HCP Targeting Considerations: A Practical Guide to Making Smart, Defensible Choices

TL;DR: Good targeting is not about finding the most complex formula but about starting with the right intent, choosing metrics that reflect real clinical behavior, and avoiding distortions from noise, overfitting, or unicorn writers. When you balance potential and performance, build flexibility into the list, and roll out changes in a way the field can trust, targeting becomes a strategic tool rather than a disruptive exercise.

Every time I start a new HCP targeting project, I come back to the same question that almost always gets skipped at the beginning: Why are we doing this?

It sounds simple, almost obvious. But the answer shapes everything that follows: the methodology, the amount of change the organization can tolerate, and how confident the field will feel about the final list.

If the current system is underperforming, then we have the space to rethink it. If it is working reasonably well and we are simply refreshing the data, then the appetite for disruption is much lower. Understanding this early prevents rework and sets realistic expectations.

More importantly, it keeps the human element front and center. Targeting is not just data. It affects the people who have spent months building relationships with their physicians. When a new list removes too many familiar faces, the impact on morale can be real and undermine the rep's faith in your 'updated' target list.

With that context, here is how I walk through the thinking.

1. Simplicity Is Not Limitation. It Is Strategy.

Analytics tempts us to add more signals, more models, and more complexity. But targeting is one of the few places where simplicity is not just preferable, it is necessary.

A good model should be something the sales leaders can explain without hesitation. If the logic takes more than a minute to describe or requires a diagram, it is already too complex. Reps rely on trust and credibility. Anything that feels opaque will struggle to land.

Simplicity is not about being basic. It is about being direct, transparent, and aligned with how decisions actually get made.

2. Be Selective With the Metrics You Use

Most projects begin with a long list of potential variables. Many look interesting. Some show correlation. Very few deserve a place in the final formula.

Just because a metric is measurable does not mean it is meaningful.

The real test is whether the metric makes sense in the real world and whether it produces rankings the field can stand behind. If something requires too much justification, has a meager correlation of 0.55, or creates rankings that feel strange, it should not be included.

A smaller set of strong, intuitive metrics always outperforms a large set of questionable ones. When in doubt, it is much better to create an overall market indicator like total treated patients by the core competitors than betting the house on a single 'analog' drug that may or may not be a stand-in for your product.

3. Patient Metrics vs Script Metrics: Choose the Lens That Reflects Reality

Different therapeutic areas have very different prescribing patterns. The choice between patient-level and script-level metrics depends on how physicians actually interact with treatment.

Sometimes script counts tell the story. Sometimes, patient-level activity paints a more accurate picture of an HCP’s true value.

The goal is to avoid distorted rankings driven by a few extremely high-volume prescribers. These substantial writers will appear through sales reporting anyway. Targeting should instead illuminate the broader band of physicians who drive consistent market performance day after day.

In rare diseases, this dynamic shifts because every prescriber carries significant weight. But in most categories, targeting should emphasize depth rather than extremes.

4. Normalize Utilization Before Drawing Conclusions

Raw counts can be misleading. Dosing frequency, treatment duration, and formulation differences can all create the illusion of high or low utilization that simply reflects the nature of the drug, not the physician.

Without normalization, the data can trick you.

If you choose to use claims, dispensed claims offer more grounded insight because they capture who is actually able to get therapy into patients’ hands. In a targeting context, that signal matters more than theoretical intent.

5. A Two-Value Model Creates Clarity

There is a natural temptation to collapse everything into a single score. It looks neat. It feels efficient. But it hides crucial distinctions.

If two physicians have the same composite score, you cannot easily tell whether they are:

• strong writers today

• high potential for tomorrow

• or some mixture of both

A two-value matrix, usually Potential by Writer Value, keeps those distinctions clear. It protects against forcing different types of physicians into the same bucket, and it creates segments that are easier to act on.

Weighting schemes, like 70-20-10, matter far less than people think. The smaller weights mostly act as tie breakers. They do not fundamentally change the ranking a lot.

6. Always Add Padding to the Target List

Once the model produces a list, the real world takes over.

You will encounter no-see offices, geographic whitespaces, specialties the field is not prioritizing, and territories with uneven density. A perfect model on paper rarely survives its first encounter with practical constraints.

This is why I always build a target list that is slightly larger than the required count. The extra pool gives you room to account for territory imbalance and access issues. If a territory lacks sufficient high-value HCPs, the padding allows you to rebalance without compromising the model's philosophy.

If the imbalance is persistent, the issue is almost always the territory map rather than the targeting logic.

7. Claims or Sales Data: The Real Question Is Strategy

Teams often debate which data source is better. The truth is that neither is universally superior.

Sales data shows what is happening now and is highly reliable. Claims data provides broader insight and captures earlier signals, but can fluctuate with vendor changes and lacks transparency about its inputs.

The choice depends on what the organization values more: stability or early detection.

Being clear about this upfront prevents confusion later when results diverge from expectations. I wrote an entire blog post about this sometime ago.

8. Deciling Is Where Philosophy Becomes Practice

Deciling looks like a mechanical step, but it is actually where the values of the model take shape. If the underlying logic is sound, deciling reinforces it. If the model is overly complex or built on questionable metrics, deciling amplifies the distortion.

The field will always question unusual rankings. If you cannot clearly explain why a moderate prescriber ranks above a much larger writer, the system will lose credibility.

Good targeting is not just accurate. It has to make sense.

9. Rollout Matters as Much as the Model

Even the best-designed model fails if the rollout overwhelms the field. Large swings in target lists can feel abrupt and demotivating.

Two strategies tend to work well:

Gradual introduction.

Introduce a portion of new targets each quarter so the field has time to adjust.

Full refresh with controlled trimming.

Deploy the full model, then regularly remove a small percentage of low value targets as new data arrives. This keeps the system fresh without generating too much volatility.

Both approaches recognize that targeting is not simply analytical. It is also cultural and behavioral.

Closing Thought: Targeting Is a Human Exercise Informed by Data

After many cycles of targeting work, I have learned that the most successful models are not the ones with the most variables or the most sophisticated math.

They are the ones that:

• start with a clear purpose

• use metrics grounded in real behavior

• balance potential and performance honestly

• remain simple enough to be trusted

• and treat the field with respect during change

When those pieces come together, the output is more than a list of names. It becomes a strategic guide that helps the organization grow with confidence and clarity.