Most cycle meetings happen in a hotel conference room. This one happened on phones, across the New Year weeks, and covered well over a dozen brands in one go. The program ran in the winter of 2022 to 2023.
A pharma affiliate covering Moldova and Ukraine needed its field force ready to detail a wide primary care portfolio at the start of the year: pain and anti-inflammatories, cardiometabolic care, gastroenterology, thyroid therapy, neurology and urology. Getting everyone in a room for that much content was never realistic. Instead we ran four campaigns over 17 days, from late December 2022 into early January 2023. 219 participant places across audience rounds produced 203,000 answers, and 59% of the 16,181 knowledge gaps found at first attempt were verified closed. A slot is one person's place in one audience round, so the number of unique individuals was lower than 219: some people held places in more than one round. Readiness rose from 77% to 90%. The more interesting finding was which kinds of facts closed and which refused to, because that pattern applies to almost any portfolio cycle meeting.
A whole bag, three audiences and the holiday weeks
A primary care rep in this kind of affiliate doesn't carry one brand. They carry a bag. On a single day they might detail a topical anti-inflammatory to a GP, a digestive enzyme to a gastroenterologist, a beta blocker and an antidiabetic to a cardiologist, and a thyroid hormone to an endocrinologist. Each conversation needs its own dosing, its own pharmacokinetics, its own positioning against competitors and its own safety messages.
The cycle meeting is where that bag gets refreshed for the next period. It's also where companies usually discover, a few months later, that the refresh didn't stick for half the brands.
The constraints here were unusual:
- The timing sat over the New Year weeks, when face-to-face sessions were out of the question. That was a constraint of the business calendar, not a design choice we would recommend.
- The audience came in three separate rounds, with medical representatives and area managers both taking part.
- The volume was enormous: 934 questions across the portfolio, in Ukrainian.
- Management wanted more than a completion rate. They wanted to know, molecule by molecule, what the team could detail and what needed a second pass.
That last requirement shaped everything else.
What we designed: four campaigns, battles and gap fixing
We split the program into four campaigns, each a battle in its own window, so no single stretch asked people to master the entire portfolio at once. Players competed in one-minute challenges, five questions in 60 seconds, against colleagues in their own audience round. Notifications kept the rhythm going. The answer volume is not an argument for training over the holidays: some of that play will have happened in people's own time, and we would not plan a program that depends on it.
Wrong answers came back later in each campaign, which is how we measure closure. A gap is one person wrong on one question at first attempt. It counts as closed only when that same person answers it correctly on a later attempt. After the battles, gap fixing sent each person the correction for their own mistakes rather than a generic recap of everything.
Why this shape, rather than one long always-on game? Three reasons:
- A portfolio this wide overwhelms people if it arrives in one block. Campaigns let each therapy area have its moment.
- Competition concentrates effort. In our experience, people play hardest inside a defined window with a visible leaderboard. At peak, a single player slot reached 2,259 answers, far above the typical one.
- Separate campaigns produce separate baselines, which means the commercial team could see which therapy area was weak, not just that "the team" scored 77%.
A fair objection is that 934 questions is a lot, and that volume can turn a learning program into a memory contest. We took that seriously. The questions stuck to what a rep actually says in a call: doses, onset, time to peak concentration, indications, the positioning line against a competitor, the one safety point the GP will ask about. Nothing was there purely to catch people out.
Why area managers played too
One design choice is easy to overlook. Area managers took part alongside their medical representatives, in the same campaigns, answering the same questions.
There were two reasons. The first is credibility. A manager who has played the battle can coach from the same questions, and knows from experience which ones are hard. A manager who has only seen a dashboard is coaching from a spreadsheet. The second is honesty. Managers are not immune to the same misconceptions as their teams, and some of the most persistent beliefs in any field force started as a manager's shorthand in a team meeting years ago.
There's an obvious risk: nobody wants to lose to their own rep on a leaderboard. In practice, we have found that managers who play along, and occasionally lose, get more engagement from their teams, not less. It signals that the program matters and that getting things wrong in a game is safer than getting them wrong in front of a GP. It also meant that when the coaching list arrived after the campaigns, the managers holding it had already seen where the questions bit.
What happened: 203,000 answers and a clear pattern
Over 17 days the field produced 203,000 answers. At first attempt we identified 16,181 knowledge gaps. By the close, 9,510 were verified closed, 59%, and 6,671 were flagged for follow-up coaching. Readiness, the share of questions answered correctly, went from 77% on first attempt to 90% on last.
