagnosis/Prognosis Drivers Analysis allows us to create models that capture what affects time-to-diagnosis and progression to a more severe stage of the disease. These models can be used for the following:

• Understand what clinical, SDOH, exogenous or HCP characteristics seem to accelerate or delay time-to-diagnosis.
• Understand what clinical, observational or treatment and treatment compliance related characteristics might lead to a more severe stage of the disease area.
• Find which HCPs seem to have Patients at risk of progression to a more severe disease stage to coordinate awareness efforts.

Data Assets

• Which HCPs could address the unmet needs?
• What are the drivers of HCP prescribing behavior?

Key Business Questions

Developed on

Claims, MDM/ Onekey, C360/CEM, HCP Channel Affinities / Adoption Segments

• Modelling Feature Clustering
• Multi-collinearity Elimination
• Over and/or Under Sampling for balancing (SMOTE)
• Logistic Regression based modelling and quantification of effect of independent variable to dependent variable

Advance Analytics Pipeline

As the local FF activities Orchestration Lead:

• Understand which HCP Segments and specific HCPs within them have a high probability on transitioning to a higher or lower value Adoption Stage and orchestrate content and FF reach out to address the emerging Opportunities or Threats.

Brands

Audience/User Marketer, Sales

User Stories

HCP Brand Loyalty/Churn Analysis

Capability OverviewThis HCP Brand Loyalty/Churn Drivers Analysis, allows us to understand/capture the main differentiators between HCP of associated Adoption Stages (Aware to Trial, Trial to Usage while active in Class, Tria to Usage while active in prior Class of Treatment, etc.). Those differentiators may be FF engagement, exogenous events, HCP Specialty or Patient Archetypes evolution, delivered content among a few. This analysis can be used for the following:

• Identify which HCPs are likely to switch to or away from a particular Adoption Stage of interest and appropriate orchestrate FF activities.
• Identify tactically which factors can be affected to facilitate or prevent Churn to a particular Adoption Stage of interest to orchistrate FF activity and improve KPIs related to Adoption.

Across Brands in Europe, LATAM, Gulf regions along with Netherlands & Saudi Arabia

Data Assets

• Which HCPs are interested in F2F promotion and have engaged well with all Pfizer channels?

• How can I define a target list to optimize HQ-level & FF-driven promotion

Key Business Questions

Developed on

C360 MDM

• Cluster Analysis (K-means, Hierachical, business-rules driven)

Advance Analytics Pipeline

As a Sales Lead / Marketer

• Understand the channel affinities of HCPs to tailor promotional activity to their channels of preference, thereby optimizing customer experience & driving overall sales

Brands

Audience/User Marketer, Sales

User Stories

Channel Affinity Segmentation

Capability OverviewChannel Affinity Segmentation looks at historical activity done with an HCP for a specific brand or overall, looks at Depth (Interaction & Engagement) & Consistency over a 6-,12- or 24-month period & generates a channel score for each HCP by summing normalized depth & consistency. Different clustering models are applied to create segments of customers who have similar affinity to one or more channels & can then be manually combined / tweaked further & renamed.

• How do the unmet needs vary by geography?

Key Business Questions

• Features’ Hierarchical clustering
• Multicollinearity analysis
• Multiple Clustering analyses (matrix approach to segmentation)

Advance Analytics Pipeline

As the local marketing lead:

• Understand which Consumer Behavior Segments are most appropriate to the disease area in question and their distribution across channels and regions to tailor content and channel investment appropriately.

Developed on

Data Assets Experian Data Demographics & Census Data

Brands

Audience/User Marketer, Sales

User Stories

Consumer Archetyping for DTC Marketing

Capability OverviewThe Consumer Behavior Archetyping for DTC Marketing Analysis, allows us to capture those Consumer characteristics that might affect their responsiveness or preferred channel to DTC marketing campaigns. This analysis can be used for the following:

• Identify the approach and channel investment per subnational region (or nationally) tailored to those channels and content most appropriate for the consumer segment archetypes present in those areas.

Data Assets

• Which HCPs could address the unmet needs?
• What are the drivers of HCP prescribing behavior?

Key Business Questions

Developed on

Claims, MDM/ Onekey, C360/CEM, HCP Channel Affinities / Adoption Segments

• Modelling Feature Clustering
• Multi-collinearity Elimination
• Over and/or Under Sampling for balancing (SMOTE)
• Logistic Regression based modelling and quantification of effect of independent variable to dependent variable

Advance Analytics Pipeline

As the local FF activities Orchestration Lead:

• Understand which HCP Segments and specific HCPs within them have a high probability on transitioning to a higher or lower value Adoption Stage and orchestrate content and FF reach out to address the emerging Opportunities or Threats.

Brands

Audience/User Marketer, Sales

User Stories

HCP Brand Loyalty/Churn Analysis

Capability OverviewThis HCP Brand Loyalty/Churn Drivers Analysis, allows us to understand/capture the main differentiators between HCP of associated Adoption Stages (Aware to Trial, Trial to Usage while active in Class, Tria to Usage while active in prior Class of Treatment, etc.). Those differentiators may be FF engagement, exogenous events, HCP Specialty or Patient Archetypes evolution, delivered content among a few. This analysis can be used for the following:

• Identify which HCPs are likely to switch to or away from a particular Adoption Stage of interest and appropriate orchestrate FF activities.
• Identify tactically which factors can be affected to facilitate or prevent Churn to a particular Adoption Stage of interest to orchistrate FF activity and improve KPIs related to Adoption.

