6 actionable tips for improving ecommerce site search

On-site search is often a fundamental part of a user’s journey when browsing and completing a purchase online, yet it’s so often overlooked or undervalued by merchants. Users who complete a search on an ecommerce site are said to be up to 30% more likely to convert, which is something I’ve definitely seen with clients in the past. This guide is designed to provide actionable guidance around improving your search results, search experience and understanding of how your search is performing.

The tips are based on my experience working with ecommerce clients of varying sizes from lots of different verticals – that said, there is no such thing as a one-size-fits all approach and you should carefully consider which of these approaches is most suited to your business, products and target audience.

1. Make sure you’re aware of what’s working

As with any project – the first thing you should do is try to get an understanding of what is working and, of course, what isn’t. Understanding how your search facility is being used, the revenue it’s driving and the comparative conversion rates against users who don’t use search, is a good start. These key metrics will help you understand whether search is working for you and whether you should be investing in improving your results.

An easy place to start with this is using the search usage report in Google Analytics, which provides data on users who have used search and users who haven’t, as can be seen below.

In the example above, users are almost twice as likely to convert when they’ve performed a search. It is exactly information like this that can, and should, guide the decisions you make around how much you want to push search on your store. If you’re in this position, it might be that you look to promote the use of site search in order to drive more conversions. You could also monitor the search terms that are performing well and use this data to inform navigation changes. This kind of information can give you a lot more insight into opportunities to drive additional revenue via improving your your site, as well as your search results.

2. Merchandise your results

In order to get the most out of search, you need to ensure that you’re optimising your results, particularly in terms of serving the right products and categories. For example, if someone is search for “coats” and you notice that the results don’t represent your full range or they don’t promote your key sellers or highest margin products, you’ll want to be able to increase visibility of specific products on high volume queries. You might also want to promote specific categories if you’re serving related links or category listings as well – this is all part of merchandising search results.

It’s important to also be aware of the performance of different products against different queries and test different ones based on what users are clicking on and what users are buying, particularly for your most common and best performing queries.

Solutions like Klevu provide lots of freedom around manual merchandising (as per the screenshot below), but also auto-merchandise results based on user behaviour (e.g. if a product has a higher CTR or CR%, it’ll be ranked higher for relevant queries), which can add a lot of value. As with merchandising at a category-level, this is a really important part of search.

Something to keep in mind is that a number of the mainstream ecommerce platforms don’t offer the ability to manually merchandise or auto-merchandise results, although most of the well-known third party solutions give you a lot more freedom and flexibility around the products you’re serving.

3. Use auto-suggest

Auto suggest / auto complete has become a lot more common over the last few years (this article references that 82% of online stores use auto-complete for search), as it helps to reduce the amount of time it takes for a user to get the results they’re looking for. There are lots of extensions for the mainstream platforms that allow for auto suggest and most of the third party site search solutions support it.

Auto suggesting queries and products can help to speed up the process of finding a product massively, as can be seen in the screenshot below. That said, various usability experts have said that auto-complete can cause more harm than good – For example, if you’re suggesting dresses when a user types “dress”, it may be that they’re not looking for dressing gown products – although lots of users will continue typing, it can be confusing. It’s therefore key to ensure that the suggestions are only shown when the query is highly likely to match what the user is looking for.

The example above is based on typing “red dr” and the user being served red dresses.

This example is from AO.com, inputting “ste” –
they provide suggestions on the query and product results.

4. Ensure you’re handling common misspellings and errors

Handling of mis-spellings and synonyms is a key requirement for search, as it can dramatically impact the user’s journey. Lots of out of the box search offerings don’t cater for errors – this article states that 60% of stores require users to input the exact words for the product type that the website uses, failing to return relevant products for things like blow dryers if “hair dryer” is typed, or multifunction printers if “all-in-one printer” is typed. Klevu for example, has an enhanced keyword index that automatically allows for common errors around keywords and also allows for stopwords, it also also indexes more product data in order to support synonyms.

It’s important that you’re keeping an eye on 0 result searches and low performing keywords so you can track errors and keywords that aren’t being handled correctly.

5. Make your search box more prominent

If you have a good search solution in place, making your search box more prominent in your header can represent a good way to entice more users to search. Merchants like John Lewis (below), Amazon and AO.com have very large search boxes, which are very obvious to users.

Merchants like Amazon, who sell complex products (like cameras), have very long search boxes, to allow for more detailed queries. According to Jakob Nielsen’s recent web usability study the average search box accommodates only 18 characters, with 27% of queries not fitting in at all. I really like the example below from John Lewis – the search bar is very prominent and encourages users to search in different ways, via the text prompt.

6. Look at providing more options around your results

Lots of merchants are using layered navigation and other types of attribute filtering to make search results pages more useful and easier to get straight to their desired products, as can be seen in the example below from Oneills.com. This example provides filtering options in a search box on the page, but you can also filter on the results pages in the same way.

A number of my clients have seen positive results from adding filtering on search results pages, allowing customers to further refine results for broader queries (e.g. Nike t-shirts, refined by colour). Giving users the ability to refine results can add a lot of value if you have a big product catalog or if you’re returning a lot of products for a more generic query.


Site search is an important area for ecommerce and it’s something that needs careful attention in order to be maximised. Addressing these issues and making use of new technology can help to generate a lot of additional revenue and improve your overall customer experience. The key thing really is understanding how your search is performing and what needs to be done around optimisation – from there you can start looking at alternative solutions and more effective tools.


Paul Rogers is a London-based Digital Marketing Consultant specialising in working with the Magento platform. Paul has worked in digital marketing and ecommerce for over 8 years in various different roles and companies. He also recently launched Replatforming.com.

Further reading:

Watch a 4 minute demo
Let us demonstrate why over 2,600 leading retailers worldwide have chosen Nosto’s AI-based personalization platform.