Recommendations — Where to Start

A recommendation is three things: a catalog Personyze can read, an algorithm
that chooses from it, and a widget that shows the result. Get the first right and the other two are
choices; get it wrong and nothing downstream can save it.

Watch the 2:02 walkthrough: starting from a template, picking the page first, and what each group of algorithms recommends.

Video: Content recommendations: what to show each reader (2:11) — it plays at the top of Content Recommendation Algorithm types.

Start with the catalog

• Category & interest recommenders — recommend a section of the site rather than a single item: a product category, or an article topic.
• Recommendation algorithms — what each one picks, what it needs, and why the page you choose changes the algorithm and not just the placement.
• Score blend & presets — the account-wide weights behind every ordering, and the per-visitor price range.
• Recommendation Setup & Feeds — the feed itself, interests, filters, fallbacks, and how to review the data behind a recommendation.
• Setting the product / content feed — The one page to read first.
• Reviewing the data — What Personyze actually holds about each item, and whether it matches your site.

Then choose how it picks

• Product recommendations — algorithms, the wizard, the action guide, JSON delivery, and offline transactions.
• Content recommendations — the same for articles and content.

Then decide where it shows

• Widgets & add-ons — add to cart, bundles, social proof, upselling, auto-applied coupons.
• Outside your site — Apps, JSON and email — where to start when the recommendation is not rendered by Personyze.
• Recommendations as JSON — Take the output and render it yourself.