Surviving the ‘tinderisation’ of online retail
Love it or hate it, you can’t deny Tinder has redefined modern dating. Users are presented with a dizzying array of potential matches (sometimes thousands) they can swipe through in mere seconds.
This phenomenon isn’t exclusive to those seeking love. The rise of e-commerce has created the same overwhelming amount of choice and competition in retail.
Online retailers are just a click away from a rival and consumers have become used to darting from website to website with a sense that there are only a click away from a better product or deal.
It’s no surprise, then, that the attention span of the modern consumer is just eight seconds, according to research by Microsoft. So fail to provide them with the right content as soon as they land on your website and they’ll move on immediately.
Brands and retailers need to work harder than ever to convert browsers into buyers. Machine learning - a form of AI that spots patterns in data that would go unnoticed by the human eye - provides exactly this, creating opportunities for retailers to provide tailored, relevant experiences.
So how can machine learning be applied to ensure that your potential customers ‘swipe right’ rather than left?
Understanding your customer
All retailers are familiar with customer segmentation – separating your customers by personality or preference is nothing new. Yet the vast majority continue to only scratch the surface, dividing their customers into basic core segments by using limited datasets such as ‘first time vs repeat customer’.
With machine learning, retailers are able to add to and analyse visitor data in volumes and at speeds that would have previously been impossible to achieve, creating a hyper-detailed customer picture and supercharging segmentation. This goes beyond simple demographics and actually drills down into specific customer behaviours.
Vast amounts of customer data can be collected simply by focusing on what they’re searching for on your website. Once combined with behavioural cues and historical data, this can identify more detailed nuances about who the customer is and what is driving them during their path to purchase.
Is the shopper English speaking but living in Spain? What time of day are they most active? Are they ordering products by cost, suggesting that they’re price-sensitive, or taking time to browse items in detail? If they’re abandoning baskets, what information is available to them that forced them to back away from the checkout? What was the last bit of content they read on your blog? All this information needs to be consolidated and analysed across your channels as it happens in order to really understand who is engaging with you and what it is they are looking for.
Serving personalised recommendations
Once you know more about your customer, machine learning will allow you to present them with a personalised service and make smarter recommendations. By serving customers the products they’re most likely to be interested in straight away you increase the chance that they’ll stick around for longer.
This could be as simple as comparing the results to the customer’s purchase history and filtering out those products that they’ve bought previously. Or it could be something more sophisticated, such as establishing that the customer is actually more interested in jazz music than pop music based on what content they have been reading and engaging with and presenting them with the deals on jazz albums instead.
Customer data could also be combined with third party insights, such as a sudden change in weather or a celebrity being spotted wearing a new clothing item, to ensure recommendations are relevant and of the moment.
Of course, a personalised experience also has to drive sales.
You’ve grabbed the customer’s attention by knowing what they want and presenting them with it, but the right nudge is still needed to get them over the line.
Those who are visiting your website for the first time and are showing signs of being in the ‘research phase’, for example, could be directed to relevant recent blog posts or to a bundle of bespoke offers.
‘Social proof’ can reassure buyers further down the funnel with reviews of the product appearing next to the item, letting customers know that the item met the needs of their peers.
Countdown timers during sales are another incredibly effective tool for converting customers who are on the verge of making a booking or who are clearly very motivated by price.
And you can even utilise machine learning to tackle specific challenges during peak shopping periods such as Christmas or Valentine’s Day. Around these times many customers will be interested in how soon they can receive their items, so bringing features such as guaranteed delivery times to those demonstrating ‘gifting’ behaviour can help you stand out from the crowd.
Uncovering new opportunities
Machine learning can potentially deliver the most value when it helps you spot opportunities you didn’t even know were there, identifying detailed patterns in data that would otherwise go unnoticed by a human analyst.
For instance, you may spot that a large number of Chinese customers are browsing items, proceeding to checkout and then abandoning their baskets once shipping costs become clear. Machine learning can calculate the value of these untapped sales, so if you realise that you’re missing out on a potential windfall you can act accordingly.
You may decide to target these customers with specific deals for free shipping as a result – this allows you to convert this previously hidden group without eroding margins by offering the discount to your entire customer base.
Technology has undoubtedly caused consumer attention spans to diminish in recent years. But it can also help retailers to survive the ‘tinderisation’ of online shopping.
Machine learning empowers us to get to know our consumers more intimately and provide a personalised, hyper-relevant experience. By injecting a level of tailored customer service usually associated with bricks-and-mortar shops, today’s online retailers will be able to convert more ‘click-and-run’ browsers into buyers. Those that don’t will get left behind.