Personalisation has been at the forefront of online customer experience for several years. In order to remain competitive, websites have needed it to drive conversions and maintain customer loyalty. This has been done, mainly, by using data to provide personalised product recommendations and tailored offers.
Now, we are seeing personalisation evolve even more, using contextualisation to deliver customers’ up-to-the-moment needs. Here, we take a closer look at context-based hyper-personalisation.
The issues with traditional personalisation
Traditional personalisation uses personal data, like gender and age, together with historical data, such as browsing and purchase histories, to provide bespoke product recommendations for individual customers. What achieves this is the much-touted product recommendation engine.
While this has proven highly successful for business outcomes and is very popular with customers, it has limitations. It does not consider the up-to-the-moment context in which a consumer visits the website.
You can see the limitations of traditional personalisation in how many websites present products. If you visit a fashion store in winter and look for jackets and coats and then return in summer, the recommendation engine will still show you jackets and coats even though you may be looking for t-shirts and shorts.
This is because it only uses historical data and doesn’t understand the current context of your new visit. It is a limitation that wastes precious screen space and causes friction for shoppers by distracting them from their real intentions. As a result, sales can be lost and loyalty affected.
How context-based hyper-personalisation differs
Context-based hyper-personalisation asks the question ‘What is the visitor looking for right now?’ And to answer that, the algorithm needs to consider much more data and operate in real-time so that the content changes in harmony with the shopper’s current intentions and needs. This can be achieved in a number of ways.
Time and location
Algorithms that understand time and location can offer better results. As with the example above, an algorithm that knows the seasons would be less likely to recommend winterwear during summer. However, even the day of the week and the time of day a site is visited can help engines recommend products more suitable.
So, too, can knowing where someone is when they are visiting. Restaurants and stores with various physical locations, for example, can send personalised offers and display online information about the nearest branch.
Understanding personal taste
If someone buys a ‘punk t-shirt’ or ‘mid-century chair,’ traditional recommendation engines tend to show them more t-shirts and chairs. What these purchases also reveal, however, is the user’s taste. They like punk rock music and mid-century design. If these styles of clothing and furniture are included in product metadata, the algorithm can use them to recommend other products that are in line with personal taste.
Session-based personalisation
When users visit a site, they have an intention in mind and hyper-personalisation aims to understand that intention. To do this well, the site needs to understand the context of the latest information it has – in other words, what is happening in the current session.
If you visited a website yesterday looking for kitchenware for your home and then visit today looking to buy your friend a birthday present, the algorithm needs to react in real-time to that new intention. It has to analyse current browsing patterns to understand that the purpose of this visit is entirely different to yesterday’s and provide content that is relevant.
One technology that can help do this is visual AI as it can understand the visual attributes of the products searched for and so can better identify the products the customer is looking for. It can also assist with understanding the aesthetic preferences of individual customers.
Social context
Tapping into users’ feelings is another important part of hyper-personalisation. Understanding the things that they have empathy for can help provide them not just with more accurate products but with a better user experience.
It shows your brand understands the feelings, values and needs of others. If they have a history of shopping for eco-friendly, vegan or fair trade products, for example, then understanding these attributes and acting on them is equally important as understanding personal taste, as they provide customers with emotional satisfaction.
Joined-up data
Hyper-personalisation means providing customers with a seamless personalised experience across all touchpoints, not just your website. To put this into place, every area of your business, customer service, accounts, physical store, website, marketing, etc., needs to be aware of the customer’s current context.
This can only be achieved by joining up the data held by individual teams. And for this, silos need to be dismantled and data unified and centrally stored so that all elements of the business are on the same page.
Imagine, for example, how a banking customer would feel if they were in discussions with customer service because of financial difficulties while still getting threatening letters from accounts and receiving loan offers from marketing. Joining up the data enables a brand to understand the current context and create an unfragmented response across all touchpoints.
Conclusion
Context-based hyper-personalisation provides customers with even better experiences. It enables product recommendations based on current intentions, which look beyond historical browsing and shopping data; it understands the users’ tastes, values and feelings and ensures all parts of the business are speaking with a consistent voice.
If you are considering using tools like product recommendation engines and visual AI to provide hyper-personalised customer experiences, you will need appropriate hosting capacity and performance. Our recommendation is to use a cloud hosting solution. For more information, visit our Cloud Hosting or Enterprise Solutions pages.
