Patients, healthcare providers (HCPs) and payers are critical stakeholders for pharma. Pharma companies are therefore keen to gather as much information about these customers as possible. Much of this information—such as demographic profiles and medical and prescription data—comes from third-party sources. Insights about preferences and attitudes come directly from customers via unblinded market research, direct surveys or other feedback mechanisms, such as voice calls. Other data is gathered through pharma’s interactions with customers, including sales calls and digital engagement. These elements are used to provide insights that shape pharma’s go-to-market strategy. 

So what’s missing? Holistic insights at an n=1 level that go beyond the “what” of customer behaviors and begin to address the “why” (needs, beliefs and motivations) and the “how” (preferences and biases). If we take HCPs as an example, the “why” includes attitudes toward patient care, treatment philosophies, product beliefs and the practice environment. Data to describe these factors might be available for a small sample of doctors through primary market research (PMR) but is not scalable to the universe. The “how” for HCPs includes engagement preferences as well as cognitive biases that shape their decision-making. Data to generate these insights is typically available from PMR or other engagement data, such as CRM. Unfortunately, that data only scratches the surface. What if we can do better? What if we can predict the “what,” “why” and “how” for every customer in a connected fashion, combining everything we know about a whole person? If we can achieve holistic insights at an n=1 level, the picture that emerges is richly nuanced and can guide smarter strategy and more meaningful engagement. 

At a high level, this involves two things: (1) pharma needs to make better use of the information they already have and (2) the industry needs to expand the art of the possible when it comes to understanding our customers. Let’s talk about the specific categories of changes in the following sections.

In most situations, medical and prescription data is seen as the primary source for quantitative insights at an n=1 level. PMR provides qualitative insights for big-picture strategy, and data on customer interactions is seen as execution data used to monitor engagement. Ownership of these data sources is typically fragmented across functional groups, which creates siloes in customer insight generation. To create a holistic view of each customer, we need to move beyond the traditional approach of seeing data as primary (qualitative) and secondary (quantitative) to managing it as a connected series of inputs from different sources: first-party data (e.g., PMR, direct surveys, voice calls, social), second-party data (e.g., CRM data, marketing data) and third-party data (e.g., medical and prescription data, electronic health records).


Other nontraditional data sources can also deepen our understanding of the “whys” behind HCP decisions. Institution-level guidelines—and how tightly those guidelines are enforced—have a huge bearing on physician decisions. Local social determinants of health (SDOH) such as patient demographics, income, education and food security are also strongly correlated with care delivery and outcomes. Beyond healthcare-related insights, consumer purchasing databases linked to HCPs can provide a glimpse of how the physician makes decisions beyond the clinical context.


In addition to gathering a broader range of data sources, we need to see customers through a wider lens. Today, we tend to view customers through a narrow, therapeutic-area-specific lens. However, most HCPs—including specialists—treat patients across a variety of conditions and comorbidities, and it follows that customer preferences are informed by cross-TA experiences. Therefore, to generate truly holistic insights, we also need to look at data across therapeutic areas (TAs). Oncology is a category that demonstrates how a cross-TA perspective can be useful. If we were to base our perception of an HCP on how they treat just one tumor type, we would be constrained by the tumor-specific nuances in diagnosis and treatment algorithms. We would miss out on important signals such as a preference for biomarker-driven therapies across tumor types. Looking for consistencies across the breadth of an HCP’s experience would give us much more confidence about behaviors and underlying attitudes.


Key actions: 

  • Move toward managing data holistically across first-, second- and third-party sources to create a connected and deeper view of customers
  • Understand the context surrounding customers by leveraging nontraditional data sources, such as SDOH and institutional characteristics
  • Break the narrow, TA-centric lens by incorporating cross-TA insights

Even when we think about data in a more connected and holistic fashion, we often run into two types of limitations: (1) insights are granular, but not holistic, as they lack a complete picture of the customer that includes their mindset, beliefs and needs, and (2) insights are holistic but not granular enough, as they are subject to limitations such as small sample sizes, respondent biases and concerns about projectability.


