Skincare
Skincare Routine Finder
Help shoppers find the right cleanser, serum, moisturizer and SPF by skin type, goal and budget.
Skip to main content
Question Shape 01
Question Shape 02
Question Shape 03
Question Shape 04
Skip 01
Skip 02
Skip 03
Skip 04
Product Recommendation Quiz Maker and Product Finder
A product recommendation quiz asks a few questions and points each shopper at one product. Build yours free in the editor below, no account needed.
Quiz
Themes
Settings
Results
Leads
Share
Make a Quiz
How it works
How to build your product recommendation quiz
The editor at the top of this page is already set to Recommend a Product/Service, so the quiz type is chosen for you. Three steps and it is live.
-
Write 5 to 7 questions
On the Questions tab, ask only what changes the recommendation: goal, type, fit, sensitivity, size, budget, use case. If an answer cannot change which product wins, cut the question. Five tight questions beat eight tired ones.
-
Add your products
Open the Products tab and use Add Product once per product, bundle, routine or plan. Give each a real name, an image, a one-line reason you recommend it, and the link the shopper should land on. That entry is the result page.
-
Match answers to products
On each product, click the green Match Answers to Product button. A pop-up lists every question. Click the answers that point to that product, then OK. Repeat for each product and the scoring is done. Match percentages stay hidden unless you switch them on.
Prefer a first draft written for you? The AI quiz generator writes questions and outcomes from a prompt, and you finish in the same editor. For the general flow behind any quiz format, see how to make a quiz.
Product recommendation quiz examples
Live examples across beauty, apparel, food and home. Take any one end to end and see the result page a shopper gets, then build the same shape in the editor at the top of this page.
Beauty · Skincare · Haircare
For beauty and skincare stores
Haircare
Haircare Product Finder
Match the shopper to a hair routine by hair type, main concern and styling habits.
Cosmetics
Foundation Shade Match
Routes shoppers to the right foundation shade and finish by tone, undertone and coverage.
Fragrance
Fragrance Finder
Recommends a scent family by mood, wear occasion and note preference.
Apparel · Footwear · Jewelry
For apparel, footwear and accessories
Fashion
Apparel Style & Fit Finder
Maps the shopper to a style capsule by fit preference, occasion and priorities.
Footwear
Running Shoe Finder
Picks the right shoe by use case, surface, cushioning preference and foot width.
Jewelry
Jewelry Gift Finder
Recommends a jewelry gift by recipient, occasion, style and budget.
Accessories
Watch & Accessory Finder
Picks a watch or accessory by style, lifestyle and budget.
Coffee · Wine · Tea · Wellness
For food and beverage brands
Coffee
Coffee Bean Finder
Recommends a bean and subscription cadence by brew method, roast and flavor notes.
Wine
Wine Finder
Recommends a varietal by taste profile, food pairing and occasion.
Tea
Tea Finder
Picks a tea blend by mood, caffeine preference and flavor profile.
Wellness
Supplement Routine Finder
Builds a supplement bundle by wellness goal, format and dietary preference. Not medical advice.
Home · Pet · Subscription
For home, pet and subscription brands
Sleep
Mattress Finder
Routes shoppers to the right mattress by sleep position, support preference and budget.
Pet
Pet Food Finder
Recommends a pet food formula or subscription by life stage, size and texture. Not veterinary advice.
Home
Plant Finder
Recommends an indoor plant by light, watering and care level.
Subscription
Subscription Box Finder
Matches shoppers to a subscription box by interest, budget and delivery frequency.
What's the Right Dog Breed for You?
A product recommender quiz for pet retailers, breeders and adoption platforms. Eight questions on lifestyle, living space, energy level and grooming tolerance route the shopper to one of four breed shapes, then anchor the lead to the matched profile.
8 questions
<2 min to complete
4 breed matches
Why merchants build one
A product quiz does four jobs at once
Every question a shopper answers is doing double duty. It narrows the catalogue for them, and it tells you something about them you had no other way to learn.
-
It turns a big catalogue into one clear answer
Shoppers stall when every product looks plausible, and a filter wall asks them to already know the answer. Five to seven questions about goal, type, fit, sensitivity and budget narrow hundreds of options to the one that fits. On the Products tab you map each answer to the product it points at, and the result page can lead with the best match.
-
It earns the email instead of interrupting for it
A popup asks for an address before you have given anyone a reason. A quiz asks after the shopper has answered a handful of questions about themselves and wants to see the result. Turn on Lead Capture in the Leads tab, pick the fields you need, and leave the skip option on so the shopper is never trapped.