What did that ask of each person? The records give answers, not minutes, so here is an honest estimate. 203,000 answers over 219 slots is about 930 answers per slot on average. At roughly five answers a minute in a challenge, that is around three hours of play per slot across 17 days, or ten to fifteen minutes a day, not one minute. Treat that as an order of magnitude, not a measured time on task. And an average hides the spread: the most active slot answered well over twice the average, which means others answered far less. We don't have a clean participation distribution from these records, and next time we would capture median minutes per person and the share of people who played on fewer than half the days.
By therapy area, closure looked like this:
- Pain and anti-inflammatory range: 1,971 gaps, 66% closed
- Gastrointestinal range: 1,522 gaps, 60% closed
- Remaining modules across the program: 6,986 gaps, 59% closed
- Cardiometabolic portfolio: 1,916 gaps, 57% closed
- Neurology and urology range: 2,635 gaps, 56% closed
- Thyroid and iodine therapy: 1,151 gaps, 50% closed
The spread between 50% and 66% is narrower than you might expect for a portfolio this diverse. The weakest area, thyroid and iodine dosing, is also the one where dosing is most individualised and where a rep most needs to understand the logic rather than memorise a number.
The averages, though, hide the most useful part. When we looked at individual questions, a clear split appeared between facts that repetition fixes easily and beliefs that it barely touches.
A single dose figure is what repetition fixes. A belief about how fast a drug works is what it doesn't.
Three questions, three behaviours
We tracked three questions from the pain range closely. We won't reproduce them, but the topics tell the story.
A daily dose, closed. One question asked for the daily dose of an oral NSAID in a migraine attack. Fifty-nine people got it wrong at first attempt, most of them picking a lower dose. Repetition fixed 54 of the 59, 92%. The most common wrong answer fell from 52 people to four. A single, concrete number is exactly what short, spaced repetition is good at.
Onset of effect, half closed. Another asked how long after taking an analgesic a clinically meaningful effect appears. Sixty-one people were wrong at first attempt, and repetition fixed 32. But 29 were still wrong at the end, and almost all of them were wrong in the same direction: they promised faster relief than the product delivers. That's not a harmless slip. A GP who passes on that promise sets up a patient to be disappointed, and the next prescription goes elsewhere.
Time to peak concentration, still open. The third asked how long it takes for another NSAID to reach maximum plasma concentration. Fifty-six people got it wrong at first attempt. Repetition fixed only three. Fifty-three were still wrong at close, and the most popular wrong answer, a much faster time than the correct one, actually gained followers between first and last attempt.
Look at those three together. The dose was a fact. The other two were a belief, a shared sense in the team that these products work faster than they do. Beliefs like that often come from somewhere: an old detailing message, a competitor's claim, a colleague's shorthand. They spread across a team and they don't dissolve by being asked again. They need someone to explain the pharmacology and, ideally, why the faster claim is tempting and wrong.
Four lessons for portfolio cycle meetings
1. A cycle meeting doesn't need a meeting room. For a broad portfolio, four short campaigns on phones produced more measurement, and more practice, than a day of slides would have. The meeting room is still valuable for strategy and for the conversations that need a person. It is a poor place to refresh fifteen or more dosing tables.
2. Measure by molecule, not by bag. A readiness score of 77% for the whole portfolio tells you almost nothing. Broken down by therapy area and question, it tells a first-line manager which two brands to cover on the next ride-along.
3. Separate facts from beliefs early. In these three questions, verified closure varied substantially: 54 of 59 gaps closed on the dose question, 32 of 61 on onset, and 3 of 56 on time to peak concentration. If a question has barely closed by the end of a campaign, stop repeating it and hand it to a person.
4. Watch the direction of the error. The stubborn errors all overstated speed. That is the kind of mistake that drifts towards overclaiming, and it is worth flagging to Medical and to whoever owns the messaging, not just to the training team.
What we would do next
With two years' distance, a few things stand out.
We would run a short, targeted second wave four to six weeks after the cycle meeting, focused only on the questions that stayed open. Murre and Dros's 2015 replication of Ebbinghaus, a single-participant experiment with nonsense syllables, shows that unrehearsed material fades fast and that spacing helps. It can't tell us how much dosing knowledge survives to March, but one later right answer does not rule out a lucky guess, and a second check is the only way to know.
We would give area managers a coaching list per rep on the belief-type errors, such as onset and pharmacokinetics, with a two-minute explanation they can use on a ride-along. The 6,671 open gaps were never meant to be solved by the platform alone.
And we would bring the messaging owners in earlier. When 53 people share the same wrong belief about how fast a product reaches peak levels, the question isn't only "who needs coaching?" It's also "where did the team learn this?" Sometimes the answer is an old visual aid.
If your next cycle meeting covers more brands than anyone can refresh in a day, a good question for the planning session is this: which of our products does the team believe works faster, or better, than the label says? A few weeks of short battles will usually tell you, and the answer is rarely the one you expect.