Recommendations: Reco Studio, widgets and their results
Show personalised product or article recommendations on your site — the quick start, where blocks appear, what they recommend, how they look, A/B tests and the results.
Reco Studio puts blocks of recommendations on your site — "You may also like", "Others also bought", "Read next" — chosen for each visitor from what they and others did. No code beyond the Monfri script is needed. Open it in CDP: Activate → Reco Studio.
The quick start
The quickest way to a first block:
- Choose your kind of site — E-commerce (products, viewed, purchased) or News / media (articles, reading, interests).
- Give the block a Placement slug — a short name such as
product-page, in lowercase letters, numbers and dashes. - Monfri shows the code to paste — Paste this code on your site. On a site where the Monfri script already runs, you can skip the code and give the placement a position instead, as described below.
The five tabs
Reco Studio has five tabs. The first four make up every block; the fifth shows how they do:
| Tab | What it is |
|---|---|
| Placements | Where a block appears — on which pages, at which spot, on which devices. |
| Strategies | What it recommends — the sources, the filters, the ranking. |
| Widgets | How it looks — layout, fields, colours, sizes. |
| Experiments | Which works better — an A/B test between strategies. |
| Analytics | The results — impressions, clicks, conversions. |
Placements: where blocks appear
A placement ties a spot on your site to a Default strategy, a Default widget and a Default layout. Its status is Active, Draft (not shown) or Stopped.
To put the block on the page without code, set the position:
- The element — a CSS selector of the spot, for example
.article-body. Instead of typing it, use the picker: drag Monfri Picker to your bookmarks bar, open any page of your site and click the bookmark; then click the exact spot. The top half of an element means before it, the bottom half after it. - Where to insert — Before the element, After the element, At the start (inside), At the end (inside) or Replace the element. Prefer the first four: they do not remove anything from your page.
- Pages — a path pattern, for example
/news/*for every news article. Empty means every page. The picker fills in the exact page you picked on — widen it to a pattern if the block belongs on many pages. - Devices — all, Desktop only or Mobile only.
- Frequency cap — at most so many showings to one visitor in so many hours; 0 means no limit.
Strategies: what to recommend
A strategy is a small diagram: sources feed candidates, filters remove some, the ranking orders the rest. Drag pieces from the Palette onto the Diagram, or click to add them.
| Source | What it brings |
|---|---|
| Trending | Popular right now, with older popularity counting less. |
| Similar (content-based) | Close in meaning to the product or article on the page. |
| For you (personalized) | Based on this visitor's own profile. |
| Together (co-visit/co-buy) | "Others also read / bought…" |
| Knowledge graph | Articles about the same people and topics — news sites. |
| Recently viewed | The visitor's own history — "continue reading", "recently viewed". |
| Editorial | Items you choose by hand, by their IDs. |
Filters: Exclude already seen, Freshness (articles newer than a number of days), Exclude out of stock, Limit per category, Min. words in title. The ranking weighs the source, freshness and popularity, and can mix in Diversity so the block is not ten of the same. A Fallback source fills the block when the main ones find too little.
The preview on the right shows real results — With your context for a sample visitor, or Generic (no context).
Widgets: how it looks
The widget designer has a Live preview · sample data. Choose the layout — Carousel, Grid or List — the fields to show, a Heading, the Image ratio, the sizes for Desktop and Mobile, and the Theme colours: Background, Text, Secondary text, Accent, Border. A block takes your site's look within these limits and cannot break the page around it.
Experiments: A/B tests
An experiment splits a placement's visitors between Variants, each with its own strategy and share of traffic, against a baseline. Results by variant show in A/B experiments — results by variant on the Analytics tab. Let a test run until each variant has enough clicks before you pick the winner.
Results
Analytics — also Analytics → Reports in the menu — shows Impressions, CTR, conversions, and A/B results from recommendations, overall and By placement. Clicks and impressions are counted only for visitors who allowed analytics; the block itself shows to everyone.
If something doesn't work
The block does not appear
Check, in this order: the placement is Active; the page matches its path pattern; the CSS selector still exists on the page (themes change); the device setting; and that the Monfri script runs on the page — see Check that tracking works, and fix it when it does not.
The picker does not highlight anything
Open the page first by normal browsing, then click the bookmark again. Pages behind a login or in an app may not allow it — type the selector instead.
The block shows the same items to everyone
The visitor is new, so personal sources have nothing yet and Trending fills in — that is the fallback working. It changes as the visitor browses.
The block shows products that are out of stock
Add the Exclude out of stock filter to the strategy.