In retail, however, the reverse is actually true. Going back to a time before big supermarkets and chain stores dominated the shopping landscape, people would typically buy their groceries or their clothes from a small, local business – one that remembered them and greeted them by name, and could cater to their preferences accordingly. A personalised service.

However, today it’s rare to find a retail experience truly tailored to the individual. Amazon, as I mentioned, is one shopping service that does personalisation and recommendations extremely well; but Amazon is a huge entity with access to an immense amount of data. By and large, it’s considered to be the exception rather than the rule.

How can retailers bring the familiarity and personalisation of a traditional in-store experience to online shopping? Beauty brand L’Occitane, working in partnership with personalisation and data specialist Qubit, believes it has found a formula.

The brand recently rolled out an AI-powered personalised experience on its mobile site that recommends products based on a customer’s behaviour on the site. To date, this “social newsfeed” style experience has succeeded in boosting mobile conversions by 159%.

I tried out L’Occitane’s new mobile shopping experience for myself to see how it fared, and spoke to Simon Jaffrey, Product Director at Qubit, about how the company set about replicating a personalised “in-store” experience on mobile, the importance of inspiration, and why it pays for personalisation to be transparent.

Test-driving L’Occitane’s new mobile experience

I will confess to never having shopped with L’Occitane before, so I started my shopping experience with a completely clean slate.

Browsing L’Occitane’s mobile site is a pleasant experience, with a sunny yellow colour scheme that manages to keep from being overpowering. The website has a streamlined look and feel, and is clearly well-optimised for mobile, with big product images and tap targets, intuitive navigation and minimal clutter.

New shoppers aren’t presented with their personalised “feed” right away. They can browse the website as normal, and after a few minutes, a speech bubble next to a little icon in the bottom right corner (which looks a bit like the standard “share” button with three dots and interconnecting lines) will alert them to “New suggestions for you”. This notification then turns into a little red dot until the user opens and views their suggestions.

Without the prompt, I probably wouldn’t have thought to tap this button to access my personalised feed, but it’s easy to get used to once you know where to go.

My personalised recommendation feed was populated by a mix of popular products and items based on my recent viewing history, with the latter making up more of my feed as I browsed. Every few minutes of continuous browsing, the feed would update and give me a “New suggestions” speech bubble to invite me to check out my updated suggestions.

Within the feed, each product has its own individual card (never underestimate the eternal popularity of cards in mobile experiences of all kinds), which shoppers can select to view the product page, share via email or social media, or add to their “Picks” by tapping on the heart icon in the top right corner.

Picks are a useful means of saving items that you particularly like in one place, akin to a list of Favourites or Bookmarks. However, they’re not to be confused with your Wishlist, which is a separate “wanted” list that you can save items to (also represented with a heart icon). Picks can only be added from your feed, while wishlist items can only be added from product pages.

The overlap between these two features is the only other thing I found counter-intuitive about the shopping experience – while I understand that they serve slightly different purposes, I think it would be less confusing to merge them into one, and I also wished that it were possible to save Picks directly from a product page. Instead, you have to first open up your feed, where you’re presented with the option to save the item that you’ve been browsing.

The importance of inspiration on mobile

One particularly neat feature which echoes the experience of shopping in-store is the “Complete your routine” recommendations which appear at the bottom of a product page as you’re browsing, which suggest complimentary products that can be added to your basket at the same time for a full beauty routine.

This is similar to Amazon’s “Frequently bought together” feature, but rather than drawing on purchasing decisions that other people have made, the site makes a recommendation about which products will go together well, in the manner of a sales assistant in a shop helping you to put together your skincare regimen.

Jaffrey explained to me that this type of inspiration – helping shoppers not only buy products, but decide what they want to buy – is at the heart of the shopping experience Qubit created for L’Occitane.

“Mobile really is the inspirational device,” he told me. “People are using it when they’re out and about, when they’re short on time, and so it’s all about giving people ideas for what they want to do with it.

“We’ve always taken a very customer-centric approach to product development – understanding what types of challenges our customers have, and then bringing that to our engineering team and determining what types of solutions we can build to solve those problems.

“Around 2015-2016 we started to notice a change in the way that consumers were coming on to retail websites. All of a sudden, there was an uptick in the number of consumers engaging on mobile web in comparison to desktop. That reached a tipping point towards the end of 2016, and by the middle of 2017, mobile web was the dominant channel – people were spending most of their time visiting retailers on mobile web rather than desktop.

“The question then became: should the experience on mobile be the same as desktop, or do they need to be different?”

Qubit did end up making improvements to L’Occitane’s desktop experience as well as mobile, something that I’ll come back to a bit later. Ultimately, they decided to create a dedicated mobile experience that would cater to the increasingly high expectations that consumers have regarding personalised digital experiences, while playing on the inspirational role of mobile.