• How do the unmet needs vary by geography?

Key Business Questions

• Correlation Matrix against KPI for Feature Selection
• Cluster Analysis (K-means, Hierachical)

Advance Analytics Pipeline

As the local BU lead:

• Understand which subnational regions showcase a reduced Healthcare System readiness for a specific Disease Area, to appropriately orchestrate efforts to imrpove patient outcomes.

Developed on

Data Assets Epidemiological, Demographic, Disease Prevalence, MDM & C360/CEM Sales

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Healthcare Readiness Subnational Region Archetyping

Capability OverviewThe Healthcare Readiness Subnational Region Archetyping Analysis, allows us to capture those location-based metrics that are most associated with improved healthcare access and healthcare quality, such as ICU capacity, HCP-to-Patient ratio, ICU-to-Patient availbility etc., and the naturally forming Subnational Region Archetypes. This analysis can be used for the following:

• Identify areas that are better or less able to handle emerging or seasonal trends of certain disease areas to tailor tactics, content and weight of efforts of MR and CFC activities.

Agent-Based Modelling Disease Forecasting

Audience/User: Marketer, Leadership, Sales

Purpose: Used to estimate new variant of breakout behavior and amplitude conditional on disease characteristics and mobility information. Can provide a simulated expected outcome of the severity of new infectious diseases and to understand relative impact of diseases to coordinate FF activities, replenishment and even government outreach. Name of AIDA Analytic Capability: Agent-Based Modelling Disease Forecasting Platform Developed on: VAW

EHR DataClaims Data Experian Data Census Data

• What are unmet opportunities in the market?
• How important/large are these needs/opportunities?
• What are the drivers of Brand Choice?

Key Business Questions

• Modelling Feature Clustering
• Multicollinearity Elimination
• Patient Clustering and Cluster Definition
• Over and/or Under Sampling for balancing
• Multi-stage Modelling of variable categories
• LR Regression/Decision Trees

Advance Analytics Pipeline

As the local Marketing Lead:

• Understand which major Patient Archetypes are underperforming with respect to Treatment Rate to either tailor material for DTC marketing campaigns focused on those archetypes or to tailor content delivered to HCPs that service those archetypes.

As the Medical Affairs Local Lead:

• Understand which HCPs service Treatment Rate underperforming patient archetypes to coordinate MR reach out with appropriate content related to their patients.

Developed on

Data Assets

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Treatment Initiation Analysis

Capability Overview Treatment Initiation Drivers Analysis helps understand factors such as clinical characteristics, HCP attributes, SDOH or exogenous factors that affect whether a patient will receive treatment or not. Analysis can be used to:

• Identify Patient Archetypes/Associated Treatment Rates for underperforming groups and thus MUN.
• Map Patient Archetypes back to HCPs or Subnational regions to understand the distribution of those archetypes for each HCP and location.
• Tailor content appropriately for the underperforming archetypes and adjust FF activities to HCPs that overindex on those archetypes that also seem to have low treatment rates.
• Tailor DTC marketing campaigns' content for underperforming patient archetypes.

Data Assets

• Which HCPs could address the unmet needs?
• What are the drivers of HCP prescribing behavior?

Key Business Questions

• TRx/NRx/Patient Number deciling and classification
• Unsupervised Learning Models (K-Means/ Hierarchical Clustering)

Advance Analytics Pipeline

As the local FF activities Orchestration Lead:

• Understand current Value HCPs driving Sales/ Market Share to understand which HCPs FF needs to prioritize in maintaining a relationship with

As the local FF activities Orchestration Lead:

• Understand potential value behaviors/ segments such as HCPs of higher TA Patient Share, HCPs operating in the treatment class who are splitting prescriptions or HCPs that operate in a Treatment Class of high relevance to the Treatment Class Pfizer operates in and are referencing high volume of patients to inform the approach and targetting of FF.

Developed on

Claims Data MDM (HCP/HCO) Onekey Data C360 / CEM Data

Brands

Audience/User Marketer, Sales

User Stories

HCP Value Segmentation Analysis

Capability OverviewHCP Value Segmentation Analysis allows us to understand/capture important Segments of HCPs with respect to their prescribing behavior in terms of rendering volume, pre-rending referencing, brand prescription share within the treatment class/prior treatment class than where Pfizer operates in. This analysis can be used to:

• Identify which HCPs are splitting within the treatment class where we operate.
• Identify which HCPs belong in higher deciles with respect to the volume share of patients within the TA.
• Identify which HCPs operate in Treatment Classes preceding the one where we operate and have a high reference out volume for subsequent Treatment Classes.
• Tailor prioritity and method/content of approach to those HCPs based on the above.

Data Assets

• Which HCPs could address the unmet needs?
• What are the drivers of HCP prescribing behavior?

Key Business Questions

• TRx/NRx/Patient Number deciling and classification
• Unsupervised Learning Models (K-Means/ Hierarchical Clustering)

Advance Analytics Pipeline

As the local Sales Lead:

• Understand the levels of growth in terms of Product Share that are achievable focusing on HCPs

As the local FF activities Orchestration Lead:

• Understand which HCP belong to segments and action groups that belong to Adoption Ladder segments that are appropriate for targeting which could result in market share and growth increase and what is their current status to tailor approach and content appropriately.