-
Every subscriber arrives already described
You are not collecting an anonymous address. You are collecting skin type, budget band, fit, sensitivity, use case or gift recipient, declared by the shopper rather than guessed from clicks. Export the responses and the next email already knows who it is talking to and which product it should be about.
-
It tells you what your customers keep asking for
Answers accumulate into something your analytics cannot see: which concern comes up most, which budget band dominates, which products shoppers match into and which never get chosen. That is merchandising, buying and copy research, collected as a side effect of helping someone shop.
All of it runs on the storefront you already have. The Share tab gives you one script tag that works on Shopify, WooCommerce, WordPress, Webflow, Wix, Squarespace or a plain HTML page. No app install and no theme rebuild.
Find Your Perfect Phone
A product finder quiz for phone retailers, carriers and refurb shops. Eight questions on camera priority, gaming, battery and budget sort the shopper into one of four phone categories, then route them straight to the matching collection.
8 questions
<2 min to complete
4 phone matches
Embed + sync
Embed on Shopify or WooCommerce. Sync answers to your stack.
Drop the product recommendation quiz on a product page, collection page, landing page or popup. Route shoppers to product pages, cart links or recommended bundles the moment they see their match.
Push answers and emails into Klaviyo, Mailchimp or HubSpot through Zapier-style workflows. No code, no theme edits.
Find Your Engagement Ring Style
A product recommender quiz for jewelers and DTC ring brands. Eight questions on setting, metal, stone shape and lifestyle sort the shopper into one of four ring styles, then drop them on the matching collection with their style profile captured.
8 questions
<2 min to complete
4 ring styles
What Questions to Ask in a Product Recommendation Quiz
Five to seven questions, and every one should change which product they end up with. Ask what someone actually did, not what they say matters to them.
Goal
What are they trying to solve?
Start with what they want fixed, one thing at a time. Give this answer the most weight when you match answers to products.
- ›"What do you want to work on first?"Dryness / Shine / Uneven tone
- ›"What wakes you up?"Too hot / Partner moving / Waking up sore
Type, style or fit
Who is the shopper, in catalog terms?
Ask about something they can check, not something they have to label. People guess their own hair type wrong. Nobody gets it wrong watching it dry.
- ›"Let it dry on its own. What does it do?"Beats "what is your hair type?"
- ›"On your most worn shoes, where has the tread gone?"Outer heel / Evenly / Inner edge
Sensitivity, budget or use case
What rules are non-negotiable?
The deal-breakers. One tap on an allergy can rule out half your catalog, and without them every result looks like an upsell.
- ›"Anything you want us to leave out?"Fragrance / Alcohol / Nothing I know of
- ›"What did you pay last time?"Works better than "what is your budget?"
The situation
What is going on around this purchase?
For the shopper, not the scoring. Ask about the situation, not what they say they care about.
- ›"Is this a replacement, or your first one?"Reassure the first, upgrade the second
- ›"What did you use last, and did you finish it?"Finished means repeat. Gave up means switch.
Put a short line under each question saying why you need the answer. In a test on a shopping site that lifted the share of questions people answered from 84 to 91 percent.
Anti-pattern
Questions to skip
If the answer cannot change which product the shopper sees, it does not belong on the quiz.
"How did you hear about us?"
Ask it at checkout. On question one it just tells the shopper this is really a form.
"How important is price to you?"
Most guides recommend it, and it backfires. When researchers asked people why they preferred something, they liked their choice less three weeks later.
Open-text questions on mobile
Typing kills completion, and seven in ten sessions are on a phone. Baymard Institute found field count matters more than step count, so splitting it up will not save you. Give 4 to 6 options to tap.
Trick questions and personality games
"Which animal are you?" belongs on a personality quiz. Here it adds length and hints the result is just for fun.
What's Your Hair Type?
A diagnostic product finder quiz for haircare brands. Eight questions on texture, density, porosity and styling habits sort the shopper into one of four hair types, then route them to a starter routine matched to that profile.
8 questions
<2 min to complete
4 hair types
The Numbers
The Product Recommendation Quiz Benchmark
We analysed 7,669 published product recommendation quizzes and 1.8 million completed responses, then rebuilt 700,257 sessions question by question to find the exact point where a shopper walks away. Method and exclusions sit at the bottom of this section.
7,669
Recommendation quizzes analysed
1.8M completed responses
700,257
Sessions rebuilt question by question
Not a stored completion counter
4x
What question one costs you
5.4 points lost, against 1.3 for an average question
20 pts
Completion lost to a welcome screen
The largest single build effect we measured
How long should a product recommendation quiz be?
