Quiz and Calculator Funnels That Change the Answer

A quiz costs the visitor between forty and ninety seconds. It earns that only if the answer changes something real: the recommendation, the price shown, the plan they enter, or the person who calls them. When we review an existing quiz the first thing we check is whether every path ends at the same offer, and in more than half of the ones we have looked at it does. The honest fix there is to delete the quiz, not optimise it, and we will say so in the first week rather than the last. What we build: Scored quizzes and calculators as a step inside a paid funnel — the branching, the scoring model, the result page, the CRM write-back and the per-question reporting. The scoring lives in a table your team can edit, a sheet that compiles to JSON at build time, rather than buried in a component. The model will be wrong in the first month, and the person who knows it is wrong is not us. Question design: The opening question is one a visitor can answer about themselves without thinking. Asking budget or job seniority first is the single most common reason a quiz loses its audience before it has earned anything. Depth is capped at seven questions and five is usually right. Each question has to change either the result or the routing; a question that exists only to feed a dashboard gets cut, and we will name the one your sales team asked for when it falls into that category. Calculators specifically: Inputs carry defaults drawn from your own data, so an untouched calculator still shows a number someone can defend. Currency and locale are set properly — SGD and MYR to two decimals, IDR and VND with no decimal places and correct digit grouping — because a rounding artefact in a savings figure loses a finance buyer faster than a weak headline. We will not build a calculator that produces a return figure your buyer cannot defend to their own CFO. Where the model needs an assumption you cannot source, that assumption goes on the page as a field the reader can change. The result and the gate: Result before email, or email before result, is a genuine decision and it turns on whether the result is worth withholding. Where volume supports it we test the two orders against qualified leads rather than raw fills. Our default is to show the result and gate the detail — the plan, the breakdown, the comparison — because a gated generic answer teaches the visitor that the whole exercise was a form with extra steps. Data and handover: Scored fields land in your CRM as typed properties: score, segment, branch taken, key inputs. Not appended to a notes field where nobody can filter them. Consent sits at the point of collection with the PDPA basis named. Per-question drop-off is in the dashboard from launch day, because a funnel without step reporting cannot be improved and we will not ship one. What is not included: The nurture sequence the results feed, product recommendation engines sitting on an ecommerce catalogue, and maintenance of the scoring model after handover unless you retain us for it. Who this is not for: Single-SKU businesses that want a personality quiz because a competitor has one. And teams under roughly 4,000 funnel entries a month, where branch-level numbers are too thin to tell a real difference from a slow fortnight. In both cases the money does more on the page the quiz was going to sit in front of, and we would rather build that.