Developed on

Claims Data MDM (HCP/HCO) Onekey Data C360 / CEM Data

Brands

Audience/User Marketer, Sales

User Stories

HCP Adoption Segmentation Analysis

Capability OverviewThe HCP Adoption Drivers Analysis allows us to understand/capture what are the levels of adoption (Aware, Trial, Usage, Loyalty, etc.) currently in the TA across the HCPs operating in the space. This analysis can be used to:

• Identify what are the main Adoption Archetypes in the market, grounded on business relevant groups (Primary Treatment Only, Primary/Secondary Treatment, Specific Treatment Class focused etc.)
• Juxtapose with other HCP focused analysis (i.e. HCP Value Segmentation, HCP Brand Loyalty/Churn Analysis) to create action groups for FF Approach and Prioritization.
• Capture evolution (emerging and subsiding trents) of those segments to evaluate opportunity for growth and groups needed for defence.

• N/A

Across all Markets

• How do the unmet needs vary by geogr aphy?

Key Business Questions

Advance Analytics Pipeline

As a Regional President

• Optimize allocation of resources across brands / countries to improve overall sales / profitability

As a Country President

• Optimize allocation of resources across brands / therapy areas, to improve overall sales / profitability

As a TA / BU Lead

• Optimize allocation of resources across brands in my TA to improve overall sales / profitability

Developed on

Data Assets Finance, Forecasts, margins, C360, Other country planning inputs

Brands

Audience/User Regional President / Country President / BU or TA Lead

User Stories

Budget Allocation (Portfolio / Cross-brand / Cross-market)

Capability Overview Budget Allocation exercises involve using financial data such as revenues, margins, growth / evolution, market shares & benchmark-based response curves / channel-level impact estimates to directionally estimate the optimal allocation of spend across brands & markets. They are usually done at a regional-level across markets or combinations of brands / markets in order to prioritze where to reallocate budgets from & to.

• How do the unmet needs vary by geography?

Key Business Questions

• Trained Logistic Regression Models

Advance Analytics Pipeline

As the local Medical Affairs lead:

• Observe which factors affect core medical KPIs (Treatment Initiation, Selection, Compliance and Diagnosis) to understand what are the emerging priorities for MR activities that need to be focused in the future

As the local FF Orchestration Lead:

• Observe factors that affect core commercial/ marketing KPIs (Adoption Staging) to understand what are emerging priorities for MR activities that need to be focused on and whether activities past and present have an effect on established factors.

Developed on

Data Assets Trained Driver Models for patent journey, HCP behavior Analytics

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Driver Temporal Evolution Analysis

Capability OverviewThis Driver Temporal Evolution Analysis is a meta-analysis appropriate for all the driver analytics, allowing us to capture how all the factors affecting an outcome (whether that is Diagnosis, Treatment, Treatment Selection, Adoption Churn Behavior, Value Churn, Treatment Switching, Treatment Compliance, etc.), change over time, either having an increased or decreasing effect. It is used to observe emerging or subsiding trends and to identify events that might have affected those behaviors.

• How do the unmet needs vary by geography?

Key Business Questions

• Trained Logistic Regression Models

Advance Analytics Pipeline

As the local Medical Affairs lead:

• Observe which factors affect core medical KPIs (Treatment Initiation, Selection, Compliance and Diagnosis) to understand what are the emerging priorities for MR activities that need to be focused in the future

As the local FF Orchestration Lead:

• Observe factors that affect core commercial/ marketing KPIs (Adoption Staging) to understand what are emerging priorities for MR activities that need to be focused on and whether activities past and present have an effect on established factors.

Developed on

Data Assets Trained Driver Models for patent journey, HCP behavior Analytics

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Driver Temporal Evolution Analysis

Capability OverviewThis Driver Temporal Evolution Analysis is a meta-analysis appropriate for all the driver analytics, allowing us to capture how all the factors affecting an outcome (whether that is Diagnosis, Treatment, Treatment Selection, Adoption Churn Behavior, Value Churn, Treatment Switching, Treatment Compliance, etc.), change over time, either having an increased or decreasing effect. It is used to observe emerging or subsiding trends and to identify events that might have affected those behaviors.

• How do the unmet needs vary by geography?

Key Business Questions

• Features’ Hierarchical clustering
• Multicollinearity analysis
• Multiple Clustering analyses (matrix approach to segmentation)

Advance Analytics Pipeline

As the local marketing lead:

• Understand which Consumer Behavior Segments are most appropriate to the disease area in question and their distribution across channels and regions to tailor content and channel investment appropriately.

Developed on

Data Assets Experian Data Demographics & Census Data

Brands

Audience/User Marketer, Sales

User Stories

Consumer Archetyping for DTC Marketing

Capability OverviewThe Consumer Behavior Archetyping for DTC Marketing Analysis, allows us to capture those Consumer characteristics that might affect their responsiveness or preferred channel to DTC marketing campaigns. This analysis can be used for the following:

• Identify the approach and channel investment per subnational region (or nationally) tailored to those channels and content most appropriate for the consumer segment archetypes present in those areas.

