Let the AI read every answer together, write the rules yourself, rank by tag, or score across dimensions. Then branch shoppers between pages, show or hide anything by condition, and calculate the numbers your results need.
Every results page picks its products one of four ways. Each has a sweet spot, and you can switch later without rebuilding the quiz.
Reads every answer together and picks from the catalog.
Best for: Fastest setup, catalogs that change oftenKeep in mind: Picks from the catalog, no rules to writeExact products behind if/then rules you write.
Best for: Full control over every resultKeep in mind: Needs updating as the catalog changesRank, match or exclude by product, variant, collection or tag.
Best for: Large catalogs with clean tagsKeep in mind: Updates as the catalog changesHighest-scoring products across named dimensions.
Best for: Assessments and personality-style quizzesKeep in mind: Scores across dimensions you nameJump rules branch shoppers between pages, redirect to a URL with answer tokens, or auto-advance on a timer. Any component or answer option can be shown or hidden by a condition: equals, contains, between, is answered, opted in, is highest or tied for highest. Flow View maps every page and branch so you can see the whole quiz at once.
Formulas calculate numbers from answers: portions, sizes, quantities. They can stay hidden, feed rules and marketing tools, and set the quantity a product is added with. Value tables map two answers to one value, pasted straight from a spreadsheet.
Portions, sizes and quantities worked out from what the shopper answered.
A hidden formula can drive rules, sync to marketing tools and set add-to-cart quantity.
Two answers in, one value out. Paste the grid from a spreadsheet and you are done.
Some products are a grid, not a list. A product matrix maps two answers to products or pinned variants, so shade matching is a table you fill in rather than a stack of rules you maintain.
You don't need a perfect plan to begin. Start small, and let Phoenix help you improve from there.