As site search users can be have a greater intent to purchase than the average, it pays to provide the best possible experience for these users.
In this post I’ll look at the presentation of site search results, and what retailers can do to ensure that users’ searches are successful.
An absolute must. Unless people have searched for a very specific product with a unique model number, or perhaps for a catalogue code, searches will result in a number of matching products.
For example, this search for a ‘red shirt’ on House of Fraser returns 810 possible matches. That’s a lot to go through.
However, the addition of product filtering options allows shoppers to refine the results by price, brand, reviews, and so on.
This should help the customer to find a more manageable number of results that match what they’re looking for.
Leave the search term there and let people edit it
Many sites delete the search term once you get to the results page, or show it in such a way that you can’t edit it. As here on H&M:
Leaving the term there, and making it editable allows users to append searches with other words, or to quickly correct any mistakes they may have made.
Provide non-product results
While most site searches may be product-driven, retailers should also cater for those users who are simply looking for information.
Here, Boden shows non-product results on different tabs. This is a useful way to separate the information out for different types of searchers.
Reviews should be used around the site where they can influence and help shoppers.
Showing them in site search results helps people to make a faster and more informed decision about the suitability of a product.
Here, Best Buy shows average review scores in its search results. With so many matching products, this helps a lot.
Allow filtering by customer reviews
To carry on from the previous point, Best Buy also offers ‘customer rating’ as a product filtering option, among a very comprehensive list of filters.
This allows users to quickly dismiss the worst-rated products.
Sprinkle in some social proof
Reviews work, but other forms of social proof can be incorporated into search results, as here on Booking.com:
It tells us that these hotels are in high demand, with people looking at and booking them today.
This is useful in one sense, as it tells the user that they need to book soon to secure it, and this can nudge them into making a faster decision.
Let people sort the results
Some people are price conscious, and want to see the cheapest products first, others may want to see newer stock, and so on. So let them order the results according to these preferences.
Provide viewing options
There’s no right or wrong answer for how to present search results. How many per page, do you paginate results or show them all on one page, grid or list view?
The obvious answer is to let them choose themselves.
Here, Macy’s provides plenty of choice.
Consider other useful display options
Here, M&S allows users to toggle between different types of product shots. They can view the product on its on, or worn by the model.
Learn how to deal with misspellings
Users will misspell items or make mistakes when typing search terms. The best way to deal with them is to anticipate what the customer meant (where possible) and provide the relevant results.
Avoid zero results / dead ends
There will be times when items aren’t stocked, or the customer’s search term is indecipherable.
In this case, provide some options. Here, River Island has some advice on searching, then shows its most popular searches and new products.
Consider quick view
Many sites now provide a ‘quick view’ option, such as Fossil:
This allows users to click and see a smaller version of the product page in overlay form.
The intention is to save them the time and effort they would spend clicking into and out of full product pages.
Show different views on mouseover
This is a quick way for users to see alternative product views on results and category pages.
Here’s an example from Wolf & Badger:
Learn from your data
Your site search data can tell you a lot about how people use it, and how well it performs. This information can be used to improve presentation of results.
For example, you can look at click depth. This can indicate the relevance of site search results and how deeply visitors go into the results.
So, if they’re clicking through to pages three and four of the results, does that mean they’re struggling to find what they’re looking for?
Data can also be used to identify best sellers, popular searches etc, and this can be used to preset relevant product suggestions.
This brings us to ‘searchandising’. Retailers can use site search results to give greater prominence to popular or recommended items, as Curry’s does here:
This is another feature that makes it easier for customers to see items in search results before they click through to product pages.
Macy’s allows searchers to change the colour of products shown in search results:
Try bigger images
Product images are important, and sometimes larger images will work better.
The tendency is for retailers to show plenty of products in search results, using relatively small images.
However, larger images can be more effective by allowing users to see more detail of products.
This A/B test on a Czech site demonstrates how this can work. The ‘winning’ version, which delivered a 9.46% increase in sales was the one in which the product images were largest.
Good site search is all about relevance, and doing as much as possible to improve the user experience.
Features like quick view and effective filters make it easier for customers to find the relevant products without too much hard work.
The caveat here is that these features should be tested and the best balance found for your site. While some principles are widely applicable, what works for one site doesn’t always work for others.