A lot of websites ignore on-site search. This is really strange, considering internal site search is really about prospects trying to have a conversation with the site owner, telling them what’s tough to find on the website—for free. Usually, you have to pay quite a bit to get that information.
No matter how good your analysts are at dissecting traffic and segmenting audiences, they will not get data about user intentions from just navigation data. There are three areas where intent can be inferred or analyzed:
- From search terms people use within search engines. (This is now a shadow of how useful it used to be, given how much of the search terms you don’t see since Google’s not provided change.)
- From the search terms people use once they reach your website. (These can be tracked by your standard traffic/ clickstream packages. If you are using Google Analytics, here’s how you can set up internal site search tracking.)
So the first thing you should really do if you aren’t tracking this yet is just make sure your traffic analysis tool can track this (most packages can) and take a few minutes to set it up. (For most packages, this is relatively easy, but you first need to understand how your internal site search tool creates URLs)
Once that’s done, you can improve the user experience not just for your internal site search page, but your whole website.
1. Make the things people search for easier to find
Visitors will search for the things they care about most: your products, your product specs, your categories, your price lists, the list goes on. Once you’ve identified the top searches, you can break them down into groups of things—product lines, categories, questions, etc. (Don’t judge; people will search how they search, not how you want them to search.)
Once you have groups of searches, you’ll be able to spot patterns:
- Some names of the groups you have created will not match with what people commonly search for. Evaluate how you name categories, and adjust as required to match your audience’s mental model.
- Group the searches by intent (awareness, interest, desire, action), then see where your pages need the most help. This will help you prioritize the fixes that need to be made on the website.
2. Evaluate the user experience for prominent things people still search for
Visitors will search for things that they cannot find after navigating through your website.
For some visitors, the site search feature is a catch-all, the final tool they use before they decide your website does not fit their needs. If you are finding products that matter a lot for the business, it’s time to feature those more prominently, both on the internal site search page and on the rest of the website.
3. Use internal site search data to improve the search page itself
Of course, all that data about what people search for internally can go a long way in improving the actual search results page on your site. For your 20-25 most popular searches, make sure you “feature” a result above the rest of the results. Do this for more searches, if possible. You can curate the experience for the top results and catch hundreds to thousands of searches with relatively little effort.
On top of featuring results, you can also use the data to cut down on your filters: if a section of the filters does not get used, you should evaluate why it’s there to begin with. Keeping the filters down to what most users need will go a long way in keeping your search page usable.
Lastly, you can use the data to help people recover from common errors. If there are common misspellings of a popular brand or product line and your internal search engine does not handle mistyped words well, you can specify featured results for the misspelled queries.
There are a number of other ways to make your internal search page better, but the idea is to keep using the data to cater to what visitors need.
Just knowing the issues to fix via data from internal site search doesn’t improve user experience or your conversion rates, but prioritizing projects based on the data will go a long way in ensuring that what you’re fixing is the problem that needs the most attention. Site search data could mean the difference between fixing something that very few users care about, and prioritizing the things that really move the needle.
This article originally appeared in Tim’s Monetate column March 13, 2014