Written by Oscar
Supercharging search on your website
Why do you need search on your site? Because you want to make sure your visitors don’t get frustrated once they’ve arrived. You don’t want a potential customer returning to Google to find content on your website and stumbling upon a competitor!
Yes, search is hard – Google has shown us that. But this doesn’t mean you can’t make effective use of search for your website. After all, you have the benefit of working within the confines of your product or service and most likely a good idea of what your audience is interested in.
In this piece I cover some of the broad search strategies and types of search that you can implement on your website relatively easily.
Developing your search strategy
It all starts with your users. What are they trying to achieve with the search? Are they looking for a specific technical document? A broad list of articles around a theme?
We usually separate search queries into three broad areas: navigational, informational, and transactional. The relevancy of each will depend on type of site you are managing and the type of visitor you are focusing on.
To help you understand your site content from a search perspective:
- Create a table summarising the possible search intents of different users. If you can, speak to some of your customers.
- Map out some of the possible search queries and destination pages based on your understanding of your website personas.
- If you already have a search and you’ve been tracking search terms, extract these and explore patterns in a qualitative research tool. (Or just use Excel!)
- Research the types of keywords and phrases used in Google and other search engines. They will give you an idea of the user intent.
- Audit your content and assess the density of keywords across pages. Review your most important pages for the frequency of a given term.
- If you have categorised content, how do the respective landing pages perform?
- Speak to your SEO specialist to get their expert assessment.
Your search also works in conjunction with your site navigation. If you find people are using your site search to find specific pages, it’s worth reviewing your navigation to ensure it’s intuitive.
All your findings will be very helpful for your SEO strategy too. So by optimising your website search, your SEO will also benefit.
Types of website search
There is a wide range of search functionality you can add to your website – either natively on your website or by leveraging third party search services. In our experience, the most useful categories of search include:
- Basic keyword/phrase search. Where it all starts.
- ‘Fuzzy’ search for those times when we misspell or looking for something more general.
- Autocomplete for fast look-up, and when it’s useful to cut to the chase and show what’s available.
- Facet search to enable filters specific to the search results or data using a range of attributes.
- The old trusted filters/sorting to narrow content quickly using a set of global categories. It has speed and familiarity going for it.
And now for a little more detail…
This is the most common type of search function on a site as it’s often what comes out of the box with many CMSs. But it’s usually rigid as it does not cater for spelling mistakes or variations of a word or phrase. For example, if you search for “creative innovation” you will only get results that exactly match these words. “Create innovation”, “creative tech innovation” would not be included in the results.
Sometimes this requirement for accuracy can be a useful way to find what you’re looking for quickly – if you know what you want and you don’t want any ‘guesses’ by the search function to get in the way of what you’re after.
The results can be made more helpful by showing the snippet of text the word appears within on a given page. Although, of course when the term appears multiple times, it will not necessarily show the most relevant snippet. This may lead the user to try a different page.
It can be helpful to be less sensitive to the exact search term.
At its simplest, fuzzy search works by also searching on variations of the keyword or phrase. For example, plurals, misspellings or other tenses. This is useful when someone is looking for something general rather than a specific term.
Fuzzy search works by calculating the number of changes that need to be made to the search term to match potential words in your content. These could be swapping a letter, deleting, or adding. The sensitivity of the search can be controlled by tweaking the settings of the algorithm.
But it does come at a price. As you widen the net, relevancy goes down. This especially true for single word searches but if you get the balance right for your data, you can improve the chances of a good search result for two or more words.
As you start typing your search term, the search function will try and determine the word or phrase based on what is available in the website. This has the benefit of immediately reassuring the user that they are likely to find what they’re looking for.
A common use for this is for the names of products or SKUs on an ecommerce site and it can be a powerful marketing tool: based on the product they’re typing you can also include product variations, offers, and complementary products.
Autocomplete also works well with fuzzy search. If the user has misspelled the word or product, the autocomplete will suggest the correct version.
You probably use this type of search regularly because that’s how search engines work.
This type of search is often found on ecommerce sites. You’re probably using this kind of search without even thinking about it. For example, when you filter a category of products based on a set of specific attributes such as colour, size, shape etc.
Ultimately it is a great way to avoid a “no results” message, which is a sure way to lose the interest of a potential customer.
We manage a lot of very content-heavy websites and have found facet search helpful in providing context for a list of content.
Here are two examples of different ways that we often use it on a website:
1. Dynamic filters for search results
A user searches for a keyword or phrase and is presented with a list of 100 items. To help them find what they’re really after, a set of categories and other attributes are automatically generated based on the results. The category names give them a better idea of the topics covered and a quick way to refine the results by one or more filters that match their interest.
To help the user, we may also include relevant images or video thumbnails. And while we’re at it, we can also allow them to play the video in an overlay in the search results.
2. Content landing pages
There is very little difference between a list of content generated by a search query and a list of all content in the website, so we can deploy the same facet search functionality. The filter groups generated are based on the content available. If new categories are added to the site, they are automatically included.
Filtering and sorting
For simpler content sites, a more straightforward way of filtering content is the best way to help the user. A set of top-level categories gives the user a good idea of the type of content available. Category filters most likely work one at a time – i.e. if the selected category is changed, the previous one is deactivated. If we do allow multiple categories, the filter shows content that matches one or the other. Not necessarily both.
Multi-category filtering requiring all results to match all category selections very quickly becomes less helpful when selecting too many categories leads to a “no results” scenario. At this point, we would recommend exploring a facet search solution.
I think it’s safe to say that “sort by price” or “by newest” are some of the more helpful ways to quickly finding what you want.
Once you’ve determined the type of search you want to provide for your visitors, you need to ensure your content is in good shape to deliver what you need. We talk a lot about the importance of content and this is no exception. There are no shortcuts.
Unfortunately plugging in a search solution is only part of the full set-up. Even the most sophisticated search is less useful if the content fed to it lacks context and quality meta information.
You may need to:
- Create new content to meet users’ search term expectations.
- Check the correct fields in the CMS are available to be searched and weighted accordingly. (Or, indeed, add new fields just for search.)
- Categorise content and make sure it is indexed. And use these as needed in any frontend filtering.
- Don’t forget to include assets such as PDFs, video, and images. For this to work you need to ensure the content of the files is available to the search and the video voiceover is transcribed (which, of course, you should be doing anyway for accessibility).
- Create a guide for content editors to ensure new content meets the needs of your search strategy.
Given the number of variables in a search set-up, it’s always important to test the final solution with the persona-led keywords and phrases you explored earlier – do the results provide the users with what they wanted? Does it meet their search intent?
When you’ve got everything set up correctly, it’s time to go live with your supercharged search, sit back and watch users finding what they want on your site.