Data Assets

• What are unmet opportunities in the market?
• How important/large are these needs/opportunities?
• What are the drivers of Brand Choice?

Key Business Questions

• Modelling Feature Clustering
• Multicollinearity Elimination
• Patient Clustering and Cluster Definition
• Over and/or Under Sampling for balancing
• Multi-stage Modelling of variable categories
• LR Regression/Decision Trees

Advance Analytics Pipeline

As the local Marketing Lead:

• Understand which Patient Archetypes are switching to/from Treatment Class or Pathway PFE operates to tailor material for DTC marketing campaigns focused on archetypes or to tailor content delivered to HCPs in those archetypes

As the Medical Affairs local lead:

• Understand which major archetypes affect Treatment Switching intra/inter-class, to ascertain validity of switch from a medical perspective and tailor awareness campaigns' content and targetting activities of MRs to those HCPs.

Developed on

EHR Data Claims Data Census Data

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Treatment Switching Drivers Analysis

Capability OverviewTreatment Switching/Churn Drivers Analysis allows us to understand which factors affect whether a patient will remain on a specific treatment pathway or whether they will move to a different pathway within class or to another treatment class. This analysis can be used for the following:

• Identify Patient Archetypes affecting switching treatment to/from our preferred product/treatment class
• Map archetypes back to HCPs/Subnational Regions to understand distribution for each HCP and location
• Tailor content appropriately to grow share of product in those HCPs overindexing in archetypes that switch to our product or treatment class or to defend share of product in those HCPs overindexing in archetypes that switch away from our product or treatment class.
• Tailor DTC marketing campaigns' content for underperforming patient archetypes.

EHR dataClaims Data

• What are unmet opportunities in the market?
• How important/large are these needs/opportunities?
• What are the drivers of Brand Choice?

Key Business Questions

• N/A

Advance Analytics Pipeline

As a Local BU Lead:

• Understand which Patient Journey end-points exist in the disease area in question, which of those are servicable by our products/which are servicable by competitor in-line or prelaunch to better coordinate activities to defend or grow our product's share

As a FF coordinator:

• Understand which Referral Transition nodes and the associated HCPs to target to ensure that more patients can get treatment on time with our product

Developed on

Data Assets

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Patient Journey Analysis

Capability Overview The Patient Journey Analysis shows a complete view of the Patient's Journey in a dendrical structure, with end nodes representing various endpoints of their journey and intermediary nodes their transitions to them. End-points and transitions allow us to understand untapped potential/threats in juxtaposition with our product's and the competitor's coverage of end-points and where along the journey our efforts could have the most impact to commercial and medical outcomes/KPIs. Examples of endpoints and transitions: Patients with delayed Treatment, Patients without Treatment, Patient Reference prior to Diagnosis, Patient Reference prior to Treatment, Non-Adherant/Persistent Patients, Patients where Primary Treatment sufficed, Patients that moved from Primary Treatment to Secondary Treatment in X months, Patients that moved from X Treatment Class to Y Treatment Class.

Patient-Need Subnational Archetyping

Audience/User: Marketer, Leadership, Sales

Purpose: Name of AIDA Analytic Capability: Patient-Need Subnational Archetyping Platform Developed on: VAW

• How do the unmet needs vary by geography?

Key Business Questions

• Correlation Matrix against KPI for Feature Selection
• Cluster Analysis (K-means, Hierachical)

Advance Analytics Pipeline

As the local BU lead:

• Understand which subnational regions showcase a reduced Healthcare System readiness for a specific Disease Area, to appropriately orchestrate efforts to imrpove patient outcomes.

Developed on

Data Assets Epidemiological, Demographic, Disease Prevalence, MDM & C360/CEM Sales

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Healthcare Readiness Subnational Region Archetyping

Capability OverviewThe Healthcare Readiness Subnational Region Archetyping Analysis, allows us to capture those location-based metrics that are most associated with improved healthcare access and healthcare quality, such as ICU capacity, HCP-to-Patient ratio, ICU-to-Patient availbility etc., and the naturally forming Subnational Region Archetypes. This analysis can be used for the following:

• Identify areas that are better or less able to handle emerging or seasonal trends of certain disease areas to tailor tactics, content and weight of efforts of MR and CFC activities.

Data Assets

• What are unmet opportunities in the market?
• How important/large are these needs/opportunities
• What are the drivers of Brand Choice?

Key Business Questions

• Modelling Feature Clustering
• Multicollinearity Elimination
• Patient Clustering & Cluster Definition
• Over and/or Under Sampling for balancing
• Multi-stage Modelling of variable categories
• LR Regression/Decision Trees

Advance Analytics Pipeline

As a Local Medical Affairs Lead:

• Understand which clinical markers indicate a person is high-risk for having a disease/their relevant sequence as well as the HCPs along the patient's journey to be able to generate appropriate informational content and coordinate MR activities to increase awareness and diagnosis rates.

As a FF coordinator:

• Find appropriate HCPs to target that have a large number of high-risk patients currently undiagnosed at an earlier time in their patient journey to raise awareness and reduce the time-to-diagnosis.

