Publish your content on The Drum
Dadi Awards Banner

Webinar: Why it’s time to invite the Cinderella of UX (site search) to the developers’ ball

According to Yext research, 88% of UK adults say site search is important to provide them with accurate information

Site search is broken, how do we fix it?

Brands invest big bucks in user experience (UX), search engine optimisation (SEO) and paid search but continue to overlook one of the most important touchpoints in the customer journey today – their own on-site search experience. And it’s costing brands both money and business.

According to Yext research, 88% of UK adults say site search is important to provide them with accurate and up-to-date information. Surely that would make it a top priority for marketers, right? Wrong. Over half (56%) of customers believe site search provides them with unrelated search results, 38% get frustrated at least once a week and 40% believe businesses aren’t doing enough to improve site search.

A roundtable discussion, hosted in partnership with Yext, brought together a panel of search experts including Jon Buss, managing director EMEA at Yext; Omi Sido, senior technical SEO at Canon Europe; and Stephen Kenwright, co-founder at Rise at Seven, to explore where brands are going wrong with site search and offer advice on how to update their most precious digital asset.

Watch the full roundtable discussion here.

The customer buying journey has changed massively over the past two decades. Search engines have kept pace by providing answers to users’ questions through a range of content, beyond blue links. By contrast, investment in site search has been left behind – all too often, you get no result at all.

“Site search is an underinvested part of the website and often something that’s just bought as a plugin when you’re buying your website or your e-commerce platform,” says Buss. “It isn’t something that has matured in the same way as conventional search. The contrast is now so great that we’re just noticing that this is such a broken experience – for the brand and consumer.”

If a user doesn’t get the answer they are looking for on a brand’s website, where do they go? Back to Google – where they might get the answer from the company they were looking for, click on a link and that company pays for you to come back. Worst case scenario – they get diverted to a competitor’s site.

Throughout the session, the experts cited examples of what good and bad site search looks like, highlighting B&Q, Three Mobile, P&G and Virgin Holidays [You’ll have to watch the full session to find out what falls into what bucket].

Amazon is a prime example in the e-commerce space of setting the customer expectations bar so high. Google has done the exact same for search results – and that comes down to the metadata it holds.

“[Site search] has not got that really obvious cry out for help from a customer that some other areas of the business potentially have,” says Kenwright. “Particularly for enterprise businesses that want to improve the customer experience here, they will go and buy a really great tool but they forget the strategy part that powers that tool, because the search results are only going to be as good as the metadata of the products or the articles, the quality of information that you put into them.”

Blue links are ‘just not good enough any more’

Most businesses will invest in product search, but it’s not solving the specific nuances that come with search today – a search bar that will allow users to search across multiple directories or databases, including FAQs and relevant content related to the search query. That requires machine learning, the power of which is “not being utilised by many search solutions nowadays,” says Sido.

“By using open-source technology that just plugs in, all it does is crawl every page and look for that keyword based upon your search and delivers a blue link,” adds Buss. “But it’s not how search works these days. Not how Alexa works. Not how Siri works. They deliver direct answers and that’s why the experience of getting blue links just isn’t good enough anymore.”

Google is constantly testing through a ‘testimony approach’ – it has a hypothesis, it has spent years of user testing and research to determine what happens when someone makes a search. Having a multi-algorithm approach that presents other products beyond primarily blue links – something that hasn't happened on a brand’s own website – utilises the true depth of analytics.

Most searches start with Google. By working in tandem with search engines and utilising their insights, brands have an opportunity to “close and continue the customer journey through to a conversion on your website” says Buss. “It’s crazy, especially when the content is quite possibly sat somewhere on the website but you’re not harnessing it and putting it into a knowledge graph to answer the question.”

The site search fitness test

To assess their site search fitness, Kenwright outlined a simple 15-question site search audit framework focused around three sets of questions: technology, response and function. The experts then offered a step-by-step plan of what brands should do next, including the need to look at data coming from Google Analytics to make informed decisions and take practical steps to replicate the searches people are posing on Google on their own website.

“Go to Google Analytics to try and understand how many people are on your website, searching for your products, and how many of them actually will end up buying from you,” adds Sido. “If you have a revenue in mind, you can then go to your CEO and ask them for money to improve your site.”

Key to this is measurement, setting a KPI framework and getting your data in order.

“When it becomes a data problem, people start to care, people start to get involved when you can say, we have these products and no information about them, or we have X customers dropping off,” says Kenwright. “Measurement is where this all starts.”

“Why does Google do a good job at delivering answers on search? Because it creates a knowledge graph of data and answers, it understands using natural language processing, what questions are being asked and how to deliver an answer,” says Buss. “Every business can have its own knowledge graph – it already has to some degree – it’s just that they’re in different repositories of data.”

“Find where those repositories of data are and make sure that your site search is connected into them so that it delivers an answer and not just random blue links – because that’s where the problem starts.”

By continuing to use The Drum, I accept the use of cookies as per The Drum's privacy policy