Understanding User Intent
I think we all know the feeling of typing something into a site search bar, only to be met with an onslaught of links that have nothing to do with what you wanted. And we try to rephrase our keywords or just search somewhere else - often a different website altogether. That’s why it’s so important for eCommerce retailers to understand what users want when they use the search bar.
If someone types in “headphones”, it could mean a variety of things. And this is where knowing your users comes in very handy. Using cookies and behaviour data, it becomes easier to anticipate what users want by showing them items based on their browsing history, purchases, location, and even gender.
I would say that this is quite necessary if you want people who visit your store online to keep coming back. Once users see that your site gets them and even helps them find things faster (or inspires them when they’re not sure what they’re looking for), they’ll want to come back again and again. The key thing here is that understanding user intent means letting go of assumptions and seeing each customer as unique individuals, instead of a faceless mass.
Optimizing Search Algorithms
We’ve all had the exasperating experience of typing ‘blue jumper’ into a fashion site’s search bar and being greeted with anything but. Or worse, nothing at all. I find it especially annoying when retailers seem to have no clue about what’s in their own stock.
That’s where a well-oiled search algorithm steps in. A search function that throws up predictable results or, even better, clever suggestions based on what you seem to be looking for is somewhat what keeps us coming back. It could be that the algorithm picks up on your search for ‘something blue’ and shows you blue everything - from dresses to shoes to bags - or, more precisely, ‘blue knitwear’. The second one makes it easier to keep your customer hooked and more likely to check out with something they like.
AI-powered algorithms can help retailers learn about how shoppers are rather searching for things. You may find that most shoppers land on a certain style of dress by typing in a synonym for it or by describing it a certain way and not the way you would expect them to. If you’re able to track these patterns and update the tags on those items, you’re going to be saving your customers a fair bit of time and money.
A fashion site worth its salt will also give you ideas on what you could be looking for if you aren’t sure yourself. So a shopper looking for ‘90s grunge’ would be shown outfits that are put together across categories or ‘looks like’ options across price points.
Enhancing Search Filters and Facets
Picture this - you're in the market for a new winter coat. You visit a site and search ‘winter coats for women’, only to be met with hundreds of irrelevant results that make you want to give up entirely. I’ve been there, and it’s not great.
Sort of. You sort of end up feeling slightly annoyed, even though it isn’t the coat’s fault. Filters are powerful tools that help shoppers narrow down their options based on specific criteria.
Sometimes these are kind of things like size, colour, material, and price. They also allow shoppers to specify their preferences quickly, helping them find exactly what they’re looking for without having to browse through irrelevant products.
It’s a sort of way to save time and help users make purchasing decisions faster. More or less. So how do we decide what filters to add.
Because a ton of filters just makes things more confusing. Sort of. The best way to do this is more or less by identifying your target audience and using data.
It seems like for example, if you’re selling dresses, size is usually a big deal for most people, so it would make sense to have that filter as an option. The way I see it, something i’ve found that works well is letting users select more than one filter at once. This allows them to use several factors as reference points when looking for something specific. For instance, if they want a dress in a particular style and fabric or shoe in a certain colour and material combination.
Implementing Autocomplete and Suggestions
Happens to the best of us - typing a search query and forgetting the actual name of the product we want to find on an e-commerce site. Or a customer’s typo coming in the way of their search being successful. Or those moments when we're unsure of how to spell something, hoping for a shortcut that will get us what we want.
These are arguably all opportunities for a store’s autocomplete functionality to shine. Autocomplete uses data analysis and machine learning algorithms to predict a customer's intended search query and displays suggestions as they type. These are usually based on site-wide search history, keyword popularity, and seasonal trends.
Product pages can also be included in the suggestions, presenting customers with quick results that don’t require them to move forward from the search box. It's quite handy at helping customers re-phrase their searches if they're getting zero results too. It saves time and gives them ideas for related products and keywords that might be more relevant for their needs.
The way I see it, this is one thing i never quite miss having on e-commerce sites, even though i usually know what i'm searching for. Feels like a little brain nudge each time my muscle memory doesn’t kick in while typing out a query.
Analyzing Search Data for Continuous Improvement
You know when you have that one customer who always comes into your shop and asks for the same thing. And every time, you wish you could just direct them to a shelf where all those things are kept. That's what site search is like for digital shoppers, except much more intricate and data-driven.
Businesses need to analyse their search data for all sorts of reasons - not least because it helps them learn what people are looking for. When you use software like Algolia or Unbxd, there are several ways you can look at this information. With Unbxd's Site Search Report, you can see common searches and which products were clicked on the most as well as which searches yielded no results.
This makes it easier to identify trends in shopping behaviour, such as high-performing search queries or synonyms that customers are using to find what they want. You can then use this information to improve your product descriptions or blog copy to make it easier for customers to find things. The report also gives you an overview of how well your filters and facets are working, which can help you decide how best to optimise them.
You can adjust how certain categories appear in the facets section of a web page based on data from the report. Another great feature of Unbxd's Site Search Report is that it allows you to see how long people spent looking at specific products after searching for a particular term. This is a great way to track trends in the market and stay on top of what your customers want.
And it's not only about the words they're typing in the search box - it's also about what they're clicking on. If a customer lands on a search result with five results and spends time looking at three of them but quickly moves away from two others, it may be worth investigating why those two weren't appealing. I think that's quite clever - never thought about it till now, but this type of analysis has endless possibilities, doesn't it. It's not only limited to finding out what customers want - it's also about how we can make better decisions when marketing or selling our products online.
Creating a User-Friendly Search Interface
Appears To Be picture this. You have a product in mind, and you know a website that sells it. The only thing left is to find it, but you can’t.
You try multiple search terms, but it doesn’t yield what you want. Sometimes, you may see an empty result page. More or less. No suggestions or options for further browsing.
Sort of. A user-friendly search interface can help avoid situations like this and help customers find products easily. It can be tricky sometimes, but it helps to keep the site search elements prominent. More or less.
A visible search bar with easy access options on both desktop and mobile devices can be helpful. If the website has multiple subpages, having a search icon on each page will help them go back to the main site search when required. The shopping journey is seemingly incomplete without filters and sorting features.
After customers have input their search queries, it’s ideal to have filters and sorting options they can use to narrow down their choices. Filters may look different for each category but common ones include - size, price range, colour, fabric type, design style or pattern, material used etc. Price filtering usually comes in the form of a slider. There’s usually also an option for the customer to sort their search results by relevance, price range or recency of product addition.
Something I’ve noticed quite often is how websites completely change up their filter options with minimal overlap between them across categories and departments. While it makes sense if there’s little room for overlap between certain categories - like shoes and t-shirts - filters like price range, colour or other attributes that appear across should remain the same for ease of use. I think e-commerce websites should make relevant recommendations on empty result pages rather than leave the customer confused about what to do next. Some websites have quick view pop-ups on results pages which may help the customer look at a product’s details more closely before committing to buy it right away or adding it to their cart or wishlist for future consideration.