“The expectation of the consumer regarding what a good experience should be was changing, and mobile was at the very heart of it,” said Jaffrey. “So the challenge was: How could we give customers coming to the mobile website the most personal and also inspirational experience, so that they would be engaged and more likely to spend time exploring that product catalogue?”

Drawing from bricks and mortar

As I mentioned earlier on, retail is unusual as a sector in that it was highly personalised before digital came along, and now is much less so in the digital age, which is the opposite trajectory to the one most other sectors have taken.

For this reason, Qubit set out to learn directly from bricks and mortar retail when creating its personalised mobile feed, talking to customer assistants on the shop floor in order to learn how to replicate this service in the digital realm.

“We’re fortunate in that our office is based in Covent Garden, and so we had our professional services team go to a number of retailers in the area – including L’Occitane – to get a feel for what that personal shopping experience looks like.

“What they discovered is that there are two parts to providing a personalised experience in retail. First of all, the assistant is trying to gain an understanding of what types of products you might be interested in. They do that by asking about what products you’ve tried in the past, your likes and dislikes.

“We’re lucky enough in that we have access to data on L’Occitane’s website that will tell us what customers have looked at or engaged with or bought in the past. That all goes into a machine learning model that we can simplify to make sure the customer is getting the best recommendation.

“The second part of the equation is the reasoning behind recommending products to customers. So, “You should try this product because…” Because of your skin tone; because this is popular right now; because this is on-trend. We call this “transparent personalisation”.

“You can see a similar thing in how Netflix approaches user design. When you’re watching something on Netflix, their recommendations tend to be accompanied by a short description: “We recommend you watch this film because you’ve just watched this film.

“So when a customer goes into the personalised discovery platform, not only do they see recommendations based on things they’ve already been engaged with, but they’ll be given a description which tells them that this product is being shown to them because it’s popular right now, or because it’s popular in a specific category that they might be interested in.”

Two products being recommended to a shopper under the heading "Customers also liked".

Giving this context for a sales recommendation is something that comes very naturally to human interaction, but takes a lot more work to implement in the digital realm. However, Qubit has the numbers to prove that it’s worthwhile.

“Having that level of transparency increases the likelihood that a customer will interact with a product by around six percent,” said Jaffrey. “And thanks to that additional engagement, you’ll know that the customer is interested in this category, or in products like this one. Which means that the next time the customer interacts with their personal shopping feed, the recommendation gets better next time around.

“This creates a virtuous cycle of continually improving the personalised experience.”

A digital in-store experience

Not only does focusing the mobile experience on inspiration help customers to discover new products and encourage them to engage; Jaffrey explained that it can also help them to navigate a retailer’s mobile website more easily, especially if they don’t already know exactly what they want to buy.

“If you look at the typical mobile website today, you have a hamburger menu and you have a search icon. The hamburger menu works if you know exactly what category you’re looking for, and you can navigate through the categories, though it might take you two or three clicks to get there.

“Search, on the other hand, works well if you know which product you want. But what we’re discovering from the experience with a personalised shopping feed is that often people don’t know what they’re looking for. They know a thematic area that they’re interested in, and that’s where inspiration really plays a role.

“In some ways, the mobile web is the closest thing you can have – from a digital point of view – to an in-store experience. We see this having a knock-on effect on other channels: people will use their mobile for the discovery part of their journey, and then they’ll carry on and complete it elsewhere. We’ve found that of all the transactions that are completed on desktop, 19% of them were influenced on mobile.

“That means that if you can have a positive impact on the mobile experience, it’s going to have a dramatic impact on every other channel.”

Product inspiration and discovery plays an important role on mobile – and may be the closest thing we have to a digital in-store experience.

To date, Qubit has primarily focused on creating an engaging and personalised shopping experience for L’Occitane on mobile, though it has also implemented a number of enhancements to the desktop site, such as improving the recommendations that were presented to customers browsing at night, based on the product lines which tended to be more popular at that time of day.

Another improvement made to the desktop site has been the implementation of transparent personalisation, providing a brief explanation as to why customers were being shown each product, which had the effect of increasing revenue by 2.8% for every visitor who shopped on desktop.

The long-term goal, said Jaffrey, is to integrate different channels to create a truly omnichannel approach, such as using the data gained from online shopping to improve the in-store experience.

Another goal for L’Occitane is to give customers more control over what appears in their recommendations, to ensure that they’re as relevant as can be.

“We want consumers to be able to dictate what their insights are, and change what type of recommendations they’re getting, based on what they really care about. I don’t see why we couldn’t also expand the personalised experience to other channels like email, or even in-store, and we’re building our APIs in that direction so that we can support that type of use case.

“Ultimately, we want personalised shopping to be ubiquitous, so that customers can access it on any channel they want to.”

Econsultancy runs a popular one-day training course in online merchandising, and the Festival of Marketing 2018 (London, Oct 10-11) features a whole stage dedicated to personalisation.