How to Build an NPI Target List From Brand Strategy

A target list is a strategy document disguised as a spreadsheet. Every row is a decision about where money goes. The teams that get it wrong usually do so because the list was pulled before anyone agreed what the campaign was for.

Step 1: Start from the brand objective, not a specialty code

Write the objective in one sentence. "Increase trial among high-potential prescribers who have not yet written" leads to a very different NPI target list than "defend share among existing writers" or "build awareness ahead of an indication expansion."

Then ask who, specifically, makes or influences that decision. For some brands it is a single specialty. For others it includes nurse practitioners, physician assistants, and pharmacists. For hospital products it may be a pharmacy and therapeutics committee, which changes the targeting unit entirely. The site's piece on choosing between individuals, practices, and health systems is worth reading at this point.

Step 2: Define the selection rules

Turn the objective into rules a data analyst can apply without guessing.

• Specialty. Decide whether you mean reported NPI taxonomy, claims-derived specialty, or a licensed specialty field. They disagree more than people expect. See how to handle specialty and subspecialty ambiguity.
• Prescribing volume or potential. Deciles from licensed prescribing data are standard, but deciles on the market measure tell you about category volume, not about your brand's opportunity. Consider whether a potential score, patient mix, or new-to-brand behavior is a better cut.
• Geography. Territory alignment with the field force, payer access by region, or launch markets.
• Account type. Independent practice, group, integrated health system, academic center.

Step 3: Apply exclusions

Exclusions are where lists most often go wrong, mostly by omission.

1. Deactivated NPIs and providers with no recent activity.
2. Duplicates (same provider under different records, or organizational NPIs mixed into an individual list).
3. Providers who opted out of promotional contact, where your CRM tracks that.
4. Providers in states or institutions with restrictions your legal team has flagged.
5. Out-of-scope specialties that slipped in through taxonomy codes.
6. Holdout cells reserved for measurement. Decide these now, before the list goes to partners, so nobody accidentally targets them.

The existing HCP target list audit has a fuller set of checks to run after the list is built.

Step 4: Do the list size math

A list that is too large for the budget spreads frequency thin. A list that is too small will not spend. A rough check, with hypothetical numbers:

If the budget is $200,000, the list is not the constraint. The constraint is how many of the 12,000 can be found at all, and you may want to add channels rather than impressions. If the budget is $5,000, cut the list to the top tiers rather than spreading thin. The reach assumptions here depend heavily on specialty, which the reach by specialty size piece covers.

Step 5: Tier the list

Few brands treat every NPI the same. A simple three-tier structure works for most:

• Tier 1: the providers who account for most of the opportunity. Highest frequency, every channel, field coordination.
• Tier 2: meaningful but secondary. Programmatic and email.
• Tier 3: broad coverage at lower frequency, or contextual only.

Tiers make reporting much more useful, because reach in Tier 1 matters more than total reach. For more on tiering without overfitting to volume, see HCP tiering.

Step 6: Hand off in a format partners can use

Keep the file simple. A CSV with these columns covers most needs:

• NPI (ten digits, stored as text)
• Tier or segment label
• Test or holdout flag, if partners need it to suppress
• List version and date

Do not send licensed prescribing values, decile numbers, or patient counts unless the partner needs them and your data license permits it. A label like "Tier 1" is enough. Send the same version to every partner, log who received it, and set a refresh date. Lists go stale; providers move, retire, and change practice.

What usually goes wrong

Lists built only from reported taxonomy. Deciles that describe the market instead of the brand opportunity. Holdouts forgotten until after launch. Different list versions at different partners, which makes reach impossible to compare. And no owner for the refresh, so a list built at launch is still running eighteen months later. The NPI targeting guide shows how each of these flows through to match rate and reach.

Practical takeaway

Before you pull a single NPI, write the selection rules and exclusions on one page and get the brand lead and analytics lead to sign off. Then build the list to that page, version it, and send the same file to every partner.

Editorial note. Analysis and frameworks are the author's own and do not represent Acxiom or any current or former employer, client, or named platform. Examples labeled hypothetical or illustrative are not results from real campaigns. Nothing here is legal, regulatory, or medical advice.