Completion and mean time to finish by question count, averaged across quizzes so one viral quiz cannot carry a band. 2,800 quizzes with 100 or more sessions each.
| Quiz length | Quizzes | Mean time to finish | Completion |
|---|---|---|---|
| 1 to 4 questions | 180 | 1m 01s | 88.3% |
| 5 to 7 questions Best fit | 600 | 1m 49s | 81.6% |
| 8 to 10 questions | 640 | 2m 20s | 76.9% |
| 11 to 14 questions | 280 | 2m 21s | 81.4% |
| 15 to 19 questions | 220 | 4m 07s | 76.6% |
| 20 or more questions | 880 | 7m 06s | 69.5% |
Length costs minutes for certain and completion on average. The time column climbs without a single reversal, and the longest band completes 18.8 points below the shortest. Roughly a point goes with every question you add.
The 11 to 14 band runs against the trend at 81.4 percent; we report it as measured. The 1 to 4 band completes highest, but four questions rarely carry enough signal to route someone well, which is why 5 to 7 is marked best fit.
How many recommendations should the quiz have?
Number of distinct outcomes a shopper can land on, restricted to quizzes of 5 to 12 questions so quiz length is held roughly constant. 1,410 quizzes.
| Recommendations offered | Quizzes | Completion |
|---|---|---|
| 2 or fewer | 100 | 63.3% |
| 3 | 220 | 78.4% |
| 4 | 410 | 81.7% |
| 5 to 6 Best fit | 330 | 84.3% |
| 7 or more | 350 | 78.8% |
Two outcomes is the worst thing you can do: 63.3 percent, 21 points below the five-to-six band. If the shopper can guess both endings from question one, the quiz has no reason to exist. Past six it falls back again, because more outcomes need more questions to separate them.
Seven build choices, measured
Every setting below is a switch in the Quiz Maker editor. Completion is the unweighted mean across quizzes, split by quiz length. An effect that flips sign between the two columns is not a finding, and we mark those.
| Build choice | Setting | 5 to 12 questions | 13 or more |
|---|---|---|---|
| Let them skip the email ask | Skip allowed | 86.9% +10.1 | 67.6% -4.5 |
| Email required | 76.8% | 72.1% | |
| Progress bar | Hidden | 90.6% +12.8 | 73.8% +2.8 |
| Shown | 77.8% | 71.0% | |
| Result redirects to a URL | Redirect set | 84.8% +5.5 | 74.8% +3.7 |
| No redirect | 79.3% | 71.1% | |
| Back button | Blocked | 80.1% +1.0 | 74.3% +5.2 |
| Allowed | 79.1% | 69.1% | |
| Recommendation has an image | Image | 80.4% +1.9 | 73.4% +3.3 |
| Text only | 78.5% | 70.1% | |
| Images on most questions | Yes | 80.2% +0.8 | 72.0% +0.6 |
| No | 79.4% | 71.4% | |
| Recommendation has written copy | Described | 79.3% | 73.3% |
| Name only | 81.0% | 71.1% |
Written copy on the recommendation is the one choice with no verdict: it loses 1.7 points on short quizzes and gains 2.2 on long ones. A difference that reverses direction is noise.
Fix
Turn on "Allow respondent to skip lead capture" on the Leads tab, and add a redirect URL to each recommendation on the Products tab. Two switches, about fifteen points on a short quiz.
What this means for your quiz
Five decisions, taken straight from the tables above.
- Open with a question, not a welcome screen. Worth 20 points.
- Make question one the easiest one you have. It costs four times what any other question costs.
- Build 5 to 7 questions and offer 4 to 6 recommendations.
- Let people skip the email ask if your quiz is short. Worth 10.1 points.
- Send every result to a product URL. Worth 5.5 points, and it is the only setting that ends in a purchase.
Free, no account needed. The editor is at the top of this page.
Methodology, sample and exclusions
Dataset
The benchmark draws on every quiz published on Quiz Maker that routes a taker to a named recommendation and asks for their contact details. In product terms that is the outcome engine with lead capture enabled, the configuration the Recommend a Product or Service quiz type creates by default. A quiz qualified for inclusion if it had reached a published, active state and attracted at least one completed response.
Final sample
310,310 published outcome quizzes were scanned. 23,105 had lead capture enabled, and 7,669 of those had collected at least one completed response, drawing 1.8 million completed responses in total. The corpus is fully anonymised: no quiz contents, no taker identifiers and no creator identifiers are reported.
