How do you design effective decision points?

Every scenario stands or falls on its decision points. The dialogue can be sharp and the setting convincing, but if the decisions ask for recall instead of judgment, learners recognize the quiz underneath within seconds. This guide is about the craft of the decisions themselves: where they come from, what makes the choices plausible, and why the strongest decisions have more than one optimal option. It is the design-side companion to decision-point mentoring, the coaching that arrives at the moment of choice. Here the subject is the moment itself.

Where do decision points come from?

Not from the content outline. The most common design mistake is deriving decisions from the material: the course covered five steps, so the scenario asks about the five steps. Decisions derived from content test whether people were paying attention. Decisions derived from performance test whether people can act, which is the thing anyone actually cares about.

The raw material is real situations where practitioners struggle. You find them by asking the people who watch performance up close, such as subject matter experts, coaches, and frontline managers: Where do people go wrong? What do typical performers do in that moment, and what do your best people do differently? This is an old and reliable technique. Flanagan's critical incident method (1954) formalized the insight that experts give far better information about specific incidents than about general principles. "Tell me about the last time someone mishandled a price objection" produces usable detail; "what should people know about pricing conversations" produces bullet points.

What those interviews describe is a gap: the distance between what people typically do in a situation and what top performers do in the same situation. That gap is the decision point, essentially pre-assembled. The situation becomes the scene. The typical missteps become the tempting wrong choices. The expert behaviors become the optimal choices. And the expert's explanation of why each behavior helps or hurts becomes the coaching. Very little needs to be invented; the design work is mostly careful listening and faithful transcription.

One discipline keeps this clean: one decision per gap. Each distinct struggle gets its own decision point. Packing two gaps into one decision muddles the choices, muddles the coaching, and muddles the data about which skill learners actually lack.

What does a strong decision point look like?

Strip away the surface details and a strong decision has three components: a situation with genuine stakes, a moment where judgment is required, and a set of choices the audience would recognize as real. The first two determine whether the decision is worth building at all; the third, covered in the next section, determines whether it works once built.

Stakes first. Something the learner's character cares about must be at risk: a deal, a patient's trust, a team member's confidence, a deadline. Stakes do not require drama. A customer going quiet on a call has stakes. What stakes provide is a reason the choice matters, which is what separates a decision from an opinion poll.

Judgment second, and this is the harder test. A judgment moment is one where reasonable, informed people could plausibly choose differently: where there are competing goods, tradeoffs, or a tempting shortcut. If everyone who read the manual would get it right, the moment is a knowledge check, which is a legitimate tool but a different one. Clark's evidence-based guidance on scenario-based e-learning points the same direction: build scenarios around the tasks where errors are common and consequential, not around the content that was easiest to write about (Clark, 2013).

A quick test for any drafted decision: could you defend more than one of the options out loud without embarrassment? If not, no judgment is being exercised, and the decision will play as a quiz.

What makes the choices work?

Choices feel plausible for exactly one reason: they are drawn from what people actually do. The list of observed missteps from your expert interviews is a distractor goldmine, because none of it is invented. "Immediately offer a discount" is not a hypothetical wrong answer; it is what a meaningful fraction of salespeople do under pressure, which is precisely why it tempts learners on screen the way it tempts them on the job.

The heart of the design is the tempting-but-suboptimal option. Every strong decision has at least one choice that feels right, looks professional, would draw no criticism in many organizations, and is quietly the wrong move. This option is where the learning happens. A learner who selects it, watches it start to falter, and then hears why it falls short experiences something no bullet list delivers. The research on learning from errors supports investing here: studying incorrect approaches alongside correct ones can deepen understanding, particularly for learners with enough prior knowledge to explain why the error is an error (Große & Renkl, 2007).

The corollary is ruthless: if a distractor would not tempt anyone with a week of experience, cut it or replace it. An option nobody selects is dead weight in the interface and a blank row in the analytics.

Can a decision have more than one right answer?

Yes, and the best ones usually do. Real decisions frequently admit several defensible approaches. Recognition-primed decision research suggests that experienced decision makers evaluate options by whether they will work, not by comparing them against one correct answer (Klein, 1998). A design that forces a single right answer onto a situation experts would handle several ways teaches a false picture of the work, and experienced learners notice immediately. Nothing loses a veteran audience faster than marking their field-tested approach wrong.

Designing for multiple optimal choices changes three things.

Respect. A decision with several strong approaches signals that the scenario was written by someone who understands the job. Learners extend trust to the rest of the experience accordingly.

Discovery. When more than one option is optimal, the task shifts from "find the answer" to "find all the strong approaches." That is a better task. It asks learners to map the option space rather than guess the author's preference, which is much closer to how judgment actually develops.

