Playable ads guide

Quiz Ads: Building an Interactive Quiz That Ends on a Recommendation

The quiz is one of the oldest interactive formats on the web, and it transfers to advertising better than almost anything else, for one reason: it is the natural shape of a purchase decision for any product that comes in more than one version. Which shade, which mattress, which plan, which course, which food for this particular dog. The shopper's real question is which one is for me, and a quiz is a machine for answering it.

It also solves a problem other playables cannot. Plenty of products have no slice that can be sampled in fifteen seconds, which rules out the usual demonstration mechanic. But many of those same products open with a questionnaire (goals, preferences, experience level) and the first three questions of onboarding are a perfectly honest sample of the product. The ad can be the start of the app.

When a quiz is the right mechanic

  • The range is wide enough that choosing is the hard part: shade, size, fit, flavour, tier.
  • The product is personalised, so the answer really differs between viewers.
  • The app's own onboarding is a questionnaire, and the ad can reproduce its opening.
  • The purchase normally starts with a comparison the viewer finds tedious, and the quiz does the comparing for them.

And one case where it is the wrong mechanic: when every path ends on the same product. A quiz with one outcome is a banner with extra steps, and viewers notice faster than you would expect. Somebody always retakes it with opposite answers to see what happens.

Three questions, not ten

Each question costs two or three seconds of attention, and attention in an ad does not renew. Three questions and a result lands around ten to twelve seconds, inside the range where a well-paced playable reaches its call to action. Five is the ceiling, and only for a rewarded placement where the viewer has already agreed to stay. Past that, completion collapses and the result screen, the only part that sells, is seen by a fraction of the people who started.

  • Answers are tiles, not text. Two to four options, each with an image or a colour where the subject allows, tappable anywhere on the tile.
  • No free text and no form fields. Typing inside an ad is slow and error-prone, and where the input is personal data it is a privacy problem the creative has no way to handle.
  • The first question is the easiest and the most self-descriptive. What is your skin type, how big is your dog, where are you flying from. It should take no thought at all, because its job is to start the momentum.
  • Order from identity to need to preference. Who you are, what you are trying to fix, what you like. It is the order a good salesperson asks in, and it produces results that feel reasoned.

Branching or scoring

There are two ways to turn answers into a result, and the choice decides how much you can maintain later.

A branching quiz routes each answer to a different next question and treats the final leaf as the result. It feels tailored, and it explodes: three questions with three answers each is twenty-seven paths, every one of which has to be written, tested in both orientations and translated. Branching works when the tree is shallow and converges quickly: one branching question that picks a track, then the same questions for everyone on it.

A scoring quiz keeps one linear set of questions and has each answer add weight to one or more outcomes; the heaviest outcome wins. Any number of questions share a fixed set of results, adding an outcome does not multiply the work, and the logic is easy to audit. For anything past two questions it is the better default. It needs one rule decided up front (what happens on a tie) or the result becomes whichever outcome happened to be checked first.

Every answer has to matter

Test the scoring on paper before anything is designed: write out which result each combination of answers produces, and look for answers that never change anything. An option that cannot move the result is a question the viewer answered for nothing, and the ones who retake the quiz to check will find out.

The result screen is the ad

Everything before the result is qualification; the result is the pitch. It has to name one specific product, show it large, and say why. Echoing the viewer's own answers back is the most persuasive copy available, because it is evidence the recommendation was made for them: because you picked oily and matte, because your dog is large and over seven.

Keep the call to action on the result screen and make it about the result rather than the brand: see your shade, get this plan, shop the match. And stay inside what the quiz can honestly claim. A three-question quiz is a recommendation, not a diagnosis. In health, finance and beauty, the difference between suggesting a product and appearing to assess someone is the difference between normal creative and a policy review.

Carry the answer into the click

A viewer who has just been told which product suits them should land on that product, not on a homepage where they have to find it again. Where the click-through goes to a website, the result and the answers can travel as ordinary query parameters, and the landing page opens on the recommendation. Where it goes to an app, the value has to ride on a measurement partner's tracking link rather than the store URL, which cannot carry it. The deep linking guide covers the mechanics.

Measuring a quiz ad

Quiz ads flatter their interaction rate, because answering a question is the easiest ask in advertising. The informative numbers sit further down.

  • Completion to the result screen. If people start and do not finish, the quiz is too long or one question is confusing; where the drop-off sits tells you which.
  • Result distribution. If eighty per cent of viewers land on one outcome, either the scoring is skewed or the audience really is: worth knowing which, because the second is market research you did not pay for.
  • Downstream conversion by result. Some recommendations convert far better than others once people arrive, which is a merchandising decision as much as a creative one.

Mistakes that repeat

  • Too many questions, usually because every stakeholder wanted theirs included.
  • Answer options written as sentences, which nobody reads at the speed an ad is consumed.
  • A result that is a generic brand page instead of a named product.
  • Asking for an email address or a postcode inside the ad. Lead capture belongs on the landing page, where there is a privacy notice.
  • Hiding the call to action until an animation on the result screen has finished.

Building one

In Playable Ads Maker the Quiz / Finder interactive module is a ready-made question flow that ends on a recommendation: re-skin the tiles, write the questions and the results, and it exports like any other playable. Quiz / Finder – Shade Match is a working example. A scoring quiz can also be built from ordinary objects: answer tiles as buttons, a Number variable per outcome, a Set Variable action on each tile adding its weight, and result cards revealed with a Show If Variable gate. All of the text stays live, so the same quiz can be translated rather than rebuilt.

Create playable end cards in minutes—no code required.

Open Playable Ads Maker