Developed on

EHR data Claims Data

Brands

Audience/User Marketer, Leadership, Sales

User Stories

High Risk Patient Identification

Capability Overview The High Risk Patient Identification Analysis allows us to create models that estimate the probability an individual, given their clinical records, might have an undiagnosed disease. This model can be used to:

• Find Patients earlier in their clinical journey and track back to the HCPs to coordinate awareness reach-out on the disease by the FF, lead to earlier diagnosis and thus improved outcomes, better diagnosis rates and greater LTV of the patient.
• Make a sizing estimation of the market to understand estimated disease prevalance and not reported ones.
• Identify HCPs that have significant number of undiagnosed high risk patients to coordinate FF reach out.
• Identify key clinical markers and their sequence that would indicate the existence of the disease.

• How do the unmet needs vary by geography?

Key Business Questions

• Features’ Hierarchical clustering
• Multicollinearity analysis
• Multiple Clustering analyses (matrix approach to segmentation)

Advance Analytics Pipeline

As the local marketing lead:

• Understand which Consumer Behavior Segments are most appropriate to the disease area in question and their distribution across channels and regions to tailor content and channel investment appropriately.

Developed on

Data Assets Experian Data Demographics & Census Data

Brands

Audience/User Marketer, Sales

User Stories

Consumer Archetyping for DTC Marketing

Capability OverviewThe Consumer Behavior Archetyping for DTC Marketing Analysis, allows us to capture those Consumer characteristics that might affect their responsiveness or preferred channel to DTC marketing campaigns. This analysis can be used for the following:

• Identify the approach and channel investment per subnational region (or nationally) tailored to those channels and content most appropriate for the consumer segment archetypes present in those areas.

Data Assets

• Which HCPs could address the unmet needs?
• What are the drivers of HCP prescribing behavior?

Key Business Questions

Developed on

Claims Data MDM (HCP/HCO) Onekey Data C360 / CEM Data

• Modelling Feature Clustering
• Variance Inflation Factor (VIF)
• Over and/or Under Sampling for balancing (SMOTE)
• Supervised Learning
• LR Regression/Decision Trees

Advance Analytics Pipeline

As the local FF activities Orchestration Lead:

• Understand which HCP Segments have a high probability of moving to Higher Value Segments and through which levels that I control such as FF volume of approach, channel of approach or content, to orchetrate the FF activities appropriately.

Brands

Audience/User Marketer, Sales

User Stories

HCP Headroom Driver Analysis

Capability OverviewThe HCP Headroom Drivers Analysis, allows us to understand/capture the main differentiators between HCPs of comparable/compatible HCP Value segments to understand what differentiates them and which of those HCPs can be moved to a more valuable Value Segment via interaction activity and content. Those differentiators can be their own characteristics and current activity within the TA, their Channel Affinities, our pior engagements on other brands, our prior engagements within this TA, etc. This analysis can be used for the following:

• Identify which HCPs are more likely to move to a higher value segment within the TA given an increase in FF activity or given specific content delivery to inform FF prioritization and approach to increase market share.
• Estimate potential growth opportunities within the TA.

Major Brands acrossUS & EU region along with Japan

C360 CDM & NBA Consumption / Application Layer

• Customer Value Index (xGBoost with normalization)
• Verso OOTB CNN & GA models for optimal sequence generation / recommendation

Data Assets

• What recommendations/insights can I provide to CFC colleagues to drive engagement with HCPs?
• What is the optimal sequence of promotional activity to drive sales?

Key Business Questions

Developed on

Advance Analytics Pipeline

As a CFC:

• Insights & suggestions to use the right channel with the right content to reach out to Target HCP / Account at the right time in order to drive an optimal customer experience & maximize the impact

Brands

Audience/User CFCs

User Stories

Next Best Action Engine

Capability Overview The Next Best Action Engine is:

• Built on AI models that listen, assess & generate targeted recommendations
• Supplemented by business-rule-driven recommendations
• Integrated in their Veeva user interface

... to drive excellence in execution to meet the needs of HCPs & optimize outcomes. It comprises capturing brand/market-level requirements, curating a fit-for-purpose AI-engine & deploying a change-management program to provide relevant recommendations optimized to drive adoption and better outcomes.

• k-value analysis
• GLM / Baseline Models
• Feature Transformation
• SHAP & Univariate GLM
• Bayesian Models

Across Major brands inUS, IDM & EM Markets

• What are the ROIs/Impact of promotion channels?
• What is the responsiveness of different channels & the optimal frequency to optimize revenue / profitability?

Key Business Questions

Advance Analytics Pipeline

As the local BU Lead/ Sales Lead / Marketing Lead

• Understand the impact of promotion across channels to better allocate brand budgets & resources to maximize overall sales / profitability

Developed on

Data Assets C360 Promotional costs / budgets

Brands

Audience/User BU Lead, Sales, Marketing

User Stories

Promotional Impact / Mix Modeling & mROI

Capability Overview Promotional Impact Assessment involves evaluating the relationship between promotional activity at brand-level & sales, to:

• Understand the level of overall sales impacted by promotion
• Attribution of those impactable sales to channels / segments
• mROI at channel-level

... in order to further optimize promotional spend / mix in the future.

Data Assets

• Which HCPs could address the unmet needs?
• What are the drivers of HCP prescribing behavior?

