Why we ask eight questions before drawing a screen

When you open Quiny and describe your business, nothing is drawn until you have answered eight questions. That is deliberate, and this post is about how we chose them.

The test

A question stays if a different answer produces a different app. Not a different color, a different app. That sounds obvious, but most onboarding questions fail it. They exist because someone thought the information might be useful later, or because a form looked empty without them.

We wrote down every question we could think of and for each one asked: if the answer flips, which screen changes and how? Then we watched real people answer them and measured two things, how long they hesitated and how often they went back to change the answer later.

The eight that stayed

1. Who buys from you? Changes the home screen more than anything else. Repeat customers get reordering up front. New customers get the story of the shop.
2. What are they buying? Products, bookings, subscriptions or a mix. This decides whether there is a basket, a calendar or a plan picker.
3. How do they pay? Card, Apple Pay, on collection, or by invoice. Decides checkout, and whether accounts are required.
4. Do people need an account? Most small shops say no and get guest checkout. Services usually say yes because bookings need a name.
5. How does it get to them? Delivery, collection, digital or in person. Decides the last screen of checkout and what the confirmation says.
6. What should it look like? Not a color picker. Quiny reads your photos and links and shows three finished directions. You tap one.
7. What is the one thing you want people to do? Reorder, book, browse, join. This becomes the primary action on the first screen.
8. Anything people always ask you? Opening hours, allergens, returns. This becomes the small text that saves you support emails.

The four we cut

"What is your brand color?" People know their color and still gave us the wrong hex code. We read it from the logo instead and get it right more often.

"How many products do you have?" We can count. And the number changes the day after you answer.

"Do you want push notifications?" Everyone says yes and nobody had thought about what the notification would say. We ask about it later, when there is a real reason to send one.

"What is your business name?" It was the first question in the flow and it made the whole thing feel like a form. Now we take it from the sentence, or from the domain, and confirm it on the first screen.

Why not zero

We could build from the sentence alone and let people correct the result. We tried it. The corrections took longer than the questions, cost more credits, and left people with the feeling that the app was ours rather than theirs. Eight taps at the start is the cheapest place to be wrong.