Key takeaways
- A useful scenario holds three elements: who the prospect is, in what context, and what is on their mind that the sales rep does not know.
- Resistance is calibrated one notch above real life, never three: an impossible exercise discourages instead of training.
- The scorecard must focus on observable behaviors, not impressions.
- The debrief is half the exercise: without it, you have played a scene, not trained.
What separates a useful simulation from a decorative exercise
Most sales simulations fail for the same reason: they are too easy. The colleague playing the customer wants it to go well, gives in after two arguments, and everyone congratulates each other. A pleasant moment, and nothing has moved.
A useful simulation has one simple property: you can fail at it. The prospect may stay unconvinced, the conversation may end, the meeting may not be won. Without that possibility there is no learning — only a demonstration.
The rest of this article covers the four settings that make the difference. They hold whatever the format: role play between colleagues, a trainer-led workshop, or a voice simulator.
1. The scenario: three elements, not thirty
A two-page scenario will never be read. Three elements are enough, provided they are precise.
Who the prospect is. Not “a sales director”, but the sales director of a forty-person industrial SME, eight months into the job, who inherited a team they did not hire. The level of detail changes everything: it is what makes the objections credible.
The context of the approach. A cold call on a Tuesday at 9 a.m. and a follow-up after a white paper download have nothing in common. The same opening line works in one case and rings false in the other.
What the prospect has in mind and the sales rep does not know. It is the element most often forgotten, and the one that separates a simulation from a conversation. The prospect has just been let down by a similar provider; their budget is frozen until March; they are mainly looking for reassurance ahead of a committee. The sales rep will have to uncover it by asking the right questions — or miss it entirely.
2. Resistance: the setting that decides everything
It is the setting most often mis-calibrated in practice, in both directions.
Too easy, and the exercise is pointless. Too hard, and it discourages: a sales rep who fails three times in a row stops trying and just endures it. The useful rule is the one-notch rule: a prospect slightly harder than the average of those you actually meet. Enough to unsettle, not enough to crush.
Resistance is set with three independent dials, and it pays to tell them apart: availability — does the prospect have two minutes or twenty; openness — are they ready to listen or already closed; how hard the objections bite — “it's expensive” is not “we already tried your competitor and it went badly”.
Working one dial at a time beats turning everything up at once. One session on a rushed but open prospect, the next on an available but wary one: you learn what works in each case instead of hitting a wall.
3. The grid: look at behaviors, not impressions
“It was good” makes no one better. A useful scorecard lists things you can tick or not, and that two observers would agree on.
A few examples of genuinely observable criteria: did the sales rep announce the length of the call within the first thirty seconds? Did they ask at least three open questions before talking about the offer? Did they reformulate the need before proposing anything? Faced with the price objection, did they ask what “expensive” covered before answering? Did they close with a dated commitment?
Five to seven criteria are enough. Beyond that, the observer loses track and the grid turns into a form filled in afterwards, from memory, which means guesswork. And the same grid must serve from one session to the next: consistency is what makes progress visible.
4. The debrief: half the exercise
This is where most programs collapse, for lack of time. A simulation without a debrief is a scene played out: enjoyable, and without effect.
Three principles make it work. Start with the person who has just played: what they spot themselves sticks far better than what someone points out to them. Stick to two points: one that worked and should be repeated, one to correct. A list of eight remarks produces no change at all. Phrase the correction as an action, not a principle: “ask the budget question before offering the demo” is actionable, “listen more” is not.
The debrief gains enormously from the recording. Replaying the fifteen seconds where the conversation turned beats discussing it from memory — everyone remembers what suits them. It is also what makes it possible to debrief a real sales meeting with the same method.
The mistakes that come up most often
Having the customer played by someone who wants to help. Politeness disarms the exercise: nobody dares be genuinely unpleasant with a colleague. That is the structural limit of peer role play, and the reason a prospect simulated by an AI holds the role better: it isn't trying to be nice.
Running a simulation once a quarter. A behavior does not take hold in one session. What anchors it is spaced repetition, not the intensity of an annual workshop — the same mechanism described by the forgetting curve.
Working on everything at once. A simulation that covers the opening, discovery, pitching and closing works on nothing. One stage per session.
Not measuring. Without a consistent scorecard, you have no idea whether anyone is improving, and the exercise dies on its own after a few weeks — for lack of proof that it serves any purpose.