Skip to content
Dino Discover Dino Discover

Why Dino Discover?

Dino Discover is a guided product-finder for WooCommerce stores. A shopper opens a page on your storefront, answers a few questions, and lands on a short list of products that fit them — instead of scrolling through a long catalogue, applying filters, or guessing at search terms.

If you sell more than a few SKUs, or if your shoppers don’t already know exactly what they want, a quiz tends to convert better than a search box. Dino Discover is the part of that puzzle that lives in the storefront.

What makes it different

Section titled “What makes it different”

Built around WordPress, not around itself. The Builder runs on @wordpress/element. The storefront bundle uses Preact through the preact/compat alias to hit a tight gzipped budget. Everything that talks to the database goes through HPOS-compatible APIs. There is no parallel admin universe — it’s a WordPress plugin all the way down.

Drafts are real drafts. Edits in the Builder stay in a draft table. The storefront never reads drafts. The only thing on the site that can read a draft is the admin preview tab, gated by a one-shot nonce that expires after 30 minutes. So you can rework a quiz mid-day without worrying that a shopper will catch you mid-edit.

Every publish is a snapshot. Each time you click Publish, the draft is serialised, hashed, and stored as an immutable row. The History tab lists every snapshot — restore any one of them back to draft, tweak, republish. The undo is the workflow.

Swappable engines. Four engine interfaces — Question, Recommendation, Coverage, Statistics — sit in includes/Engine/. A site-wide filter is enough to replace any one of them. When a smarter recommendation backend exists, the storefront and the REST API don’t have to change to use it.

Accessible, low-bandwidth storefront. The shopper-facing quiz is WCAG 2.1 AA, keyboard-navigable, reduced-motion-aware, and tuned for low-end devices. The page that hosts it loads only a small bootstrap; the full bundle fires on the first interaction.

Who it’s for

Section titled “Who it’s for”

Merchants with a deep catalogue. Hair care, supplements, wine, garden plants, tools, gifts — anywhere “browse all 400 SKUs” is the wrong answer.

Stores that want to capture interest, not just orders. The lead capture surface lets you ask for an email at the end (or before) results. The data routes to Klaviyo, Mailchimp, a webhook of your choice, or just the WordPress admin.

Agencies building custom storefront experiences. Engine interfaces, ~55 REST routes, roughly thirty named hooks, and theme-overridable templates mean you can build something opinionated without forking the plugin.

What it isn’t

Section titled “What it isn’t”

Not a survey tool. Quizzes here exist to make a recommendation. If you need to collect arbitrary survey responses for research, this is the wrong plugin.

Not a recommendation widget for browsing. Dino Discover starts with an explicit “ask me some questions” interaction. If you want a “related products” carousel that just appears beside everything, there are better shapes for that.

Not “drop in and forget”. A good product-finder rewards real thinking about which questions actually filter your catalogue. The Catalogue Coverage page is there for exactly that — to show you which products your current question set can reach, and which it can’t.

A few things we deliberately don’t do

Section titled “A few things we deliberately don’t do”
  • No FX conversion in recommendations. Currency is treated as a display concern; the recommendation engine doesn’t second-guess what WooCommerce reports. If you run a multi-currency setup, the engine honours the merchant-configured prices and rates.
  • No fork-required customisation. Everything you can usefully change on a production site is reachable through a filter, a hook, a template override, or a REST payload. The PHP code in includes/ is the canonical source of truth — if a filter isn’t enough, the path is “extend the interface”, not “patch the plugin”.
  • No black-box AI in v1. The default recommendation engine is a tag-weighted upvote sum. Transparent, fast, debuggable. The interface is ready for a learned backend, but the merchant always controls what ships.

What’s next

Section titled “What’s next”