Resource Allocation / Optimization

Princeton Brand Econometrics (PBE) brings important new thinking and management tools to the mission of improving sales force productivity. Executives who take advantage of PBE’s tools can readily produce double-digit increases in rep productivity (incremental sales). Depending on the size of the company, these productivity increases will translate into tens of millions up to hundreds of millions in annual incremental sales.

There has always been a stumbling block when dealing with sales rep productivity—namely, that companies don’t know how many prescriptions were actually generated by sales reps. They know how many prescriptions each doctor wrote, they know how many times each doctor was seen by a company rep, and they know how many samples were dropped off. However, they don’t know how many prescriptions would have been written if the reps hadn’t shown up and how many were written because they did.

This no longer needs to be the case.

After six years of fruitful research, PBE discovered how to quantify a given doctor’s zero-details baseline prescribing and the additional scripts generated by each individual call. Representatives’ productivity for the entire territory can now be quantified with the lowest possible margin of error.

PBE’s work is not confined to the analysis of past performance. It represents a true breakthrough in helping representatives be more productive in the future. PBE can quantify the number of incremental scripts that will most likely result from one detail, two details, three details, etc. for every doctor in the territory and every brand of interest. Now, significantly greater sales can be achieved by showing representatives how to target call opportunities, not just doctors. The two approaches are different and reps will quickly appreciate the differences.

Currently, companies use prescription audit data to determine how frequently individuals should be seen. If only one company had this data, it would have a tremendous advantage over its competitors. After all, a bump in a heavy writer’s prescribing will mean more business than the same per cent increase from a light writer. However, the prescribing data is a commodity. Even the smallest companies have it. So, doctor-targeting based on historic prescribing gives no company a noticeable competitive advantage, although each probably believes its manipulation of the data is superior.

PBE’s approach accounts for both the volume of prescribing as well as each doctor’s unique responsiveness to the number of sales calls. PBE’s approach creates a significant competitive advantage for those who exploit it.

Having quantified the incremental value of each potential call, (and validated the results) PBE is then in a position to work with companies in a number of vital undertakings:

  1. Determining the optimal field force size based on the number of profitable call opportunities.
  2. Making it much easier for reps to plan their call activities. They simply have to check which doctors represent the most profitable call opportunities in an area and call on the ones who will see them. In fact, the planning process for representatives, to include coordination among mirrored territories, can be reduced to minutes per week through computerization.
  3. Sales territories can be configured based on the volume of profitable call opportunities area-by-area as opposed to the volume of past prescribing. The two approaches yield different configurations, with PBE’s configurations being superior in terms of addressing the true sales potential.
  4. After PBE has quantified how many scripts doctors would have written or will write in a territory assuming zero sales calls, it becomes possible to tie rep compensation and quotas much more closely to their true production. By tying compensation more closely to productivity and productivity more closely to effort, companies will see a significant increase in realized productivity and a modest increase in actual call averages.

PBE has proven through client-sponsored blind validation tests that its method of working at the individual sales call level produces forecasts of national TRx’s that fall within less than +/- 2% of actual. This can even be done for new products, with an error margin only about one point higher.

Having looked at the outcomes of current practices and the productivity that is possible, we conclude that productivity increases—above the zero call baseline—of 25-75% (compared to current allocation) are in the cards.


For new products and for re-examining the targeting for existing products, an arm of primary research is incorporated: 

Stage 1: Segmentation Analysis of Prescribing Universe

The data-collection technique employed by this methodology is designed to recruit very large physician samples in a cost-effective manner.  Samples are typically in the n=5,000 to n=15,000 range.  Nonetheless, it is not possible to recruit every physician to take part in the primary research.  Therefore, prior to conducting the primary research the prescribing universe must be segmented to enable PBE to reliably ascribe information collected on responder physicians to non-responder physicians.  The clustering technique most appropriate for this purpose is Latent Class Analysis, which better handles, among other things, categorical and ordinal variables relative to other clustering techniques.  The outputs of this analysis are groups of physicians that behave homogeneously.

There is an important reason for doing the segmentation analysis prior to fielding the primary research: the segmentation analysis often identifies at least one segment that is likely to have great potential (i.e., a group that is likely to show high promotional responsiveness in the primary research), but is relatively small in headcount.  If these smaller groups are not sampled initially at a disproportionately high level relative to their size, it is unlikely that they will be adequately represented in the sample, or even discovered.  By identifying every segment before fielding the research, PBE can ensure that each segment is sampled sufficiently.

Stage 2: Using Primary Research to Forecast Physician-Level Promotion-Responsiveness

The greatest strength of this resource-allocation methodology is that it leverages ScriptCASTTM—PBE’s new-product forecasting methodology—which has been demonstrated to be the most accurate new-product forecasting model in the industry.

It is critical that resource allocation be determined using a methodology that can forecast with pinpoint accuracy.  After all, determining the optimal allocation of resources is simply a matter of reverse-engineering a model that can accurately forecast sales under any promotional plan.  Certainly, any methodology that is unable to forecast with precision cannot be used to allocate resources since it can’t reliably tell you what the impact would be of executing one allocation versus another.

ScriptCAST utilizes an in-market validated methodology for forecasting how individuals will respond in the real world to a proposition.  The forecasts produced using ScriptCAST have, on average, fallen ±2.67% of actual Rx’s—greater accuracy than even test markets can achieve.


The massive samples generated by this data collection modality combined with the forecasting accuracy of ScriptCAST enables PBE to allocate promotional resources at the individual physician level.  This results in a resource allocation that produces significantly greater profits than using any other approach.