eCommerce Search - Making it easier for people to shop online with you

Rob Chettleburgh
By Rob Chettleburgh
14 Nov 2013

People like to shop in different ways. Sometimes they are happy to browse around your eCommerce store looking for inspiration and ideas. Sometimes they know what they are looking for and want to get straight to it.

There are many ways to help speed up the shopping experience when that’s what your customers want. Logical product categories and product filters are fundamental features of good eCommerce solutions that we’ve discussed before. But the most rapid and convenient of all can be simply to use a search box.

I say ‘can be’ because a poorly executed search facility can leave a customer bewildered and your eCommerce business no better off.  I sometimes get the impression that search boxes are included on sites just because businesses think they need to be there (which is most cases is true, they really do need to be there but it’s the wrong mindset), or because the eCommerce platform they are using makes it easy to add one.

What’s often lacking is a thorough thinking-through of the process to predict all of the ways that a real customer might use, or misuse the functionality. 

As an aside, I wonder whether the challenge of executing search effectively explains why both H&M and Primark don’t have search boxes on their sites but Asos, a company that only trades on the internet, has an excellent search facility.

Anyway, here’s a rundown of the main considerations if you want to offer your customers a superior experience with your search facility and, as the title suggests, make it easier for them to buy things from you.

Position

One thing you shouldn’t have to search for is the search box. At the top of the page, above the main navigation and usually on the right hand side seems to be becoming the standard. It’s where your customers will be expecting to find it so it’s a good idea to put it there. Ensure it’s included on all pages (except perhaps your checkout).

Reinforce the message by putting some text in the box. ‘Search’ is a simple option but more creative invitations like: ‘What do you want to find?’ or ‘I’m looking for…’ could be a bit more customer-friendly. And make sure the text disappears automatically when somebody clicks on the box.

Size

Generally you want your search box to be big enough to get noticed but not so prominent that it becomes a distraction (unless it converts brilliantly for you, more on that shortly). If you keep to a standard location then it should be easy enough for people to find without having to make it too eye-catching. It’s a good example of where A/B split testing can help you find the best approach.

One thing to plan for is the length of queries that people are likely to type in. It’s really helpful if people can see the whole query, even if it’s a few words. That way they can see whether they have misspelt anything and whether what they typed in is actually what they thought they were typing.

Tracking

And now down to the nitty-gritty. How many people use your search facility? How many results do they typically get? How many find what they were looking for and go on to complete a purchase?

If you can answer these questions you’ll know whether your search facility makes a positive contribution to the shopping experience and to your online sales, or not.

Event tracking in Google Analytics can be used to log the number of results found against a search phrase. When this is used in conjunction with Google Analytics’ inbuilt search tracking you can get everything you need to improve your search further. If a large number of people get no or few results then you clearly have a problem. 

From our experience, we find most clients are surprised at how poorly search performs. Experience also tells us there is always room for improvement. Look at Google after all!

Analyse your search traffic through Google Analytics. If you see a trend, as we did on a recently on a site, where we found far higher volumes of traffic convert via search then any other means, make it more prominent and measure the results. While this is a very bold search example I’m showing you, it worked amazingly well for the client. And that’s the point , if you’ve refined your search to a level where it really works then make sure it’s used.

eCommerce Search

Some possible solutions

You don’t want your customers to have to pass a spelling test before they can buy something from you. Often there are common misspellings that you should predict and allow for, here is a great example of this in action. The customer has misspelled Samsung here but we know what they’re after and still return results as if nothing had gone wrong.

eCommerce search misspelling

Predictive search is great. Here your eCommerce system is matching strings of letters to the product database and throwing up likely matches. The chances of an incorrect spelling being typed are reduced and your customer may not have to type the whole term before getting the match they want.

Think about plurals

Products tend to be indexed as singular terms eg ‘garden spade’. But it’s natural when somebody is performing a search for them to type ‘garden spades’. I’ve seen examples where searches can’t cope with the difference and return few results even when there were many matches for what the customer wanted on the site.

And remember, even though you call a product by a particular name, customers may use a variety of synonyms. A good search system won’t expect customers to use the exactly the same terminology you do. Your eCommerce system should be flexible enough to allow commonly used synonyms. Again, tracking customer behaviour will reveal what they are really typing into your search box. Making allowances for this is easier than trying to re-educate them.

Too many results?

Returning too many results can leave your customers no better off. Options you can look at here are the capability to narrow results by department before the search button is clicked and, again, predictive search, which can display more specific searches based on the same word. But unless you track the results people are getting, you’ll never know if getting too many results is an issue.

No results

What to do when you get no results? Even with the best search facility this can happen. Do you just say ‘better luck next time’ or do you do something to try to keep your customer engaged?

You could recommend other popular searches people have made based on data you have collected. Or you could make product recommendations based on their search and viewing history. Anything is better than an empty page after all.

Hopefully there are a few ideas there about how you can make search work more effectively for your eCommerce site. As always you’ll only get the best results after methodical experimentation and measuring.