A customer walks up to a pharmacy counter in Spain with a sore throat and a six-year-old in tow. She wants something for the child's cough and something for her own. The counter assistant has perhaps forty seconds, three competing brands on the shelf behind them, and no sales rep in sight. Whatever they recommend in those forty seconds is the sell-out. The program in this case ran from February to June 2026.
That is the reality this program was built for. A consumer health company in Spain wanted to reach the pharmacists and counter assistants in a target base of 2,830 pharmacies, in a crowded OTC market where its previous training had reached almost nobody. Over 139 days, 1,421 pharmacy staff joined the app, and 778 of them became active players. The knowledge numbers are more sobering: 27% of 23,378 gaps verified closed. Reach and verified correction are two different results, and the difference between them is the most useful lesson here.
In OTC, the counter is the field force you do not employ
Consumer health companies face a problem that prescription pharma mostly does not. The person who decides what the customer takes home is not on the payroll. It is a pharmacist or a counter assistant, working for an independent business, serving customers back to back, often recommending across dozens of categories in a single shift.
Traditional training struggles to reach them. A rep visits a pharmacy, talks to whoever is free, leaves materials that may or may not be read, and moves on. E-learning modules are built, launched and ignored, because nobody at the counter has forty minutes. Meanwhile the competing brands are doing exactly the same thing, and the customer at the counter sees none of it. They only see the recommendation.
The client's objective was concrete: turn counter knowledge into orders, brand by brand. The portfolio covered five ranges in quite different categories:
- a cough and throat range (syrups, sprays, lozenges) for adults and children
- a magnesium and energy range
- a heat therapy range for pain
- a sleep support range
- an eye care and intimate care range
Each of those needs different knowledge at the counter: ingredients, age thresholds, which format suits which customer, how a product works, when to recommend it and when not to. That is a lot to ask of people who are not your employees and owe you nothing.
We built it as an always-on program with brand cycles
We did not design a course. We designed a season.
The program ran always-on for 139 days, organised into five brand cycles. Each cycle focused on one range. It published short micro-contents and a podcast about that brand, then measured the counter's knowledge of it through one-minute challenges (five questions in 60 seconds). Across the program there were five games and 29 micro-contents and podcasts. Two competitions punctuated the season, and two reinforcement waves followed them, with Gap Fixing sending each person the fix for their own mistakes.
Reach was the first problem to solve, and notifications carried much of the load: 82,625 were sent across the program. Of 1,607 people invited, 1,421 joined the app and 778 became active players. Those are counts of people, not pharmacies: the target was a base of 2,830 pharmacies, and the records we used do not say how many of those pharmacies had at least one player. For a voluntary program aimed at staff in independent pharmacies, it is a meaningful base, but it is not the whole target.
The cycle structure was chosen for three reasons.
First, it matched how brands are managed. Each brand team could own its cycle, its content and its results, instead of competing for space in one large program.
Second, it kept the ask small. A counter assistant does not have to learn five ranges at once. They learn one, get tested on it, and move to the next.
Third, it paired content with competition on purpose. We expected the micro-contents and podcasts to explain, and the challenges and competitions to make people come back. That expectation turned out to be the central finding.
A fair objection is that pharmacy staff will not play a game about a manufacturer's products. Some won't, and it would be dishonest to pretend the 778 active players were the whole target base. But the people who did engage answered 181,000 questions. That is a great deal of voluntary attention on a category where attention is the scarcest resource there is.
Content explains. Competition gives people a reason to come back. In this season, far more gaps closed where people competed.
27% closed overall, but closure concentrated where the counter competed
Across the program, we identified 23,378 knowledge gaps at first attempt. A gap is one player wrong on one question at first try, closed only when that same player answers it correctly later. 6,197 closed, which is 27%. The other 17,181 were flagged for follow-up.
By brand range, closure was consistently modest. The five ranges below account for 23,158 of the gaps; a further 220 sat in other content outside them.
- The cough and throat range, the widest by far, produced 9,671 gaps (about 41% of all gaps found) and closed at 26%.
- The magnesium and energy range: 4,763 gaps, 22% closed.
- The heat therapy range: 3,492 gaps, 30% closed.
- The sleep support range: 3,144 gaps, 32% closed.
- The eye and intimate care range: 2,088 gaps, 29% closed.
