books, literature, reading, sale, street, market, discount, old, buy, shop, movie titles, scripts, discount, discount, discount, discount, discount, scripts, scripts
Photo by Life-Of-Pix on Pixabay

Shoppable Catalogues

Part of Social commerce conversion

Testing whether a shopping collection helps users choose

Set a useful choice question, compare collection and direct routes fairly, and judge outcomes beyond collection clicks.

This is an editorial pen name used by the publication, not a claim of an independently verified personal identity. Original editorial credit: Desk Social Commerce.

undefined

Compare a website shopping collection with a direct route to the same products to learn whether the extra page helps shoppers choose. Collection clicks alone cannot show whether shoppers found a suitable item or completed an order.

Define the choice problem

Choose products shoppers might reasonably compare, then write a task in plain language, such as choosing a bag for a stated use. Before the test, define which actual product attribute makes a choice suitable, such as size or intended use.

Build the collection labels around that attribute. If size is the distinction, for example, test labels such as “compact” and “larger capacity” only when the product information supports them; compare the collection with a direct route to the same products without that extra guidance. If the collection only repeats product tiles already visible in the post, there may be little useful guidance to test.

Keep product eligibility, prices and availability current in both routes. The question is whether the collection helps a decision, rather than which products deserve promotion.

Compare routes fairly

Create an entry URL for the social post and configure your website to assign each eligible visit at random, before navigation, to either the collection or a direct product route. Record the route assigned to each visitor and keep returning visitors in the same group where the setup permits. Keep both routes live at the same time with the same eligible products, prices and availability.

Make the primary outcome the suitable-choice rate: the share of assigned eligible visitors who select a product that matches the task’s pre-written criterion. Track completed orders per assigned visitor separately as a commercial outcome; product-page visits and cart actions can help explain the result. Keep assignment consistent for returning visitors where the testing setup permits it.

With ecommerce events configured, Google Analytics can measure actions such as selecting an item, viewing item details, adding a product to a cart, starting checkout and making a purchase.

If your site cannot randomise visitors at entry, track the same outcomes for each route and record differences in traffic source, device, stock and promotions. Compare the route groups with those differences in mind, but label the result observational: visitors reaching each route may differ, so the comparison cannot establish that the collection caused a change.

Record stockouts, promotions and audience differences. Set a stop rule against a sample size agreed before launch, rather than ending the test when early clicks rise. Optimizely lists setting sample size in advance and not peeking at results among guidelines for valid experiments using classical statistics (Optimizely).

For a formal significance claim, review the p-value and confidence interval with statistical support (Optimizely). Treat results without that support as directional, not proof that the collection caused better choices.

Collection vs Direct Route: Measuring Effective Decision Support

Primary Outcome: Suitable-Choice Rate
Share of eligible visitors selecting a product matching the task criterion
Commercial Outcome: Completed Orders per Visitor
Track via Google Analytics ecommerce events
Supporting Metrics
Product-page visits, cart additions, checkout start

Judge whether choice improved

If shoppers visit the collection but few select a product that meets the task criterion, inspect whether its labels help them compare the relevant attribute. If suitable selections rise but orders do not, inspect the product page and checkout. These are questions for investigation, not conclusions from one metric.

Keep the collection if it helps the chosen outcome without creating material problems elsewhere. Simplify or remove it if the extra screen adds work without useful guidance. Revisit the decision when the products or shoppers' questions change.

More from Shoppable Catalogues