The Future of Site Search: 5 Trends for 2018 and Beyond

2017 was a big year for site search. Voice search truly moved into the mainstream, m-commerce continued to eat into traditional e-commerce’s share of the market, and chatbots were everyone’s favourite subject. But what does 2018 have in store? Here are the 5 trends that'll shake up e-commerce this year.

1. Many Site Search Experiences to Remain Sub-Par

It may seem a little curmudgeonly to begin with a negative, but the fact remains that many retailers just aren’t getting e-commerce site search right.

A 2017 report by the Baymard Institute titled ‘The Current State of E-Commerce Search’ makes for astounding reading. Some of the study’s key findings include:

• 18% of websites provide no useful results when the user types just a single character wrong in the product’s name
• 70% of sites require visitors to search using the exact terminology for the product type used by the website
• Searches with symbols and abbreviations are not supported by 60% of e-commerce websites
• Autocomplete suggestions are found on 82% of e-commerce websites, but 36% of them do more harm than good, either by autocompleting before the user has input enough of the search term or by simply defaulting to the most popular suggestion and returning irrelevant results

This failure to cater to user behaviour is damaging for a couple of reasons. Firstly, if a consumer is willing to take the time to type in exactly what they’re looking for from your site, they’re likely to be further down the funnel than any other visitor. So, making it difficult for them to find what they need is tantamount to passing up a cast-iron sales opportunity.

Secondly, it can have very real ramifications for your bottom line, such as decreased conversion (on both desktop and mobile) and lower purchase values for those sales you do manage to convert. In addition, a poor search experience will leave potential customers with the impression your brand doesn’t prioritise customer experience.

There are signs this is improving as retailers grow more aware of the importance of site search, but given the scale of the problem, we’re unlikely to see an industry-wide shift in 2018.

2. M-Commerce to Continue Its Rapid Rise

Despite the problems detailed above, m-commerce is likely to continue its meteoric growth in 2018. According to research by Criteo, Q4 2017 saw a 21% increase in transactions completed on a smartphone from the previous year. What’s more, mobile accounted for half of all European e-commerce transactions during the same period.

For retailers, this means a need for an even greater focus on getting the little things right with their mobile offering. Top of this list should be the quality of your mobile site search.

Now is the time to begin asking questions like: can potential customers find the products most relevant to them quickly and accurately? Does your site allow for human error such as mistyped characters and aspects of human language such as abbreviations when returning results? Does your site structure make the range of products on offer clear, without overloading the user with irrelevant options?

The alternative is to risk becoming increasingly outmoded as the future of search moves away from desktops and onto smartphones and tablets.

3. Machine Learning to Become Commonplace in E-commerce

The use of machine learning in retail has been steadily growing for a few years now, however, for most of that time, it’s been the preserve of early adopters and the sector’s big players (who have deep pockets to match). 2018 could be the year this begins to change, as machine learning becomes more widely obtainable and affordable for businesses outside of the e-commerce giants.

Utilising machine learning offers retailers 2 crucial advantages:

Automated Search Merchandising and Personalisation

Machine learning allows retailers to move away from manual, labour-intensive search and category merchandising and towards a fully automated approach based on users’ behavioural data.

A machine learning algorithm can automatically merchandise, interpret customer behaviour in real-time, and even learn new or misspelt words. This gives customers highly personalised and relevant results way beyond their initial search query, driving conversions and higher value purchases. What’s more, due to the algorithm’s ability to learn as it works, retailers have no need to manually merchandise search and category listings.

Easy to Implement 

Because of its abstract nature and frothy media reports about the ‘cutting edge’ of technology, it can be easy to assume that machine learning is extremely complicated to implement. However, it can be relatively simple; all it really requires is a model of the product catalogue to train the algorithm on before releasing it on your live system.

4. Natural Language Processing

Human communication is complex; our daily speech is a cacophony of idioms, metaphors, neologisms and rhetoric, much of it extremely hard to understand without lived experience. To muddy the waters even further, we often change the way we communicate depending on medium—you probably don’t use the same language when you post on Twitter as you do in an email to a prospective client. 

This variety and complexity aren’t hard for a human to understand—after all, we’re conditioned to it but for a computer, human quirks can be very difficult to process. Which is where natural language processing (NLP) comes in.

NLP sits astride machine learning, statistics and linguistics and aims to help computers decipher the context of written or spoken language as communicated by humans. It’s useful to e-commerce, because if taken to its logical conclusion, it will allow customers to search your platform using naturalistic language: “cheap dressy shoes” for example.

We’re not there yet and a lot of commercially available NLP is still pretty rudimentary; it’s also difficult to know just how far research will develop it. However, given the great leaps made in AI and machine learning over the last few years, we wouldn’t bet against NLP being a huge part of e-commerce in the future.

5. Visual Search

Finally, if 2017 was ‘the year of voice search’, could 2018 be the turn of visual search?

Visual search allows customers to use an image as the query term for a search, rather than the age-old standard of words. The potential this represents for retailers can hardly be overstated; imagine a world in which you have all the contextual information for precisely what your customer is looking for, even an image of the thing itself. Customers using the wrong keywords or struggling to find exactly what they’re looking for would be all but eliminated. This is the future visual search is attempting to usher in.

Offerings like Pinterest Lens, Google Lens, and Bing Visual Search are winning new converts all the time, as well as honing the capabilities of the technology. And, while we’re not quite at the point where the use of visual search is about to become ubiquitous, it’s certainly on the cusp of becoming more widespread.

Visual search represents a potential paradigm shift in the way consumers interact with the world around them, so it almost goes without saying that now is the time for smart retailers to start seriously considering the technology and its implications.

Conclusion 

All in all, 2018 looks set to pick-up where the previous year left off: with m-commerce very much in the ascendency and innovations such as visual search and machine learning really beginning to gain ground. However, to benefit fully from the technology, e-commerce retailers need to do more than just embrace the latest trends, they also need to be open to understanding user behaviour and the parts of their offering that act as roadblocks to great user experience. 

If you're one of those retailers, a great place to start is with your site search and navigation. Our free guide can help your organisation get started and includes a short list of the 8 best tips for designing a frictionless search and navigation experience.