01
Step one: choose the few inputs that matter
Start by deciding which facts actually change the outcome. For a shirt, height, weight or chest measurement, and preferred fit are usually enough. Every extra question adds friction and lowers completion, so each one should earn its place. Worked example: a shopper reports a chest of 100 centimeters and wants a relaxed fit. The quiz asks nothing about age or shoe size because those do not affect a shirt. Mark which questions are required and which are optional, and offer plain units with a simple way to switch between metric and imperial.
02
Step two: map answers to a product size chart
Next, connect the answers to the garment's own measurements, not to a generic label. Each product should carry a chart listing its real chest width, length and sleeve. In the example, the shirt in medium measures 104 centimeters around the chest, and the large measures 110. A relaxed preference allows a few extra centimeters of room, so the quiz favors the medium. Keep the rule visible in your documentation: body measurement plus ease allowance compared against garment measurement. Update the chart whenever a supplier changes a pattern, because stale data produces confident but wrong results.
03
Step three: show the result with context and an exit
Finally, present the recommendation in a way the shopper can judge. Show the suggested size, the reason in one sentence, and the nearest alternative. In the example the result reads: medium suggested for a relaxed fit, large if you prefer extra room. Add a link to the full chart and the return terms so people can verify the advice themselves. After launch, compare recommended sizes with later exchanges to see which questions need adjustment. Treat the quiz as a guide that improves with feedback, not as a promise that every order will fit.