After a site searcher submits their query, what appears next can make or break their journey. Should you show traditional search results? Redirect exact matches to their corresponding category pages? What should you show if no matches are found?
Optimizing your search result merchandising is about knowing best practices as much as it is about knowing what works for your customer.
Let's unpack four common search strategies that make great candidates for A/B testing.
If you’ve invested in building beautiful browse pages, consider redirecting exact match searches directly to them to leverage enhanced UI and curated merchandising.
But sometimes you’re unsure whether a category redirect or a search experience truly performs better while shoppers are in a search context.
Search results pages enable you to show suggested search refinements and show users a familiar UI that matches their expectations. Search and browse rankings may also be influenced by separate merchandising rules and algorithmic factors.
Let’s take the following American home furnishing site for example. This retailer and manufacturer’s search and browse pages display different grid and filter UI. Search returns 553 matches with a 4 results per row and minimized facet menu, and browse includes ‘Shop by Shape’ filter links and displays 3 products per row with only 267 results.
Furniture store’s search results for “coffee table” with 553 results
Furniture store’s Coffee Table category with 267 results
Banners and other graphics on browse pages can also dramatically change mobile experience. Because redirects affect every type of device, you should run head-to-head tests that compare pages with redirects to equivalent pages without them.
Sally Beauty mobile search results for “shampoo” |
Sally Beauty mobile browse page for Shampoo |
Before running your test, review which search redirects you have in place and their respective browse page experiences. You may choose to include all redirected searches in your test group, or only ones with similar UI elements on the browse page to keep like-with-like (provided your anticipated sample size will be large enough, by aggregate search volume).
The power of fuzzy matching means you never have to send visitors to an empty search results page. Today, it’s common practice to just show results even if there are no exact matches.
However, some of these experiences will be better than others. And there's a risk that matches can be so poor for certain queries that users lose trust in your entire website.
Sometimes search results have no relevance to a search query.
Consider testing fuzzy matching against the following alternatives:
If you’ve set your site to always show results, consider testing this strategy against serving a traditional ‘no results’ page with a prompt to refine their query and/or a few alternative recommendation pods.
A ‘no results’ strategy showing related items and best seller recommendation pods
You may also wish to test strategies on individual or groups of queries (such as brands you don’t carry) with a specific strategy, like showing a ‘no results’ page with explainer text and recommendation pods versus showing similar matches, or redirecting to the nearest browse page.
For example, if Adidas receives many searches for other consumer brands, it may test its ‘no results’ page against redirects across a bucket of search terms like Lululemon, HOKA, On, and Nike.
Adidas’ existing ‘no results’ experience
Adidas’ near-match for Lululemon search
Alternatively, the leading footwear and apparel retailer below displays brand tiles and best sellers for brands and products it doesn’t carry.
A leading footwear and apparel retailer’s ‘no results’ page
Note: before attempting a ‘no results’ A/B test, ensure that you incur enough no results events per month to support a split test within a reasonable time frame
Mastering search merchandising is an ongoing journey, but with these A/B testing strategies, you're equipped to refine and optimize your approach. Remember, even minor adjustments to category redirects and "no results" pages can yield major improvements in conversions and overall customer satisfaction.
If you want more testing ideas across all facets of ecommerce search and discovery, subscribe to our Split/Test Digest. You’ll receive the raw data, behind-the-scenes math, and real-world results straight to your inbox.