Analysis

GUEST COMMENT The future of ecommerce site search

Ecommerce site search has been evolving dramatically in recent years, at last moving away from a legacy keyword-driven approach, that often yielded inaccurate and confusing results, towards a leaner, more intuitive approach that puts user experience at its heart. Whilst there have been some significant leaps already in this field, it’s likely that we have only just started to scratch the surface, in terms of what ecommerce search can achieve, and what the benefits reaped by merchants could look like. In this article, we take a look at some of the key developments currently coming through into mainstream eCommerce, which are likely to shape the direction of search in the next few years.

NLP and self-learning technologies

Natural language processing (NLP) has already started to make a big impact in the enterprise eCommerce arena, and its takeup is likely to accelerate in the coming year or so, with a number of powerful third-party solutions already offering integrations for all the major eCommerce platforms.

NLP takes search away from the simplistic “is this keyword present in the product title or description” approach, and starts asking “what is the customer really asking for?”. By taking a semantic angle to evaluating search results, the results produced are far more accurate and relevant to the customer’s actual search intent. Factor in the self-learning capabilities of many NLP-based engines, and suddenly eCommerce site search starts to look ver different. When faced with the conclusive proof of how powerful this kind of search strategy is, enterprise merchants have been quick to adopt the technology, and it looks likely that 2017 will see NLP search start to trickle down to the mid-level and SME markets, with lots of smaller merchants already using some of the known third party solutions.

Increased levels of personalisation

Another aspect of ecommerce search that looks set to improve dramatically is in the area of personalisation. It’s logical that the more personalised that search results are to an individual customer, the more likely that customer is to complete a purchase. As such, many of the new third-party search apps have invested heavily in developing best of breed personalisation techniques. Search tools are increasingly focusing on ‘user intent’ when evaluating search results, so that search terms that, at first glance appear to be very similar, could generate very different results, depending on where on the purchasing journey the customer is deemed to be. A customer searching for ‘red pullovers 38inch chest’ for example, knows exactly what they want, and should, accordingly, be served very specific results. Another customer searching simply for ‘pullovers’ may be better served by offering a category-level result, a pullover size guide, and a snapshot of the bestsellers from the pullover category.

As search technologies progress, and more and more data passes through this new wave of search engines, the levels of personalisation that can be achieved are going to rise.

Chat agents and voice activated search

Most of us are already familiar with online chat support. Our banks, mobile phone companies and broadband suppliers, for example, all typically utilise chat support systems to handle routine online enquiries. These can take the form of fully automated ‘bot’ systems, which parse the customer’s query and find the best match from an index of pre-written answers, or they can involve real people, providing real-time human customer support in a text-based format.

It’s likely that some larger retailers could soon start trialling ‘shopping assistants’ – real human beings who will help customers find exactly what they need online. Whilst they could, of course, answer support questions and provide information such as returns policies and delivery options, they could also offer much more of a ‘personal shopper’ role, going through the customer’s exact requirements, likes and dislikes and so on, in order to guide the customer to the most suitable products. In the bricks and mortar retail world, the personal shopper is usually only employed for high-end, luxury goods, but there is no reason why the approach could not be adapted successfully to a broader range of retail environments online. Supported by smart search technology to back up the operator’s own retail experience, this level of personalisation could boost the customer experience and the retailer’s sales volumes, especially if the merchant’s product catalog is readily available to be used for recommendations.

Another option in the personalised search field is for voice-activated search to become the norm. We have already seen voice-activated personal assistants entering our everyday lives via Siri, Cortana and Alexa. Indeed, with Amazon’s Echo/Alexa service, customers can already place orders for Amazon purchases solely using their voice. It follows that retailers and customers would like to see ecommerce stores offering similar functionality, so that customers simply ask the website for what they want, and the relevant results are then displayed, with voice-based options to choose which items to view in detail or place in the shopping basket.

Another big recent development in ecommerce has been the introduction of Facebook chatbots, which provide a great way of engaging with a customer whilst helping them find what they’re looking for. Burberry is a good example of a retail brand who were very early to launch a chat bot (as can be seen below) and several other brands have integrated their product catalog to provide easy access to product recommendations. This is often grouped into conversational commerce, which is a big buzz word in ecommerce currently.

Predictive analytics & data-driven marketing

Merchants who encourage users to perform a search can also benefit from the data this provides them with – with queries providing detail around what users are looking for. As eCommerce stores adopt search tools based on smart technology and machine learning, it will become more and more possible to mine this data to analyze patterns, to model potential scenarios and to inform personalised marketing activity.

By modelling patterns of behaviour, for example, it would be possible to establish exactly what kind of intent a visitor had, and to personalise the visit accordingly. For example, it could be possible to ascertain that the visitor was at the early stages of their purchasing journey and unlikely to buy on that visit. This data can then be passed into other areas such as email marketing – Klevu’s integration with Dotmailer is a very good example of this, which allows for automated personalised recommendations based on a user’s search behaviour. So if a user searched for “Nike trainers” but didn’t make a purchase, product recommendations in email campaigns would automatically be focused on Nike trainers.

Ecommerce site search is getting a lot more important, leaving behind the frustrating days of simple keyword matching, and heading confidently towards a new world of smart systems that ‘understand’ what customers really mean and want. Whilst some of our predictions may be a little further away than others, in terms of mainstream adoption, it’s likely that all of them will see significant takeup within the next year or so.

Further reading
https://baymard.com/blog/ecommerce-search-query-types
http://www.klevu.com/blog/benefits-of-using-natural-language-processing-in-ecommerce-site-search/
https://paulnrogers.com/optimising-magentos-search-function/
http://www.practicalecommerce.com/articles/133067-13-Chatbots-on-Facebook-Messenger-for-Merchants

Paul Rogers is an ecommerce & digital marketing consultant, working primarily with the Magento platform. Paul is very focused on ecommerce technology and has been working in the site search space more recently whilst working with Klevu. You can read more posts by Paul on his ecommerce blog.

Image credits:
  • Paul Rogers
  • Paul Rogers

Leave a Reply

Your email address will not be published. Required fields are marked *