Poor site search results that return irrelevant results is just a bad user experience.
When implemented effectively, site search provides a fast and easy way for customers to bypass other navigational options and quickly find what they are looking for.
By entering a product type or name into the search box, users are indicating an intent to purchase. They may have come to the site looking for a specific item, and are therefore more likely to convert.
Stats on site search are quite hard to find (some of the few I found were from me back in 2013). This is partly because site search usage depends on the type of website, how well it works, and how prominently search is promoted to visitors.
For example, many fashion sites tend to make site search less prominent, to the extent that the option is easily missed.
Other sites, especially those with large product rages like Amazon, often choose to make site search more prominent, as it offers a shortcut around standard navigation. For this reason, site search usage on such sites is likely to be higher.
Site search user experience is mixed. As Baymard finds in its study of the top 50 US ecommerce sites, some basic features are often missing. For example, 22% of the sites studied didn’t support searches with colour variations.
Others simply don’t work in the way users would expect, especially as they’re accustomed to the relative accuracy of search engine results. This includes dealing with synonyms and thematic searches.
The bottom line here is that poor site search results that return irrelevant (or even zero) results is a bad user experience.
First of all, if you want people to use search, make it easy to find. Here, RS Components puts the search box in a prominent position, where it can be easily seen.
In addition, the text in the search box is also helpful, advising shoppers they can search by keyword, manufacturer or RS part number.
RS also makes its site search more user friendly by auto-suggesting possible results as a user types their search query. This has the effect of speeding up the process and producing more accurate results, as it reduces the chances of shoppers misspelling product names.
Showing images of matched products is also a useful visual cue which helps people find what they’re looking for.
Speed and accuracy are key here. Users shouldn’t feel like results are taking too long to load, but the relevance of results is the most important thing.
If results are inaccurate, and not relevant to the search query, shoppers won’t be able to find what they want, but will also lose trust in the site’s search function.
Ensuring that results are accurate takes some work. By thinking about how people search for products and the terms they use, you can help them to find more accurate results.
People don’t necessarily use the strict product names and terms, so it’s about adapting and anticipating the search queries shoppers will use. By looking at analytics data and the site search terms people actually use, you can learn a lot.
Site search results pages should help customers to make sense of the results presented to them, learn a little about the products shown, and allow them to narrow their selection down to a more manageable number.
The ability to sort and order results according to preferences like price and customer rating, to change the number of results shown and more, and help the shopper to customise the results page according to their preference.
Filter options are also invaluable so that users can further customise results. As the product options narrow, the customer should hopefully be close to finding the product they want as they’re only viewing those which match their selected preferences, and so the reducing number of options helps them move towards a purchase decision.
It’s also important to make it easy to select and deselect filters, as customers want to explore different options. Showing the number of matching products for each filter type also helps customers to avoid selecting so many filters that no products are shown.
Showing summarised product information on results pages helps customers to decide whether to take a closer look at or to dismiss certain products.
Here Curry’s shows five of the key technical specifications for these TVs, and adds key selling points such as free delivery, store collection options, and average review scores.
Other useful features include the ability to add items to a comparison table as users scan results, as this enables them to study products in more detail side by side.
Some sites help users to quickly interpret results to help them find the best deal. For example, Trainline picks out the cheapest fares for each option. This means users can quickly see the best fares without spending too much time scanning the price table.
Good site search isn’t just about designing one that works well, it’s about learning how people use it, and testing and optimising to constantly improve the experience.
Site search term data can tell you how people are using the search, and the kind of language they use to search, which can help you understand and improve.
Likewise, analytics data on site search performance can show how many searches are successful, how many users drop out after searching, and so on.
It’s also important to test site search with users to identify any problems they may have and to test improvements.