There’s massive opportunity in the $1,200 billion world of pharmaceutical marketing, and substantive shifts are underway as big data increases scope and narrows focus.
Pharma marketers now face the challenge of combining high-level, abstract knowledge with personalized recommendations tailored to patient and health care provider (HCP) needs. Existing sales strategies and current categorization methods are no longer enough — instead, marketers must leverage predictive modeling and machine learning to understand and anticipate behavioral change, categorize uptake rates effectively and pinpoint ideal HCP connections.
Uptake Issues
In a market driven by data and tempered by individual requirements, marketers face an uphill battle: How do they connect with HCPs to create engagement and promote new pharmaceuticals effectively? As noted by one recent study, physicians face “cognitive dissonance” as they attempt to balance benefit claims made by sales reps against existing scientific evidence — and in the face of payer influence, in an era when payers are stronger than ever in dictating challenges and opportunities for patients. This creates a scenario where physicians are unsure what to think, which drives down overall sales numbers.
Meanwhile, there’s also a market shift happening as traditional mass-market sales are replaced by more nuanced and specialized approaches. In practice, this means many pharma reps are facing sales-number slides as consistent HCP clients reduce order amounts, while smaller-volume clients are often sidelined altogether.
Leveraging License Plates
Numbers matter. People matter. Typically, pharma marketers have connected the two using “scores” — numbers that predict the value of a given HCP customer. The problem? In a world driven by highly personalized health data, these scores lack the depth to be actionable or precise.
The solution is simple: applying to AI to find the right math to fit the business problem. By considering a host of HCP features — including geographical location, patient demographics, historic order volumes and personal sales preferences — it’s possible to create “license plates” for providers, with each digit representing a specific feature score. By comparing hundreds or thousands of features, companies can evaluate the potential sales value of HCPs along with their preferred method of interaction. Even better, the math doesn’t care that assessing value in three-dimensional space is almost impossible; comparison across license plates lets pharma reps better understand HCP needs and develop strategies to mitigate specific concerns.
Boxes and Balloons
Armed with accurate HCP scores, pharmaceutical sales reps can make sure they’re prioritizing the right clients and putting in the work to boost order numbers with segmentation. The current legacy process effectively assigns each HCP a “box” shared by providers of a similar score. Reps might have five boxes, with the first representing low-volume, low-value prospects and the fifth representing top-priority contacts.
While this streamlines the sales process, it also creates a problem: Boxes can’t respond to dynamic changes. Providers in the third box are likely to stay there all year until reps have time to sit down and reevaluate existing prospects, even if changing circumstances create an opportunity for upward movement to box four or five.
There’s a better option: Balloons. Using predictive modeling, it’s possible for companies to incorporate emerging data, such as patient demographic changes or staffing concerns. By using analytical processes that update this data weekly or monthly, sales reps can escape the closed box problem and instead create balloons — clusters of similar-value HCP data that can be inflated, deflated or popped on demand.
Rewriting the Rules
Sales professionals depend on interpersonal relationships to drive pharma volumes and develop long-term client connections. Cluster-based predictive modeling can empower these connections by leveraging hard data to create provider and patient summaries that are descriptive, relatable and actionable for sales staff.
For one consulting project, the IBM analytics team converted data into words to create a “narrative” for what the numbers were predicting about these health care providers. By translating numeric connections into plain-language summaries, IBM was able to merge analytics and linguistics, giving sales reps the insights and information they needed to cultivate personal relationships without putting patient confidentiality on the line. The result? An 11 percent increase in sales across 40,000 psychiatrists.
The Winds Are Turning
There’s a change in the air underway: Mass-market email campaigns no longer deliver substantive value, and low-volume HCPs are frustrated by the lack of attention from pharma reps. Companies need a way to personalize prescription marketing — and the new wind has the potential to carry aloft these balloons of patient and trial data that pharma marketers can use to their advantage. By combining dynamic data with predictive modeling, it’s possible to replace static scoring systems with descriptive, evolving clusters and deliver math-based HCP summaries that empower human-to-human connections.