A blood bank tender is decided in a spreadsheet long before anyone signs a contract. Somewhere in a transfusion centre, a lab manager and a procurement lead line up four or five screening platforms side by side: hands-on maintenance minutes per day, calibration frequency, tests per pack, throughput, what happens when a donor sample is reactive. The sales specialist who gets one of those cells wrong does not lose an argument. They lose the column. The program in this case ran from March 2024 to July 2025.
That was the brief behind this program. A global diagnostics company wanted its blood screening sales and application specialists to know the competitor platforms as precisely as they knew their own, and to keep that knowledge current for more than a year. Over 16 months, six short capsules and a final battle, 1,569 player places across the capsules recorded 79,000 answers, and half of the 6,621 knowledge gaps we measured at first attempt were verified closed within the assessment. The more useful story is in the other half: which gaps repetition fixed on its own, and which ones stayed open and belong in a manager's deal review.
In blood screening, the comparison table is the sales conversation
Blood screening is not a market where you win on a clever opening line. The customers are transfusion services, blood banks and large central labs. They run high-volume serology and nucleic acid testing on donated blood, they answer to regulators and haemovigilance systems, and they buy through formal tenders that often lock in a platform for years. Switching is expensive and risky, so the evaluation is meticulous.
In that setting, the specialist has to explain the assay technology, handle the operational worries every lab has about switching, and hold their own in a number-by-number comparison with the incumbent, often in front of people who have run it for a decade.
The client's own training content reflected this. About a third of the tagged questions were direct competitive comparison: daily maintenance times, calibration frequency, assay pack sizes and the known weaknesses of four competing platforms. The rest covered assay technology and serology, screening algorithms, reading the instructions for use, a simulation of donor pool behaviour, and how to run the conversation when a customer pushes back.
Two things made this hard to train with a classic approach.
- Competitive facts go stale. A competitor launches a new assay, changes a pack configuration or updates a maintenance procedure, and last year's battle card quietly becomes wrong.
- The audience was small, specialised and spread out. A few hundred people per wave, sales and application roles mixed, working in English across many countries. Classroom sessions every time the picture moved were never realistic.
So the question we were asked was not "how do we teach this once?" It was "how do we keep a specialised team sharp on a moving target for well over a year, and prove it?"
We designed capsules, not a course
The answer we proposed was a series of short, repeated competitive windows rather than one large program.
Each capsule opened for five to seven days. Inside that window, players took one-minute challenges (five questions in 60 seconds) against real colleagues, drawn from a question bank that grew to 398 questions over the program. Six capsules ran between March 2024 and the spring of 2025, and a final battle in July 2025 closed the series. The largest single capsule had 277 players. The 1,569 figure counts each person once per capsule they played, so regular players appear several times; the records we used do not give a count of unique people or show how many played every capsule.
We chose this shape for three reasons.
First, a short window creates a moment. People join a battle that ends on Friday. They postpone a course that is available "any time".
Second, repetition across capsules gave errors a chance to show up again. Because the same competitive facts came back in later capsules, a person who got the maintenance time wrong in one window and played the next could be tested on it again, and we could see whether the error survived.
Third, the final battle was a late check. Sixteen months is a long time in a competitive market, and ending with a battle showed what the people who played it knew at that point, rather than what people remembered the week after a training day. We should be clear about its limits. The people in the final battle were not exactly the people in the first capsule, and each person's exposure in between varied, so it is not a matched retention study.
A fair objection is that one-minute challenges trivialise complex content. You cannot teach a donor pool simulation in five quick questions, and that is not what the questions were for. The deep content lived in the client's own training material. The challenges found, person by person, which facts had not landed, and forced a later attempt so we could tell whether the correction held.
A single wrong number in a tender comparison does not lose an argument. It loses the column.
Half the gaps closed. The other half told us where to coach.
Across the program, we identified 6,621 knowledge gaps at first attempt. A gap is one player wrong on one question at their first try. It only counts as closed when that same player answers the same question correctly later, verified in the platform. On that strict definition, 3,282 gaps closed, which is 50%. The remaining 3,339 were flagged for follow-up coaching.
The split by content block is where it gets interesting.
- The general modules across the program (the largest block, 3,599 gaps) closed at 53%.
- Competitor strengths and weaknesses produced 2,038 gaps and closed at 47%.
- The donor pool simulation produced 371 gaps and closed at 48%.
- Assay technology and serology produced 484 gaps and closed at 44%.
- Objection handling produced 115 gaps and closed at only 29%.
- Blood bank fundamentals barely produced gaps at all (14), which is what you would hope from an experienced team.
Read that list from the bottom up and a pattern appears. The basics were known. Facts closed at around half. Knowing the right response to a customer's specific worry closed least. Note that this is still knowledge tested in a multiple-choice question; whether someone can actually handle the objection in a meeting takes role play or observation to see.