The more important split was not by brand but by format. In the two competitions, closure reached 41% and 36%. Outside them, when content was published and the questions were simply available, closure was much lower. The overall 27% blends two very different patterns: people who came back to compete, and people who read a micro-content, answered once and moved on. This is an association within one program. The people who chose to compete may also have been the most engaged to begin with, so we can't say how much of the difference the format itself caused.
It is worth being honest about what 27% means. It is well below what we see in programs anchored to a single high-stakes moment, such as a launch or a certification, where people have a clear reason to get every answer right. An always-on channel program asks for something harder: sustained, voluntary attention from people who do not work for you, across five categories, for almost five months. Some drop-off is built in. What matters is where the closure happened, because that tells you what to change, and here the answer was unusually clear.
That fits what retrieval practice research would predict, and it is striking to see it this clearly in a real channel. Content on its own created few second attempts here; the competitions created many more. And without a second attempt, there is no way for a gap to close.
Three questions from the cough and magnesium ranges
The per-question data shows where the counter struggled most. We describe the questions rather than quote them.
Choosing between formats in the magnesium range did not move at all. Asked which product in the range suited a customer asking for a particular form of magnesium, 153 people were wrong at first attempt, and none of them was recorded as correcting it. Here the answer counts at the last attempt were identical to the first, which suggests the question was barely retested at all. So this zero may say more about the lack of a second attempt than about a persistent misunderstanding. Either way, it is the error that puts the wrong box on the counter: the right brand, the wrong product within it.
Ingredients of the children's cough syrup were widely confused. 243 people were wrong at first, most mixing up one ingredient with one from a similar herbal formula. Only 27 corrected it; 216 were still wrong at close, though here too the answer counts barely moved between first and last attempt, so many were probably never retested. When the question is which product is right for a child, confusion between formulas is more than a sales gap. It touches on whether the recommendation is appropriate at all.
How the adult syrup works on a dry cough was also stubborn. 139 people were wrong at first attempt, most of them choosing a mechanism that belongs to a different kind of cough product. Fourteen fixed it; 125 did not, and again we can't tell how many of the 125 saw the question a second time.
Those numbers are uncomfortable, and we are publishing them because they are useful. They tell the client something a satisfaction survey never would: in an always-on program, the most confusable product facts often did not come back to the same person at all, and when they did, few corrections followed. Those need a different treatment, starting with making sure they are retested.
Four lessons from a season at the pharmacy counter
- Reach and learning are two different results. The program reached pharmacy staff that previous training had not. The knowledge data says the verified correction was uneven. Report them separately, and keep any commercial figures in a third column with their own source, denominator and period; none of the knowledge data here is evidence of sales lift.
- Closure followed the competitions. Closure of 41% and 36% in the competitions against a much lower rate elsewhere is the clearest pattern in this program, even if it is an association rather than proof. If you want gaps to close, schedule more moments where people have a reason to come back.
- Within-range confusion is the costliest counter error. Most of the stubborn gaps were not about the brand at all. They were about which product in the range, which formula, which age. Design questions and content around those choices first.
- Some gaps belong to the rep, not the app. A rep visiting a pharmacy with a list of the exact product confusions that staff member has shown is far more useful than a rep with a generic brand message. The 17,181 open gaps are, in effect, a visit plan.
What we would do next season
We would put a competition in every brand cycle, not just two across the season. The pattern is clear enough to act on: far more gaps closed in the competitions than around published content on its own, and the cost of adding a short, time-boxed battle is small compared with producing more content.
We would build a dedicated round on within-range choice and age thresholds. Instead of scattering those questions across cycles, a focused challenge on "which product for which customer" would target the errors most likely to put the wrong product on the counter, and it would give the brand teams a single, clear measure of recommendation accuracy.
We would connect the open gaps to rep visits. The field team visiting these pharmacies could receive, for each pharmacy, the handful of product confusions its staff still showed. That turns a visit from a reminder into a two-minute fix, and gives the rep something specific to say.
And we would give counter staff a way to ask. An assistant that provides answers based on approved product information (ingredients, ages, formats, when to refer) would help at the moment of recommendation, and the questions it receives would show the brand teams what the counter is unsure about in real time.
The question for your next channel planning meeting
If you sell through pharmacies, you already know the counter decides a large share of your sell-out. The harder question is this: do you know what the people at that counter actually get wrong about your products, person by person and product by product?
Most consumer health teams measure reach, and some measure orders. Few measure whether their counter staff know the right recommendation. At Atrivity, this season gave a pharmacy team and its brand managers that third measure, including the parts that were not flattering. For the next season, that is the part we would build around.