This is where artificial intelligence (AI) can help—it can help create new inferences about the customers at an n=1 level based on available data, and it can help scale insights available only for a subset of customers. There is a wide and continually expanding array of AI models that can mine the demographics, real-world data, engagement touchpoints and attitudinal signals that make up the customer information landscape. Today, pharma is using AI predominantly to drive execution via mechanisms such as next-best action (NBA). There is a big opportunity to use AI to drive new inferences and scale insights, leading to a clear competitive advantage by bridging gaps between strategy and execution.


Consider the following recent case study ZS conducted with a leading pharma company. The company launched a disease-modifying therapy for treating an autoimmune neurology condition in a highly competitive market with 15-plus products and generics. A traditional use of AI in this setting is to predict early adopting HCPs based on analog product data. We took a different approach where we used PMR data from a small sample of neurologists along with AI applied to multiple HCP-level data points to enhance our predictions of early adopters at an n=1 level. This required us to go beyond traditional AI to find newer, smarter ways to generate stable predictions using small, noisy data. The results were then abstracted to a microsegment level to drive messaging and tactics and used at an n=1 level to drive execution. This resulted in three times more writers and five times more new patients for this product launch.


Key actions:

  • Go beyond the traditional use of AI to drive execution and invest in AI to drive holistic insights
  • Actively experiment with using AI to both create new inferences using existing data (e.g., physician mindshare and treatment philosophies) and scale insights by using AI to project rich data from a subset of customers (e.g., using market research data to infer physician beliefs and cognitive biases)

To date, customer insights have largely had a clinical or brand focus. However, each physician is a person who also happens to be a doctor treating many conditions, one of which is the condition your brand is focused on. Even among specialists, many individual conditions account for less than 2% of the time a physician spends, per workday. Not only do we need to appreciate customers’ needs, beliefs, motivations and preferences, but we also need to think about the decision-making that takes place beyond that. We need to think about the cognitive biases and shortcuts that physicians take to make decisions. This will truly help us move beyond the “what” to the “why” and the “how.” This necessitates a broader use of behavioral insights and cognitive science.


Key actions: 

  • Think about the whole person as opposed to having a clinical or brand lens to customer insights 
  • Use behavioral science to not only understand beliefs and cognitive shortcuts people take to make decisions but also identify ways in which pharma can intervene and shape customer thinking

A key step is to make insight collection timelier and more agile because insights have a shelf life. Nowadays, many organizations are investing in higher levels of responsiveness, such as event-based triggers and dynamic targeting. These capabilities are hungry for real-time insights and predictive modeling. To compete in this environment, it’s important for pharma companies to accelerate insight collection and synthesis across the organization to avoid spoilage of insights.


It’s also necessary to learn from the data continuously so we can keep refining our inferences and predictions at an n=1 level. While this happens today in a more automated fashion for promotional execution related decisions (on which customer to target and via what channel and frequency), we also need to adopt a more automated approach in refining our insights on why a customer behaves a certain way and what we can do to influence those behaviors.


Key actions: 

  • Put all available insights under one roof and build a democratized capability to both gather facts and generate inferred insights about HCPs
  • Shift from project-based insight generation toward agile platforms that are always on and continuously building on rich data sources like real-world data

We need to turbocharge how we enable strategy and tactical planning today by providing deeper insights into the customer. To ensure there is a strong, connected approach to how we use our customer understanding end to end, the insights framework needs to be common across strategy and execution. This requires the following:

  • A nuanced understanding of the customer at an n=1 level—the “what” (behaviors and opportunities), the “why” (needs, beliefs and motivations) and the “how” (preferences and biases). This requires connected data across all sources, going beyond a clinical or brand lens to think about the whole person and leveraging AI in a scalable way to create new inferences and scale insights.
  • Putting mechanisms in place to ensure these insights can be leveraged to drive both strategy (by abstracting to a microsegment level) and execution (by using the more holistic n=1 insights). This requires a series of capabilities to be built both upstream and downstream in the commercial process to bring this to life.

To stay competitive, pharma companies must do more than just engage customers in a timely way with compelling messages via mechanisms such as NBA and dynamic targeting. The key to success in the years ahead will be to drive better connections between insights, strategy and execution. Companies that don’t take the time to invest in developing a holistic view of their customers will miss out on the nuanced opportunities that are critical to thrive in a competitive market.