Analytics. When several options are valid, decision data stops being a percent-correct score and starts describing behavior: which approaches learners gravitate toward first, which strong approaches they never consider, where coaching was needed. That is the kind of evidence that supports real conclusions about capability.

There is also a practical authoring benefit. When two subject matter experts disagree about the best response, the disagreement usually dissolves once multiple optimal choices are allowed: both are right, and the decision should say so.

How do you evaluate each option's outcome?

Three categories cover nearly every case: optimal, suboptimal, inappropriate.

• Optimal choices are what top performers do. There is often more than one.
• Suboptimal choices are what reasonable people do that falls short: plausible, tempting, and carrying a real cost. They deflect ownership, delay the hard conversation, or solve the wrong problem competently.
• Inappropriate choices are actively damaging or unprofessional, but still drawn from observed behavior. People really do get defensive, blame other departments, and argue with customers. The tier exists because those behaviors exist, not to supply cartoon villains.

The categories only become methodology when every choice carries a written rationale: why does this option lead to the outcome it does? Getting that rationale means pushing experts past "that's just wrong" to the mechanism, to the why. What does this choice cause? What does the customer, patient, or colleague on the other side experience? What happens three sentences later?

The rationale does double duty. During design it is a quality check: if you cannot articulate why an option is suboptimal, it may not be, and it may belong in the optimal tier or out of the decision entirely. During the experience it becomes the mentoring, and feedback earns its effect by being specific and tied to the action the learner actually took (Shute, 2008). A category without a rationale hands the learner a verdict; the rationale is what turns it into a lesson.

What are the common failure modes?

Three patterns account for most weak decision points.

Quiz questions disguised as scenarios. "The customer seems upset. Which of the following is a best practice for de-escalation?" Narrating a setting does not make a recall item a decision. The tell is grammatical: the question asks what is true about the domain instead of asking the learner to act. A decision reads "What do you do?", never "What is correct?"

Options nobody would choose. Insult the customer. Ignore the alarm. These pad the option count, lower the challenge, and quietly tell learners the experience was not written for professionals. Every option should be something a real person has actually done in that situation, which is another argument for sourcing options from interviews rather than imagination.

Decisions about trivia. Which form number, what the policy paragraph says, which year the regulation passed. If the answer can be looked up in ten seconds on the job, the moment involves no judgment and deserves no decision point. Reserve decisions for moments where the difficulty is choosing, not remembering, and let reference material stay reference material.

What does this look like in a worked example?

Here is the full arc, compressed, for a generic customer-escalation decision.

The interviews surface a gap. Situation: a long-standing customer calls after a second missed delivery date and demands a manager. Typical behaviors: transfer the call immediately, lead with policy, offer a discount right away. Expert behaviors: own the failure specifically, ask about the impact before proposing a fix, and stay attached to the resolution even when escalating.

The situation becomes the scene. The learner guides a service rep. The customer finishes recounting the second miss, pauses, and says: "I want to speak to your manager." The decision asks: what do you do?

Notice the shape. Three optimal choices, because experts genuinely handle this moment several ways. Two suboptimal choices that are things competent people do every day, including one (the immediate transfer) that literally grants the customer's request. One inappropriate choice that is regrettably common rather than cartoonish. And every rationale traces back to a single interview question: why?

Where does the methodology lead?

Everything above is tool-agnostic. You can build a decision this way for a live role-play, a traditional branching simulation, or a paper case study, and it will be better for it. But the methodology does favor certain structures: it wants several optimal choices learners are expected to discover, categorized suboptimal options that can play out safely, and coaching delivered at the decision itself. That combination is the native shape of AliveSim's Guided Scenarios. A suboptimal choice is allowed to play out just far enough for the learner to see why it falters, a mentor character explains the reasoning in place, and the learner returns to the same decision to find the stronger approaches, including ones a different expert might have picked first. AliveSim

The craft, though, starts long before any tool opens: interview for the real situations, derive each decision from the gap between typical and expert behavior, build the choices from what people actually do, and write down the why. Do that, and the decision points will carry the scenario.

References

• Clark, R. C. (2013). Scenario-Based e-Learning: Evidence-Based Guidelines for Online Workforce Learning. Pfeiffer.
• Flanagan, J. C. (1954). The critical incident technique. Psychological Bulletin, 51(4), 327–358.
• Große, C. S., & Renkl, A. (2007). Finding and fixing errors in worked examples: Can this foster learning outcomes? Learning and Instruction, 17(6), 612–634.
• Klein, G. (1998). Sources of Power: How People Make Decisions. MIT Press.
• Shute, V. J. (2008). Focus on formative feedback. Review of Educational Research, 78(1), 153–189.

Published August 5, 2025 · Updated July 14, 2026 · 11 min read