Improving on-site search is vital for winning new customers and retaining them, particularly now shoppers are switching to mobile devices.

On-site search is so vital to helping website users find what they are looking for that it is surprising so many brands are not taking the channel more seriously.

It is even more surprising given the rise of m-commerce. It has never been more important to pre-empt and personalise results to get the most out of smaller screens. That was the message delivered to Econsultancy Live delegates by Andrea Polonioli, product manager for Coveo’s ecommerce division.

He began the presentation by underlining the importance of on-site search. It varies from site to site but Polonioli reports that between 30% to 60% of visitors typically conduct an on-site search. At the same time, those who do conduct a search are between two to four times more likely to convert than visitors who do not.

Polonioli warned brands that in the digital age, it is very easy for shoppers to go elsewhere. “We used to say if they can’t find it, they can’t buy it,” he said. “But now it’s more a case of if they can’t find it, they’ll find and buy it elsewhere.”

To avoid losing customers in this way, retailers need to bear in mind that Google has already raised expectations that search is effortless and intuitive. Helpful search has become “table stakes” for brands and yet the rewards are high.

“The phrase ‘just Google it’ is now ingrained in the everyday lexicon, and for a good reason,” Polonioli added. “The Google experience – namely an experience that is relevant, personalised and anticipatory – is what your customers demand, and expect. And importantly, being able to deliver Google-like experiences pays off.”

Attributes to focus on

The most obvious early steps are to embrace ‘type-ahead’ suggestions where a list of potential query options can anticipate what a person would likely be looking for. This is not only helpful, particularly on a mobile device, it also reduces how hard a shopper has to think. Offering options taps into the human brain’s ability to recognise a term, if offered to them, rather than have to think up the correct combination of words, which may be industry terms they are unfamiliar with.

Another way of helping customers is to embrace typos and understand the term a user most likely meant to enter. These have increased since screens got smaller and now around 15% to 25% of searches will feature at least one misspelt character.

A third major step forward can come from more mobile-friendly faceted search.
Providing familiar controls like drop-down menus and checkboxes with natural language labels certainly empowers customers to narrow down a large set of results to a small relevant set. But faceted search was originally designed for desktop users and translating the experience onto a mobile device is difficult, as it’s hard to show both the filters and the results on a small screen.

Modern AI-powered solutions address this problem by displaying a handful of filters – also known as ‘discovery tags’ – that are known to be optimal for the user. Polonioli used the example of a female shopper on a sports store site, who searches for ‘Nike’ and can easily filter the results by selecting one of the discovery tags being displayed, namely ‘shoes’, ‘tennis’ and ‘women’.

Further, Polonioli explained that being able to personalise search results is also increasingly important. Typically, ‘short head’ queries account for a surprisingly large portion of all search activity (and revenue). These queries are also very generic, though, meaning that they would benefit the most from personalisation. But personalisation isn’t easy, as for most brands the reality is that their customers don’t log in as they do on sites like Amazon, and 70%-95% of users visit a website less than twice a year.

All this means that the data available for personalisation is fairly limited. Brands should therefore consider new and sophisticated AI-driven approaches that can deliver personalisation even when few customer data points are available.

Much room for improvement

The challenges do not provide an excuse for brands not to try to improve their search experience, and Polonioli underlined the point by drawing attention to recent research. Baymard Institute figures show just 29% of on-site search engines are able to handle even a single misspelt character in a query and 61% offer no support for synonyms. While 82% of sites deserve praise for offering type-ahead query suggestions, 36% offer suggestions that are not helpful and so do more harm than good. In addition, only 40% of websites allow customers to refine their searches using faceted navigation.

Given the huge importance of search in delivering helpful customer experiences and driving up conversion and retention rates, Polonioli made clear his belief that too many retailers are missing out on some simple steps to serve customer better as well as drive up revenue.

Join Coveo and Econsultancy for our next webinar, 25th February, 3pm GMT – Implementing AI and relevance: challenges and opportunities. Register here.