Users of site search are more likely to convert than the average user, so ecommerce sites should strive to produce the best possible experience. 

Here I’ve compiled a list of site search best practices, along with some excellent examples from retailers. 

Search boxes and functionality

There’s more to search box and site search functionality than you may think.

The placement and design of search boxes can make a difference to usage, while the addition of certain features makes for a better search experience.

Make the search box easy to spot

The prominence of the search box on the page can influence the user’s decision to make use of it to find products.

Therefore, if site search is important to your site, the prominence and visibility of the search field should reflect this.

Some sites, perhaps to maintain the clean design, tend to make their search boxes harder to spot, as in this example from Zara.

It’s very subtle and could easily be overlooked. The only mitigating factor is that it is not crowded out by other navigational elements at the top of the page.

It’s also a very interesting and unusual search function, as we explored in this article.

By contrast John Lewis, which places much importance on site search, makes its search box impossible to miss.

Make the search box big enough for typical queries

The need for this will vary from site to site, depending on the types of product stocked, but it’s important that search boxes are big enough to fit most queries.

For example, searches for things like electrical products which have long product codes can be harder when the text starts to disappear.

This means that users are less able to review the search term for mistakes as they type and makes it harder to edit it.

By contrast Amazon can handle the same lengthy product description with space to spare. Vital for a retailer with such a wide product range.

Use autocomplete for site search

This is a very useful feature which improves the search experience by reducing the work that users need to do.

As users type, products are suggested. If sites are smart enough, then these suggestions will reflect site search data and serve the most likely products first.

The use of images provides a visual appeal but also allows the user to check the products very quickly.

Autocomplete also helps when users may be unsure of spellings. This ensures that users find the product they need.

In this case, if you’re not sure how to spell the name of the Russian author of The Gulag Archipelago, help is at hand.

This is also very useful on travel sites for the same reason.

How marketers can use visual search 

Use auto-complete for merchandising

Auto-complete is very useful to help customers find the search term they want, and to avoid issues like misspellings, but it also offers opportunities for merchandising.

Here, as the site search sees that I’m looking for a wine gift basket, it starts to recommend products, complete with price, image and a snippet of text.

Site search data can be used to identify which products are most likely to appeal to searchers.

Allow users to search within a particular department

This helps users to narrow their search from the very beginning, making it more likely they’ll find what they need quickly.

It’s a great idea for sites with lots of products, like Newegg:

Place text in search box to encourage searches

The text prompts the users to search and also suggests the kinds of things they may look for.

As suggested by ConversionXL, sites can tie up these product suggestions with analytics data showing high performing products.

You can also use text which appeals to your userbase, as Spencer’s does here with ‘wut r u lkn 4?’

Make the text disappear

Retailers should also use JavaScript to ensure that the default text in the box disappears as users click to enter their own search term.

Don’t force them to delete the text before they can begin, as this is incredibly annoying.

Place a site search box on each page of the site

Having a search box on each page makes it easy for customers to get back to a product search from any point, and also provides an alternative method of navigation for users that arrive at product pages.

However, placing a site search box within the checkout process can provide a distraction for customers when they should be concentrating on making a purchase, so this is one area that doesn’t need one.

Allow users to search using product codes

This is a good option for retailers with magazines and catalogues, and these searches imply a real intent to purchase.

Here, if you search Argos with a code…

…you’re taken straight to the relevant product page:

Search results

The quality of search results is all-important.

They should be accurate and relevant to the user’s query, while the presentation of those results can have an influence on whether the visitor decides to buy


This depends on product labelling and metadata, but users will lose faith if results are a bit wonky.

Here, I search for blue shirts and this is exactly what I get.

Avoid zero results pages

This can be avoided easily by using autocomplete, which ensures that customers enter a relevant search to begin with.

If no autocomplete is present, the aim should be to avoid a dead end for users. 

House of Fraser achieves this by showing results for almost every search. It also retains the search term and search boxes so users can easily amend the search or start again.

How local languages should influence your search advertising strategies 

Throw in some social proof

Social proof can work very well, so why not use it within search results?

On I’m given review scores while the top result tells me there’s just one room left and that 21 people are looking at this hotel.

Show non-product results

People aren’t always searching for products. They may be looking for customer services, or perhaps buyer’s and how-to guides.

Here, Boden shows results from the help sections as well as style and fit guides.

Allow users to choose the way results are displayed

Allowing the user to select different views of results allows them to tailor their own search results.

Here’s an example from Kohl’s:

Searching for ‘returns’ on Three serves results that customers are most likely to want.

Use reviews as filters

Very useful. Reviews are powerful on product pages, so why not use them in other ways?

Here, users can filter by review score:

This is a great example from Abes of Maine. As well as filtering by reviews, users can choose best uses and features to narrow the search.

Filtering options

An absolute essential. Users need to be able to narrow down their searches using a variety of means to filter the product selection.

These include:

  • Product category.
  • Price range.
  • Size.
  • Brand.
  • Colour.
  • User ratings.

In general, the more filtering options the better, though this will depend on the size of the product range.

Here, has a comprehensive set of filters which help the user to narrow their search.

Sorting options

Sorting options allow the user to change the order of search results so they can view the most relevant results first.

This may be by price, showing the cheapest or more expensive first, or ordering results by relevance to the search query.

Handle common misspellings

John Lewis handles my typo on iPod well, serving results as if the mistake didn’t happen:

Here, House of Fraser serves this for the misspelling ‘siut’.

Make it easy for users to find products with synonyms

This is something that site search, and ‘no results found’ searches can tell you. Perhaps there is a common misspelling, or users are searching for a brand you don’t stock.

If so, rather than showing no results at all, serve up results that are related to the search term.

In this example, users searching for ‘Esky’ (a brand of cooler boxes) are shown similar products from different brands:

Show results in colour

Perhaps you have products in multiple colours. If a customer searches in this way, show it in that colour.

Show the search query on the results page

Showing the search term provides an instant reminder to the customer, but also allows them to append or remove words from the search in order to produce more accurate results.

Provide quick view options

Site search users often have a clear intent to purchase, and are more likely to convert than the average visitor.

The key here is to remove as many obstacles as possible from the purchase journey.

Quick view allows users to see a mini version of the product page and an add to basket button without having to load the page.

Here’s an example from Dune:

Selecting quick view allows shoppers to open up a mini-product page where they can view more details, select size and colour and add items to their basket.

Show different product images on mouseover

Mouseover effects on results pages can be useful to show products in context or from different angles.

On Bottica, hovering over product images on results pages triggers multiple product views, so shoppers can gain a better idea of the product with little extra effort.

Show technical detail

In the case of laptops, showing the specs in search results enables users to quickly compare features without having to visit product pages.

Are there any site search tips you’d like to share? Also, which features will improve ecommerce site search in future?

Let me know in the comments…

Meeting search intent: A model for measuring success