Key Business Questions

• TRx/NRx/Patient Number deciling and classification
• Unsupervised Learning Models (K-Means/ Hierarchical Clustering)

Advance Analytics Pipeline

As the local Sales Lead:

• Understand the levels of growth in terms of Product Share that are achievable focusing on HCPs

As the local FF activities Orchestration Lead:

• Understand which HCP belong to segments and action groups that belong to Adoption Ladder segments that are appropriate for targeting which could result in market share and growth increase and what is their current status to tailor approach and content appropriately.

Developed on

Claims Data MDM (HCP/HCO) Onekey Data C360 / CEM Data

Brands

Audience/User Marketer, Sales

User Stories

HCP Adoption Segmentation Analysis

Capability OverviewThe HCP Adoption Drivers Analysis allows us to understand/capture what are the levels of adoption (Aware, Trial, Usage, Loyalty, etc.) currently in the TA across the HCPs operating in the space. This analysis can be used to:

• Identify what are the main Adoption Archetypes in the market, grounded on business relevant groups (Primary Treatment Only, Primary/Secondary Treatment, Specific Treatment Class focused etc.)
• Juxtapose with other HCP focused analysis (i.e. HCP Value Segmentation, HCP Brand Loyalty/Churn Analysis) to create action groups for FF Approach and Prioritization.
• Capture evolution (emerging and subsiding trents) of those segments to evaluate opportunity for growth and groups needed for defence.

• N/A

Across all Markets

• How do the unmet needs vary by geogr aphy?

Key Business Questions

Advance Analytics Pipeline

As a Regional President

• Optimize allocation of resources across brands / countries to improve overall sales / profitability

As a Country President

• Optimize allocation of resources across brands / therapy areas, to improve overall sales / profitability

As a TA / BU Lead

• Optimize allocation of resources across brands in my TA to improve overall sales / profitability

Developed on

Data Assets Finance, Forecasts, margins, C360, Other country planning inputs

Brands

Audience/User Regional President / Country President / BU or TA Lead

User Stories

Budget Allocation (Portfolio / Cross-brand / Cross-market)

Capability Overview Budget Allocation exercises involve using financial data such as revenues, margins, growth / evolution, market shares & benchmark-based response curves / channel-level impact estimates to directionally estimate the optimal allocation of spend across brands & markets. They are usually done at a regional-level across markets or combinations of brands / markets in order to prioritze where to reallocate budgets from & to.

Data Assets

• What are unmet opportunities in the market?
• How important/large are these needs/opportunities?
• What are the drivers of Brand Choice?

Key Business Questions

• Modelling Feature Clustering
• Multicollinearity Elimination
• Patient Clustering and Cluster Definition
• Over and/or Under Sampling for balancing
• Multi-stage Modelling of variable categories
• LR Regression/Decision Trees

As the local Marketing Lead:

• Understand which are major Patient Archetypes that are being non-compliant to tailor appropriate material for DTC marketing campaigns focuses on those archetypes or to tailor content delivered to HCPs that service those archetypes to increase Patient LTV.

As the Medical Affairs local lead:

• Understand which major archetypes with respect to Treatment Compliance to tailor MR content and efforts orchestration to increase Treatment Compliance and positive Treatment Outcomes.

Advance Analytics Pipeline

Developed on

EHR data Claims Data SPP Data Experian Data

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Treatment Compliance Drivers Analysis

Capability OverviewTreatment Compliance Drivers Analysis helps understand which factors affect whether a patient wil remain adherent and persistent to a particular treatment pathway or whether they will skip doses or discontinue treatment and not complete treatment in a compliant way. Those factors could be clinical, SDOH, HCP, Socio-Economic in aggregate, Insurance related or any that have been identified by the BU. This analysis can be used to:

• Identify the most prevalent Patient Archetypes affecting compliance, whether that is adherence or persistence.
• Map those archetypes back to HCPs or Subnational Regions to understand distribution of those archetypes for each HCP or location.
• Tailor content appropriately to the identified archetypes for improving compliance either via DTC marketing campaigns (Location anchored or not) or to HCPs that overindex non-compliant patient archetypes.

• How do the unmet needs vary by geography?

Key Business Questions

• Trained Logistic Regression Models

Advance Analytics Pipeline

As the local Medical Affairs lead:

• Observe which factors affect core medical KPIs (Treatment Initiation, Selection, Compliance and Diagnosis) to understand what are the emerging priorities for MR activities that need to be focused in the future

As the local FF Orchestration Lead:

• Observe factors that affect core commercial/ marketing KPIs (Adoption Staging) to understand what are emerging priorities for MR activities that need to be focused on and whether activities past and present have an effect on established factors.

Developed on

Data Assets Trained Driver Models for patent journey, HCP behavior Analytics

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Driver Temporal Evolution Analysis

Capability OverviewThis Driver Temporal Evolution Analysis is a meta-analysis appropriate for all the driver analytics, allowing us to capture how all the factors affecting an outcome (whether that is Diagnosis, Treatment, Treatment Selection, Adoption Churn Behavior, Value Churn, Treatment Switching, Treatment Compliance, etc.), change over time, either having an increased or decreasing effect. It is used to observe emerging or subsiding trends and to identify events that might have affected those behaviors.

Data Assets

• What are unmet opportunities in the market?
• How important/large are these needs/opportunities?
• What are the drivers of Brand Choice?