How the cohort is identified, and its limit
Quiz Maker stores the outcome family a quiz belongs to, not the specific template a creator picked from the menu, so a quiz built as Recommend a Product or Service cannot be separated after the fact from a personality quiz whose creator switched lead capture on. We therefore define the cohort by what the quiz actually does, routing to a named recommendation and asking for an email, rather than by which menu item created it. This is a deliberate and slightly broad definition, and a minority of quizzes in the corpus read as personality quizzes with a lead capture attached. Where that matters we say so.
Session reconstruction
Completion, drop-off and timing are not taken from stored counters. Each session carries a per-screen timing record, and the number of entries in it equals the number of questions the taker actually answered; this was validated against the stored answer set, where an unanswered question carries an empty answer. That gives an exact question-by-question drop-off curve for the 700,257 sessions that carry a usable timing record. Those sessions are a subset of the 1.8 million completed responses in the corpus, not the same figure counted twice: completion and build-choice rates are computed across the full corpus, while the question-by-question curve is computed only on the reconstructed subset.
Completion rate
Completion is the share of sessions that reached the result screen. A session only exists once the taker has answered at least one question, so visitors who bounced off the first screen without interacting are excluded from both numerator and denominator; the rates here describe people who genuinely began the quiz. Rates are reported as the unweighted mean across quizzes, not across responses, so a single high-traffic quiz cannot carry a band.
Inclusion thresholds
Quiz-level rate comparisons use the 2,800 quizzes with 100 or more observed sessions. The drop-off chart is restricted to quizzes of 12 to 30 questions so that every question position shown exists in every quiz counted; that restriction leaves 1,340 quizzes and 322,340 sessions.
How the drop-off chart is computed
For each question position we take the share of sessions still answering at that position, then report the difference between consecutive positions as the points lost at that question. Bar lengths are that loss scaled against the largest single loss, question one at 5.4 points, and the "still in" column carries the underlying survival share so both readings sit on the page. The chart stops at question 11 because that is the last position at which every quiz in the restricted set still has a question to show.
Controls
Quiz length confounds almost every build choice, because creators who build long quizzes also behave differently in other ways. Every build comparison is therefore reported separately for quizzes of 5 to 12 questions and quizzes of 13 or more, and a difference that reverses direction between those two bands is called out rather than averaged into a single headline number.
Device mix
69.6 percent of sessions run on a phone and 30.4 percent on desktop, with mobile finishing about 30 seconds quicker and completing 1.7 points lower. We report it here rather than as a build choice, because the device split is not a setting anyone controls.
What is association and what is not
None of these comparisons is a randomised test. Creators chose these settings, and a creator who hides the progress bar may differ from one who shows it in ways we cannot observe. The question-by-question drop-off curve is a direct measurement and does not depend on this caveat; the build-choice table does, and should be read as a strong prior worth testing on your own quiz rather than as a proven cause.
Exclusions
Deleted, stopped and moderated quizzes were excluded. Sessions marked deleted or over the creator's plan limit were excluded. Sessions with an implausible duration, under five seconds or over one hour, were excluded from timing figures only, not from completion. Quizzes that show several questions per screen were excluded from the drop-off and build analysis, because a screen is not a question there.
What this benchmark does not measure
This release reports session-level drop-off, completion and timing. It does not report email capture rate, post-result click-through, revenue attribution or share rate. Contact details captured by these quizzes are stored inside the response record rather than as a separate field, so a like-for-like lead rate cannot be computed across the whole corpus without reading response content, which this benchmark does not do.
Reproducibility
Anonymised aggregate tables are available on request for academic and competitive analysis. Please contact the editorial team via the Quiz Maker contact form to request access.
Find Your Signature Cologne
A fragrance finder quiz for cologne brands, niche perfumeries and DTC scent boxes. Eight questions on scent family, season, occasion and personality sort the shopper into one of four fragrance profiles, then route them to the matching collection with their scent profile captured.
8 questions
<2 min to complete
4 fragrance profiles
Product recommendation quiz maker FAQs
What is a product recommendation quiz?
A product recommendation quiz is an interactive quiz that asks shoppers a short set of questions about their goal, preference, fit, sensitivity, budget or use case, then routes them to a matching product, bundle, routine or plan. Instead of making the visitor browse every SKU, the quiz acts like a digital shopping assistant. Quiz Maker handles the questions, the scoring, the result page and the embed in a single editor and a single account.
What is the difference between a product recommendation quiz and a product finder quiz?
They are usually the same thing under two different names. "Product recommendation quiz" describes the outcome (we recommend X to you). "Product finder quiz" describes the shopping experience (you find your right product). Both run on the same engine: a small number of preference questions feed an outcome map that points at one product, one bundle, one collection or one plan. Pick whichever phrase your shoppers and your category use.
