How many questions a quiz should have (and in what order)

The question comes up eventually: if screening is good, why not ask ten things and screen even better? Because every question costs you people. What is useful is that there are fairly consistent numbers on exactly how much.

What the form data shows

Aggregated data from form platforms and funnel analytics (Zuko, Formsort and industry round-ups) points the same way:

• Forms with 3 to 6 fields see the highest completion rates; with just 3 fields, conversion as high as 25% has been measured.
• Each additional field cuts conversion by roughly 4%, and past 7 fields abandonment sits around 68%.
• The first 10 seconds decide it: about 35% of visitors settle within that window whether to continue or leave.
• Phone and email are among the highest-abandonment fields (roughly 6% each). Making the phone number optional rather than required moved abandonment from 39% to 4% in documented tests.

Do not transplant this uncritically: it is form data, and a multi-step quiz behaves differently because each screen asks little and gives a sense of progress. But the underlying pattern holds — the marginal cost of each question is real, and it compounds.

The number that works for local services

Four or five qualifying questions, plus the contact screen. It sounds arbitrary, but it satisfies both constraints at once: enough to rule out the unviable — ownership, timeline, budget, and one trade-specific technical question — while staying inside the range where completion is still high.

If you need more data, the answer is not a longer quiz: ask the rest on the call, when the person is already with you and answering costs them nothing extra.

Order matters as much as count

• Open with the easiest, most engaging question, not the one most useful to you. The first screen sells the rest of the quiz.
• Put your biggest disqualifier second or third: early enough not to waste anyone's time, not so early that it reads as an interrogation.
• One question per screen, with a visible progress indicator. Knowing how much is left is what sustains completion.
• Save contact details for last, after you have delivered something: a price range, an estimated saving, an indicative quote.
• Mark optional anything you do not genuinely need. That single change on the phone field is, per the tests above, among the highest-return tweaks available.

How to tell you overdid it

Look at completion per screen, not just the total. A drop concentrated on one step does not mean the quiz is long: it means that question is badly worded, arrives too early, or asks for something the visitor cannot answer yet. If the drop is spread evenly, then yes — it is length.

And remember what screening is for: the goal is not to maximise responses, it is to maximise useful ones. A quiz that converts less but leaves only viable contacts is cheaper than one that converts more and fills the calendar with wasted visits — with the numbers attached.