Does NPS Work With Physicians?
No, and the reason is more interesting than the usual complaint about survey fatigue.
The standard objection is that your physicians are busy and your response rates are poor. True, and beside the point. The real problem is that the strongest evidence for using an advocacy score in pharma almost certainly has the arrow pointing the wrong way. Here is the case for it, then why I stopped believing it.
What does the evidence for NPS in pharma actually say?
It is better than critics admit, so let me put it at its strongest.
Bain's 2018 pharma work found that roughly 40% of a doctor's drug recommendations relate to overall experience rather than the product. The same work put physicians who score a company highly at somewhere between 2.3 and 2.7 times more likely to prescribe its brands. McKinsey's launch research found fully satisfied prescribers more than twice as likely to prescribe. Those are large effects. If you run a brand, they are the numbers your agency shows you.
The original logic is sound too. Fred Reichheld built the case in the Harvard Business Review in 2003 on a real observation, which is that willingness to recommend tracked growth across the industries he studied. It was honest work. Pharma simply borrowed it without checking whether the conditions held.
So why am I not convinced?
Because a doctor who already prescribes your brand has every reason to rate you highly.
Sit with that. She has met your rep more often. Your medical team answered her question about a difficult patient. Her office staff knows how to work your hub. She has seen your drug work.
Each of those is a consequence of prescribing, not a cause of it.
So when you find that high scoring physicians prescribe more, you may have discovered that prescribing produces high scores. The correlation is real. The direction is assumed. Nobody in the industry has published the study that separates the two, and until somebody does, a brand plan built on that arrow is a brand plan built on a guess.
Bain's own numbers hint at the problem. Their published range has relative score differences explaining anywhere from 10% to 70% of the variation in subsequent revenue growth. That is not a measurement. That is a span.
In a separate paper the firm put it more carefully still, calling goodwill "a necessary but insufficient condition for generating revenue growth." Necessary but insufficient is a fair description of what you are buying. It is also not a thing you can manage to.
Why does a three person market break the score specifically?
Your physician is not the only person in the transaction, and the score only asks her.
She chooses. A payer decides whether to cover. A patient uses the therapy and pays part of the cost. The advocacy question assumes those three are one person, because in the markets Reichheld studied they were. Ask a bank customer whether she would recommend her bank and you have asked the chooser, the payer and the user in a single question.
Ask a physician the same thing and you have surveyed one third of the decision. The third least exposed to what goes wrong. She writes the script and the failure happens downstream, in an adjudication she never sees and a pharmacy counter she never visits. I have written separately about why consumer CX does not translate to pharma, and this is the clearest single example of it.
Can you even close the loop on a detractor?
Usually not, which removes the mechanism that made the instrument work everywhere else.
The discipline was never really about the number. It was about what happened after a low score: you called the customer, you found the problem, you fixed it, you told them you fixed it. That loop is the engine.
Your version is constrained. You may not be able to discuss the product with a patient. You often cannot see which handoff failed. Your rep may learn about a detractor score weeks later, in a format she cannot act on, about an issue that belongs to a function she does not control.
So you collect the signal and cannot complete the circuit. A diagnostic with no treatment attached is a report, not a program.
What is the score costing you?
Access, which is now your scarcest asset.
Veeva Pulse data reported by BioSpace found 45% of HCPs accessible to biopharma, down from 60% eighteen months earlier. Every survey you send consumes a little of what remains. Ask yourself what you did with the last wave of responses. If the answer is a slide, you spent access on a slide.
There is a credibility cost too. As I put it to CXM last year, doctors do not go around asking about loyalty. Asking a physician to rate her affection for a pharmaceutical company tells her something about how you see the relationship, and it is not flattering.
What should you ask her instead?
What it cost her to work with you.
How many calls to get one patient started. How many times her staff supplied the same information twice. How long she waited to learn a case had stalled. Whether anyone told her when it resolved. Effort is specific, attributable to a named team, and fixable inside a quarter, which is three things a score is not.
Pair that with progression data on the patients themselves and you have a measurement system rather than a mood ring. I set out the full replacement set in how to measure customer experience in pharma, and the short version is that every measure needs a numerator, a denominator and an owner.
The gap this leaves open is not small. Deloitte's 2025 research found that only 28% of HCPs believe pharma's engagement strategies meet their needs, against 82% of life sciences executives who say they are satisfied with those same strategies. Those executives have scores. The scores did not warn them.
Is there any version worth keeping?
One, and it is narrower than your vendor will tell you.
Run the same question across your own regions, teams or reps, at the same moment, and the confound I described above is roughly constant. Everyone is measuring physicians who already prescribe. What is left is relative, and relative differences between your own territories can point you at a practice worth copying or a problem worth visiting.
That is an internal diagnostic. It is not a measure of your commercial health, it does not belong in a board pack, and it cannot carry a brand plan. Keep it at that size and it earns its keep.
What it cannot do is tell you whether your science reached the patient. That question needs different instruments, which is why we built Customer Excellence around progression rather than sentiment.
Key takeaways
- The strongest pharma evidence for advocacy scores cannot rule out reverse causation, because prescribing produces the experiences that produce the score.
- Bain's own published range for score to growth runs from 10% to 70%, and the firm calls goodwill necessary but insufficient.
- Your chooser, payer and user are three people, so a physician survey reaches the party least exposed to the failure.
- Without the ability to close the loop on a detractor, the score is a diagnostic with no treatment attached.
- Keep it only as an internal comparison across your own teams, where the bias is constant and the differences still mean something.
Questions to ask your leadership team
- If your score rose four points next quarter, what would you expect in your brand numbers, by when, and who would be accountable for delivering it?
- How do you know your high scoring physicians are not simply your existing prescribers?
- When a physician gives you a low score, who calls her, how fast, and what can that person actually change?
- What did you do with the last wave of responses, and was it worth the access it cost?
- Does any measure in your CX pack tell you whether a patient started therapy?
About the author
Wayne Simmons is the founder of The Customer Excellence AGENCY and the author of The Customer Excellence Enterprise (Wiley, 2024). He is founding faculty of the MS in Customer Experience Management at Michigan State University's Broad College of Business. He led global customer excellence in Pfizer's first Chief Marketing Organization and in Bayer's Customer Powerhouse. Related reading: Why pharma CX programs fail, How to measure customer experience in pharma and What is value leakage in pharma?