How does the quiz decide which product to recommend?
You decide, in advance, and the quiz follows your map. Every answer is linked to one or more outcomes, and the outcome carrying the most weight when the shopper finishes is the one that shows. Name your products first, then write questions whose answers separate them cleanly. Match percentages stay hidden on the result page unless you switch them on.
How long does it take to build a product recommendation quiz?
A first working draft takes about half an hour. The editor on this page opens already set to recommend a product, so you write 5 to 7 questions, add your products, and match the answers to them. You do not need an account to start. Editing is where the real time goes: budget an afternoon to get the wording, the images and the product links right for a catalogue you care about.
Can I build a product recommendation quiz without coding?
Yes. The whole flow is no-code: write the questions, add your products and outcomes, map each answer to the outcome it should push, then publish. The embed is a single script tag rather than an iframe, so it slots into Shopify, WooCommerce, WordPress, Webflow, Wix, Squarespace or a plain HTML page without an app install and without editing your theme files. The live examples on this page run on that same builder.
Can I use a product quiz with Shopify or WooCommerce?
Yes, and on Wix, Squarespace, Webflow, WordPress or hand written HTML too. There is no app to install and no platform lock: you paste one script tag onto a product page, a collection page, a landing page or the homepage. Results link out to your product pages, collection pages or cart URLs, so the shopper acts inside your existing store. Nothing about your theme, checkout or hosting has to change.
Does it connect to Klaviyo, Mailchimp or my CRM?
Yes. Emails and the answers behind them pass to Klaviyo, Mailchimp or HubSpot through Zapier style workflows, so a lead lands already tagged with skin type, budget or fit and your next email can act on it. You should not need a developer to wire that up. Ask for the email before or after the result, with a skip option, so the shopper sees the recommendation either way. More on lead generation quizzes.
Can I recommend specific products, or only categories?
Both. An outcome can be one SKU, a bundle, a routine or a whole collection, and you pick the level that matches the decision the shopper is actually making. Each outcome carries a product image, a short line explaining why you are recommending it, and a link or cart URL. There is no checkout inside the quiz itself, so the shopper lands on your product page or cart to buy.
Will it work if I have hundreds of products?
Yes, if you map at the right level. The quiz matches answers to outcomes you define; it does not read your catalogue and match SKUs on its own. A store with hundreds of products should point outcomes at collections, bundles or a shortlist of hero products instead of at every item. That is better shopping anyway. Three strong matches on a result page beat thirty, and hand mapping every SKU is slow work in any tool.
Will a product recommendation quiz slow down my store?
The quiz loads from one script tag and renders inside its own container instead of rewriting your theme, so the footprint is small. No third party script is free, though, so measure it rather than trust it. Run PageSpeed Insights on the page before and after you embed, and put the quiz on the pages where it earns its place instead of site wide. A landing page embed keeps the cost on that one page.
Does a product recommendation quiz work on mobile?
Yes. The quiz layout is responsive, so questions, images and result pages resize to the screen instead of being shrunk down. Test it the way your shoppers will meet it: open your published quiz on your own phone before you promote it, check the answer buttons are big enough to tap with a thumb, and watch the result page images load on a normal mobile connection. Mobile is where quizzes most often fall down.
Is the free plan actually free?
You can build a complete quiz in the editor on this page, questions, products and result pages included, without creating an account or entering a card, so you can judge it before committing to anything. Paid plans cover higher response volume and the features a live storefront needs, and the current limits are published on the pricing page. Read the response cap before you build, because that is the usual sting in this category.
How many questions should a product recommendation quiz have?
Five to seven questions is the right band for most ecommerce stores, and every question should change the recommendation. Our benchmark of 7,669 product recommendation quizzes puts that band at 81.6 percent completion, against 76.9 percent at 8 to 10 questions and 69.5 percent past 20. Skin type, goal, budget, size, fit, sensitivity, scent preference, use case and gift recipient all earn their place. Brand awareness or "how did you hear about us" questions do not, and they cost you completions. If you cannot say honestly that a question shifts the result, cut it. See quiz examples for how that looks in practice.
How do I know if my product recommendation quiz is working?
Pick your numbers before you launch. Four matter: how many shoppers start the quiz, how many finish, how many hand over an email, and how many of those go on to order. Tag the links on your result pages with UTM parameters so the orders show up in your own analytics and not only in a vendor dashboard. We do not publish an average lift figure, because the honest answer depends on your traffic, catalogue and questions.