Three questions from one capsule show the difference
We looked closely at the 277 players of the fourth capsule. Three questions tell most of the story. We describe them rather than quote them.
A pack size question closed well. Asked how many tests a new multipack configuration of one of the client's own assays contained, 44 people were wrong at first attempt. Thirty-two of them credited it with a larger pack than it actually has. That is not a trivial slip in a tender: quote the wrong number of tests per pack and the cost-per-result calculation collapses. Repetition fixed 26 of the 44. Eighteen were still wrong at the end of the capsule.
An objection handling question closed, but slowly. The scenario was a lab worried about the disruption of switching from an incumbent platform. The right answer was to offer an operational efficiency consultancy (a workflow analysis) together with a study of the switch effect. Twenty-seven people initially picked a different, plausible answer: run a serology evaluation programme. It sounds reasonable. It also leaves the lab's real worry, the workflow risk, unaddressed. Of the 36 people wrong at first attempt, 23 corrected it through repetition and 13 did not.
A competitor maintenance question stayed open. Asked for the daily maintenance time of a competing platform, 45 people picked a figure two minutes away from the correct one. Only 21 of the 46 who got it wrong fixed it through repetition, and 25 still had it wrong at the end. A near-miss like this is sticky, because the wrong answer is close enough to feel right. It is also the kind of detail a lab manager who runs that platform every morning will catch immediately.
That last example is the one we would take into a sales meeting. Repetition did fix almost half of it, so more practice is not useless. But for the 25 who kept the wrong figure, it looks like a belief that has settled, probably carried from an old battle card or a colleague's remark, and we would not bet on another week of questions. The better fix is a manager, a deal review and the current competitive document on the table.
In practice that intervention can be short. The manager gets the names of the people still holding the wrong figure, the correct figure and its source. At the next deal review for a live tender, the team rebuilds the comparison table for that account together, with the current competitive document open, and each person states the figure and where it comes from. A week later, the same question comes back to those people in the platform to check the correction held.
Five lessons we took from sixteen months of capsules
- Competitive knowledge needs a cadence, not an event. A kickoff battle card is accurate for a few months at best. Repeated capsules gave the client a way to refresh the competitive picture in the field without calling another meeting. When the content changed, the next capsule carried the change.
- Near-miss numbers are the most dangerous errors. The wrong answers that survived were rarely wild. They were close: a larger pack size, a maintenance time a couple of minutes off. Our view is that plausible distractors should be built on purpose around the numbers that matter most in a tender, because that is where false confidence lives.
- Objection handling closes slowest, so measure it separately. At 29%, objection handling closed at barely more than half the rate of the fact-based blocks. If you roll it into one overall readiness score, you hide the part of the job most likely to lose a deal. Report it on its own line.
- A late battle is a useful check, not proof of retention. It is easy to show good numbers the week after training. The final battle showed what the team playing it knew 16 months in, after staff turnover and territory changes. Because the cohort changed, it describes the team at that date rather than how much any individual retained. To measure retention properly, track the same people across capsules.
- Flagged gaps are a coaching list, not a failure rate. The 3,339 open gaps were not a verdict on the team. They were, name by name and question by question, the agenda for the next round of one-to-ones and ride-alongs.
What we would do next time
If we ran the series again, we would change three things.
We would tie each capsule more tightly to a real business moment: a regional tender season, a competitor launch, a new assay approval. In our experience, programs anchored to something the field already cares about tend to close more of their gaps than the ones that simply arrive on a schedule, although that is an association we have observed, not something we have proven causes the difference. A capsule that lands the week a competitor announces a new configuration answers a question the specialists are already asking.
We would route the stubborn gaps to managers automatically. The data showed exactly which people still believed the wrong maintenance time or the wrong objection response. Sending that list to the right sales manager before their next deal review, with the correct answer and the source document, would turn a report into a conversation.
And we would give the field a way to ask. Specialists in a tender often need a precise answer in the moment: a pack size, a calibration interval, the approved wording for a switch concern. An assistant that answers from approved competitive and product documents, and feeds the questions it receives back into the next capsule, would close the loop between what the field needs and what the program teaches. We now build that into programs like this one.
The question to take to your next sales meeting
Pick the five numbers your customers use to compare you with the incumbent. Then ask your team, individually and without notes, to state them. Not the story, the numbers.
If a third of them are close but wrong, you have a team that sounds confident and quotes figures a lab manager will correct. That group is who your managers should be talking to before the next tender.
At Atrivity we built this program with a specialised diagnostics team, and the pattern it revealed (facts close with repetition, the right response to an objection closes more slowly and often needs coaching) is one we see again and again. If you want to know which of your competitive facts have landed and which have not, that is where we would start.