Key Business Questions

• Modelling Feature Clustering
• Multicollinearity Elimination
• Patient Clustering and Cluster Definition
• Over and/or Under Sampling for balancing
• Multi-stage Modelling of variable categories
• LR Regression/Decision Trees

Advance Analytics Pipeline

As the local Marketing Lead:

• Understand which are major Patient Archetypes that are underperforming with respect to Product Market Share to either tailor material for DTC marketing campaigns or to tailor content delivered to HCPs that service those archetypes

As the Medical Affairs local lead:

• Understand which major archetypes exist affecting Treatment Class or intra-class selection, to ascertain whether they receive the appropriate treatment class and tailor awareness campaigns content and HCP targeting of MRs.

Developed on

EHR Data Claims Data Experian Data Census Data

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Treatment Selection Drivers Analysis

Capability Overview Treatment Selection Drivers Analysis helps understand factors that affect if a patient will receive treatment with a specific product/class vs treatment with another product/class. It captures decision factors (clinical characteristics, HCP attributes, SDOH or exogenous) for inter/cross class treatment selection. This analysis can be used to:

• Identify prevalent Patient Archetypes affecting Treatment Class or Intra-Class Treatment selection
• Map archetypes to HCPs or Subnational Regions to understand distribution for each HCP and Location
• Tailor content appropriately to target underperforming archetypes in terms of market share and adjust FF activities to HCPs that overindex on those archetypes that also seem to have low market share
• Tailor DTC marketing campaigns' content for underperforming patient archetypes

Patient-Need Subnational Archetyping

Audience/User: Marketer, Leadership, Sales

Purpose: Name of AIDA Analytic Capability: Patient-Need Subnational Archetyping Platform Developed on: VAW

Agent-Based Modelling Disease Forecasting

Audience/User: Marketer, Leadership, Sales

Purpose: Used to estimate new variant of breakout behavior and amplitude conditional on disease characteristics and mobility information. Can provide a simulated expected outcome of the severity of new infectious diseases and to understand relative impact of diseases to coordinate FF activities, replenishment and even government outreach. Name of AIDA Analytic Capability: Agent-Based Modelling Disease Forecasting Platform Developed on: VAW

Across Brands in Europe, LATAM, Gulf regions along with Netherlands & Saudi Arabia

Data Assets

• Which HCPs are interested in F2F promotion and have engaged well with all Pfizer channels?

• How can I define a target list to optimize HQ-level & FF-driven promotion

Key Business Questions

Developed on

C360 MDM

• Cluster Analysis (K-means, Hierachical, business-rules driven)

Advance Analytics Pipeline

As a Sales Lead / Marketer

• Understand the channel affinities of HCPs to tailor promotional activity to their channels of preference, thereby optimizing customer experience & driving overall sales

Brands

Audience/User Marketer, Sales

User Stories

Channel Affinity Segmentation

Capability OverviewChannel Affinity Segmentation looks at historical activity done with an HCP for a specific brand or overall, looks at Depth (Interaction & Engagement) & Consistency over a 6-,12- or 24-month period & generates a channel score for each HCP by summing normalized depth & consistency. Different clustering models are applied to create segments of customers who have similar affinity to one or more channels & can then be manually combined / tweaked further & renamed.

Data Assets

• Which HCPs could address the unmet needs?
• What are the drivers of HCP prescribing behavior?

Key Business Questions

• TRx/NRx/Patient Number deciling and classification
• Unsupervised Learning Models (K-Means/ Hierarchical Clustering)

Advance Analytics Pipeline

As the local FF activities Orchestration Lead:

• Understand current Value HCPs driving Sales/ Market Share to understand which HCPs FF needs to prioritize in maintaining a relationship with

As the local FF activities Orchestration Lead:

• Understand potential value behaviors/ segments such as HCPs of higher TA Patient Share, HCPs operating in the treatment class who are splitting prescriptions or HCPs that operate in a Treatment Class of high relevance to the Treatment Class Pfizer operates in and are referencing high volume of patients to inform the approach and targetting of FF.

Developed on

Claims Data MDM (HCP/HCO) Onekey Data C360 / CEM Data

Brands

Audience/User Marketer, Sales

User Stories

HCP Value Segmentation Analysis

Capability OverviewHCP Value Segmentation Analysis allows us to understand/capture important Segments of HCPs with respect to their prescribing behavior in terms of rendering volume, pre-rending referencing, brand prescription share within the treatment class/prior treatment class than where Pfizer operates in. This analysis can be used to:

• Identify which HCPs are splitting within the treatment class where we operate.
• Identify which HCPs belong in higher deciles with respect to the volume share of patients within the TA.
• Identify which HCPs operate in Treatment Classes preceding the one where we operate and have a high reference out volume for subsequent Treatment Classes.
• Tailor prioritity and method/content of approach to those HCPs based on the above.

HCP Adoption Ladder Analysis

Audience/User: Marketer, Sales

Purpose: Used to identify the historical Adoption Staging behavior, its evolution and drivers of the latter. Name of AIDA Analytic Capability: HCP Adoption Ladder Analysis Platform Developed on: VAW

Data Assets

• Which HCPs could address the unmet needs?
• What are the drivers of HCP prescribing behavior?

