Personalization is at the heart of strong customer engagement. It helps to build the shopping experience around each consumer, making products seem more relevant and attractive, helping them find what they want quicker, and creating a positive relationship between store and shopper. These increasingly individual interactions, along with the speed, ease, and convenience of multichannel digital commerce, is underpinning the success of forward-thinking online retailers.
Various forms of personalization have been used in ecommerce for a number of years now. Simple examples include sending out marketing emails with a person’s name in the subject line (“Hey Sarah, check out our latest offers!”), or recommending products in those emails based on items you’ve bought before. On-site personalization has also developed, for example, with logged-in customers seeing product collections based on their purchase history, or the purchase history of other people who have bought the same items.
But these ‘catch-all’ tactics still involve a level of homogenization, whereby users are often grouped together and given the same recommendations. The evolution of artificial intelligence (AI) has helped retailers and marketers take it to another level, moving more towards ‘true personalization’ that delivers a unique shopping experience for each customer. And there is significant value in doing so, with studies showing that nearly half of shoppers will spend more if their experience is personalized.
AI allows retailers to use intelligent algorithms that tap into the vast amounts of data they have at their disposal, building up accurate pictures of each customer, moving beyond broad-brush personas.
Truly individual product recommendations
How often have you opened an email that says something along the lines of ‘Top new products for you!’, only to end up thinking: ‘they don’t know me at all’? It’s happened to us all, ‘personalized’ emails suggesting the same products to huge swathes of people. But the use of machine learning that learns more and more about each customer’s circumstances and shopping habits is changing that.
Marketing messages can be made more relevant than ever on an individual level, by unifying the data available about each customer across various touchpoints. For example, platforms such as Moveable Ink deliver dynamic email content that changes in the moment based on a variety of factors, including CRM data, as well as contextual information such as the weather in the user’s location at the very time they open it. This sort of in-the-moment personalization can help drive customer actions such as booking a holiday, shopping for a new raincoat, or heading out to buy an ice cream.
AI is also helping bring live and in-the-moment personalized merchandising to online stores too. Amazon makes around a third of its revenue from product recommendations, using evolutionary algorithms to help identify a customer’s tastes and deliver increasingly more relevant suggestions for what they might need or want purchase. In 2016 the ecommerce giant released the artificial intelligence platform powering its recommendation engine as open source software, taking deep learning to a wider tech audience.
Deep learning allows retailers to build a profile of what their visitors are looking for, with every click, swipe and interaction contributing to this picture and helping to present relevant products that are more likely to drive a purchase. AI applications can also gather data from previous visits, purchase history and demographic signals from other data, such as logged-in social profiles.
For ecommerce sites, this can be used to present a personalized homepage or product catalog for every visitor. For customers, it’s the equivalent of walking into a real life store where every shelf and window display has been arranged just for you.
Improving product discovery through visual search
Traditionally, ecommerce sites have relied on site search, categories, tags, and menus for customers to find what they want. But AI tools are helping to replicate the way most people would look for something in a brick and mortar store – using their eyes.
Pinterest started to bring AI-powered visual search to a mass audience in 2015, using deep learning to allow users to zoom in on specific objects within a pin and then find other items, patterns or colors in visually similar pins. This technique was a progression from the earlier methods of pin discovery, such as ‘Guided Search’ and ‘Related Pins’, which, similar to many ecommerce platforms, relied on image descriptions or categories to connect related items.
Ecommerce sites have begun to follow suit. For example, eyewear retailer Sunglass Hut uses machine learning technology to power its interactive product curation tool. It analyzes the vectors in product images and matches these to other similar looking products, digging deep into the entire inventory to match items by how they look, rather than how they are described. Shoppers choose their favorite frames from a selection and are presented with more items that appeal to their tastes, offering a more intuitive way for customers to find the products they love.
Shoppable images, wherever you find them
Another emerging method of visual shopping includes making any images ‘shoppable’, whether they have been taken by a consumer or discovered on non-ecommerce sites. Similar to Pinterest’s ability to zoom in on objects within an image, Sentient Aware’s Shop The Look feature allows consumers to identify editorial images they like, for example, a celebrity wearing a particular outfit. It then uses deep learning object detection to pick out the specific items of clothing within that image, presenting the user with a range of similar products available to purchase.
Similar results can be achieved with a shopper’s own photos too. For example, CamFind is an app that uses CloudSight’s AI image recognition API, allowing users to take photos of any object and get a detailed description of what it is, including shopping recommendations.
Using chatbots for personalized customer service
For many consumers, shopping in a brick and mortar store still offers a level of reassurance, as there is usually a trained store assistant available to help you out when you need it.
Alternatives methods of live customer service have been tried out with online stores for years, for example through live chat, phone support, contact us forms, customer forums, or FAQ sections. But these methods all have their limitations, and now the rise of chatbots is helping to replicate ‘real life’ customer interaction at scale, keeping up pace with an online retail world in which an almost unlimited number shoppers from around the world can visit your store at any time of day and night.
Conversational commerce for convenience and speed
Chatbots are also helping to revolutionize online retail with their ability to deliver a personalized shopping experience to the customer, wherever they prefer to have that interaction. Consumers are no longer limited to having to visit an ecommerce site or app to browse for goods, ask for recommendations or make purchases.
Major retailers are now engaging directly with their customers – and taking orders – over instant messaging platforms such as Facebook Messenger and Kik, as well as across various social media channels. Pizza Hut uses chatbots to accept orders over Facebook Messenger or Twitter; beauty brand Sephora offers personalized beauty tips and the ability to purchase products directly through Kik, while clothing store H&M offers a similar service.
Device agnostic shopping in a connected world
Not only does AI help deliver an interactive experience across a customer’s preferred channel, but it also offers a unified and coordinated experience. It can pull through information from various touchpoints, to ensure the interaction is relevant and accurate, no matter what device they are using, helping them to get to their desired result as quickly as possible.
The advent of voice assistants, such as Amazon’s Alexa or Alibaba’s Tmall Genie, are also slotting seamlessly into the ecommerce world, allowing customers to place orders by speaking, without needing to type anything or input payment details. The integrations with other connected devices in the home, such as smart televisions, will only get more sophisticated and streamlined as more retailers embrace the connected home trend.
Advancements in natural language processing (NLP) have allowed chatbots and voice assistants to communicate in a relatable way, choosing the appropriate tone of voice to match the brand and match a customer’s mood or the nature of their inquiry. It can also recognize when the chat needs to be handed over or escalated to a human customer service rep. So bots aren’t completely replacing humans, but instead providing a more efficient service when a customer wants it, with the ability to defer to a human when needed.
It is not just the improved language that makes the AI commerce experience a positive one, but the instant access to the vast amounts of data means the responses are accurate, efficient and fast, serving a customer in a fraction of the time it might take a human to do the same. For example, a bot can search an inventory of hundreds of thousands of products instantly to respond to a query, or communicate instantaneously with other connected platforms such as a delivery provider, to get immediate answers or make arrangements quickly.
Combining the principles of personalization with the speed and efficiency of AI is helping retailers to tap into the always-connected world, delivering fast, easy and compelling ecommerce experiences tailored to each customer, whenever, wherever and however they want to access them.