Quiz Question Logic
Quiz question logic is the set of rules that determines how each answer affects scoring, which question appears next, and what outcome a respondent receives.
Key takeaways
- One answer option can carry points, tags and a path trigger at once.
- Rules evaluate in a defined order; ambiguous conditions produce silently wrong outcomes.
- Every answer combination needs a valid outcome, including an explicit catch-all fallback.
- Multi-select questions add several values at once and inflate totals unless capped.
- Rule count grows the test matrix faster than it grows funnel relevance.
In depth
Every answer option is a small record: it can hold a point value, one or more tags, and a trigger that changes what happens next. When a respondent submits a step, the engine reads that record, adds the points to a running total, writes the tags onto the session, and evaluates any conditions attached to the question before deciding what to render. Because the state accumulates as the person advances, later rules can reference earlier answers, which is what makes the questionnaire behave like a system rather than a form.
Two forces pull against each other. More rules make the experience more precise, but the test matrix grows with every condition, and interdependent rules become impossible to reason about once they reference each other across several steps. Weighting is the other lever: giving budget or timing questions more points than preference questions sharpens the signal, while flat weights keep the rule set legible. Most teams get further by writing fewer, clearer rules than by encoding every nuance they can imagine.
In practice the logic is worth writing down as a table before it is built: one row per answer option, with its points, its tags and any trigger. In a scorecard the same table doubles as documentation for whoever inherits the funnel. Teams usually start with scoring alone, confirm that respondents spread across the tiers, then add tags for routing, and only afterwards add triggers that alter the path. Every question also needs a defined behaviour when it is skipped.
Logic cannot repair a question that measures the wrong thing. If an option is ambiguous, respondents pick it for different reasons and the points it carries mean different things for different people. Answers are self-reported, so the rules encode what someone claims rather than what is true. Rule sets also age: weights written for last year's offer keep firing long after the qualification criteria changed, and nothing in the system flags the mismatch until conversions quietly drift.
Example in practice
How to measure it
Read the distribution of outcomes rather than individual sessions. Count how many respondents each tier and each tag receives over a full traffic cycle. If one outcome absorbs nearly everyone, the rules are not separating; if an outcome receives nobody at all, its condition is probably unreachable. Both are logic defects that a completion rate alone will never surface.
Then check whether the logic predicts anything. Compare the tier assigned by the rules against what those leads actually do next, such as booking a call or replying to the follow-up. A weight is doing work when higher-scoring answers correlate with better outcomes. Also count how often each individual option is chosen: options nobody picks add rules without adding information.
Common mistakes
The frequent failure is writing rules only for the paths someone imagined. Respondents whose answers match no condition land on a blank or default screen, and nobody notices because the analytics still record a completion. List every outcome before writing a single rule, attach an explicit catch-all, and check the outcome report for results that never occur, which usually means their condition can never be satisfied.
The second is silent double counting. A multi-select question adds the points of each chosen option, so a respondent ticking five boxes can outscore a genuinely better-qualified one who ticked two. The same happens when two questions both award the same tag. Cap the contribution of multi-select questions, and walk three extreme personas through the quiz: all lowest answers, all highest, and every box ticked.
Frequently asked questions
Do I need question logic for a simple quiz?
Even basic quizzes use light logic to assign scores or pick an outcome, so some form of it is almost always present. You can start with simple per-answer scoring and add conditional paths only when they improve relevance.
How do I keep quiz logic from getting too complex?
Begin with straightforward scoring, document every rule, and test each answer combination against the outcome it should produce. Add branching only where data shows it lifts completion or qualification accuracy.
What is the difference between question logic and branching logic?
Question logic is the whole rule set attached to answers, including points, tags and triggers. Branching is one part of it: the rules that decide which step comes next. A quiz can have rich question logic and no branching at all, if every answer scores and tags but everyone still sees the same questions in the same order.
How many points should an answer be worth?
Start with a simple scale, such as zero to three per option, and keep it identical across questions until real responses exist. Uniform values make the totals easy to reason about and easy to recalibrate. Add weight only to the questions that genuinely separate buyers from browsers, and change one weight at a time so you can see what the change did to the distribution.
Can one answer both add points and jump to another question?
Yes, and that combination is common. The option adds its value to the running total, writes any tags, and then its trigger sends the respondent to a different step. Keep the two effects documented separately, because a jump that skips scored questions lowers the maximum score reachable on that path and can distort the tier the person ends up in.
How should multi-select questions be scored?
Decide explicitly whether the values add up, whether only the highest selected option counts, or whether the total is capped. Plain addition rewards people who tick everything, which is rarely the behaviour you want to reward. If the question is about breadth, addition is correct; if it is about fit, take the highest value or cap the contribution at the level of a single-select question.
How do I test quiz logic before launching?
Walk the quiz as a respondent for each outcome you defined, choosing answers that should produce that outcome, and confirm the result and score match. Then test the edges: skip every optional question, pick only the lowest-value options, and pick every option in multi-selects. Anything that produces a blank result, a wrong tier or an unreachable step is a defect to fix before traffic arrives.
What happens if a respondent's answers match no rule?
That depends entirely on whether you defined a default. Without one, the respondent typically sees an empty or generic result, and the lead reaches your CRM without a tier or tag. Always set a fallback outcome that is genuinely useful rather than an error message, and monitor how often it fires; a rising fallback share means your conditions no longer cover real answer patterns.