Key Business Questions

Developed on

Claims Data MDM (HCP/HCO) Onekey Data C360 / CEM Data

• Modelling Feature Clustering
• Variance Inflation Factor (VIF)
• Over and/or Under Sampling for balancing (SMOTE)
• Supervised Learning
• LR Regression/Decision Trees

Advance Analytics Pipeline

As the local FF activities Orchestration Lead:

• Understand which HCP Segments have a high probability of moving to Higher Value Segments and through which levels that I control such as FF volume of approach, channel of approach or content, to orchetrate the FF activities appropriately.

Brands

Audience/User Marketer, Sales

User Stories

HCP Headroom Driver Analysis

Capability OverviewThe HCP Headroom Drivers Analysis, allows us to understand/capture the main differentiators between HCPs of comparable/compatible HCP Value segments to understand what differentiates them and which of those HCPs can be moved to a more valuable Value Segment via interaction activity and content. Those differentiators can be their own characteristics and current activity within the TA, their Channel Affinities, our pior engagements on other brands, our prior engagements within this TA, etc. This analysis can be used for the following:

• Identify which HCPs are more likely to move to a higher value segment within the TA given an increase in FF activity or given specific content delivery to inform FF prioritization and approach to increase market share.
• Estimate potential growth opportunities within the TA.

Data Assets

• Which HCPs could address the unmet needs?
• What are the drivers of HCP prescribing behavior?

Key Business Questions

• TRx/NRx/Patient Number deciling and classification
• Unsupervised Learning Models (K-Means/ Hierarchical Clustering)

Advance Analytics Pipeline

As the local Sales Lead:

• Understand the levels of growth in terms of Product Share that are achievable focusing on HCPs

As the local FF activities Orchestration Lead:

• Understand which HCP belong to segments and action groups that belong to Adoption Ladder segments that are appropriate for targeting which could result in market share and growth increase and what is their current status to tailor approach and content appropriately.

Developed on

Claims Data MDM (HCP/HCO) Onekey Data C360 / CEM Data

Brands

Audience/User Marketer, Sales

User Stories

HCP Adoption Segmentation Analysis

Capability OverviewThe HCP Adoption Drivers Analysis allows us to understand/capture what are the levels of adoption (Aware, Trial, Usage, Loyalty, etc.) currently in the TA across the HCPs operating in the space. This analysis can be used to:

• Identify what are the main Adoption Archetypes in the market, grounded on business relevant groups (Primary Treatment Only, Primary/Secondary Treatment, Specific Treatment Class focused etc.)
• Juxtapose with other HCP focused analysis (i.e. HCP Value Segmentation, HCP Brand Loyalty/Churn Analysis) to create action groups for FF Approach and Prioritization.
• Capture evolution (emerging and subsiding trents) of those segments to evaluate opportunity for growth and groups needed for defence.

• k-value analysis
• GLM / Baseline Models
• Feature Transformation
• SHAP & Univariate GLM
• Bayesian Models

Across Major brands inUS, IDM & EM Markets

• What are the ROIs/Impact of promotion channels?
• What is the responsiveness of different channels & the optimal frequency to optimize revenue / profitability?

Key Business Questions

Advance Analytics Pipeline

As the local BU Lead/ Sales Lead / Marketing Lead

• Understand the impact of promotion across channels to better allocate brand budgets & resources to maximize overall sales / profitability

Developed on

Data Assets C360 Promotional costs / budgets

Brands

Audience/User BU Lead, Sales, Marketing

User Stories

Promotional Impact / Mix Modeling & mROI

Capability Overview Promotional Impact Assessment involves evaluating the relationship between promotional activity at brand-level & sales, to:

• Understand the level of overall sales impacted by promotion
• Attribution of those impactable sales to channels / segments
• mROI at channel-level

... in order to further optimize promotional spend / mix in the future.

EHR DataClaims Data Experian Data Census Data

• What are unmet opportunities in the market?
• How important/large are these needs/opportunities?
• What are the drivers of Brand Choice?

Key Business Questions

• Modelling Feature Clustering
• Multicollinearity Elimination
• Patient Clustering and Cluster Definition
• Over and/or Under Sampling for balancing
• Multi-stage Modelling of variable categories
• LR Regression/Decision Trees

Advance Analytics Pipeline

As the local Marketing Lead:

• Understand which major Patient Archetypes are underperforming with respect to Treatment Rate to either tailor material for DTC marketing campaigns focused on those archetypes or to tailor content delivered to HCPs that service those archetypes.

As the Medical Affairs Local Lead:

• Understand which HCPs service Treatment Rate underperforming patient archetypes to coordinate MR reach out with appropriate content related to their patients.

Developed on

Data Assets

Brands

Audience/User Marketer, Leadership, Sales

User Stories

Treatment Initiation Analysis

Capability Overview Treatment Initiation Drivers Analysis helps understand factors such as clinical characteristics, HCP attributes, SDOH or exogenous factors that affect whether a patient will receive treatment or not. Analysis can be used to:

• Identify Patient Archetypes/Associated Treatment Rates for underperforming groups and thus MUN.
• Map Patient Archetypes back to HCPs or Subnational regions to understand the distribution of those archetypes for each HCP and location.
• Tailor content appropriately for the underperforming archetypes and adjust FF activities to HCPs that overindex on those archetypes that also seem to have low treatment rates.
• Tailor DTC marketing campaigns' content for underperforming patient archetypes.