The algorithms powering loyalty: How an AI firm is helping Tesco gamify Clubcard

Tesco’s Clubcard loyalty programme will turn 30 next year.

First launched in 1995, Tesco’s scheme was a pioneer in the realm of grocery loyalty, and today is regularly cited as a stand-out example of loyalty growth and effectiveness. It’s fair to say that Tesco’s member-oriented Clubcard pricing has driven a huge shift in the way that grocery retailers approach loyalty, with numerous other grocery players – from Sainsburys to Co-op – following in Tesco’s footsteps with member-exclusive pricing.

Now, Tesco has announced another evolution of its loyalty scheme: Clubcard Challenges, which will see individually personalised points challenges rolled out to 3 million Clubcard holders, offering the opportunity to earn up to £50 in Clubcard points.

“[It’s a] value proposition for customers that is unique,” says Cédric Chéreau, Managing Director of EagleAI, an AI-based retail personalisation solution by loyalty specialist Eagle Eye. Tesco has partnered with EagleAI to launch Clubcard Challenges.

I spoke to Chéreau along with Sarah Jarvis, Communications and Propositions Director at Eagle Eye, about the technology that makes individual loyalty personalisation possible, how to measure success with a scheme like Clubcard Challenges, and how to communicate a new loyalty proposition.

“A game-changer for personalised offers”

“Loyalty leaders … have value propositions that are very tailored to their business and the behaviour that they want their customers to deliver,” Jarvis says.

“Grocery is an amazing space for loyalty because you can see so much of the customer – you can understand how to make their lives a little bit easier, a little bit cheaper, by giving them a personalised experience.”

Clubcard Challenges, which will launch on 20^th May and run for 6 weeks, will present participating users with a set of 20 individualised challenges encouraging them to spend more on products and brands that are suited to them. Tesco gives the examples of ‘Spend £20 on our summer BBQ range over the next 6 weeks’ or ‘Spend £10 on plant-based meals’.

Successful challenge completion will earn users up to £50 worth of Clubcard points, and Tesco has promised that the value will double to £100 if the points are spent with more than 100 brands who are the scheme’s reward partners.

a notoriously tricky element of individualised loyalty offers… is that they can be difficult to communicate in a universal way

“One of the secret ingredients of the Challenges is that the value proposition is a mass value proposition,” Chéreau explains. This gets around a notoriously tricky element of individualised loyalty offers, which is that they can be difficult to communicate in a universal way. The base concept of Clubcard Challenges, by contrast, doesn’t change.

“That is something very strong – and that’s a game-changer for personalised offers. The way to get [to 50,000 Clubcard points] is completely unique for each customer – but the value proposition is the same for everyone … [which is] very, very important from a marketing perspective.”

“The massive shift to member pricing in the market [was] led by Clubcard Prices – which is a mass marketing message,” says Jarvis. “But if I walk into my local supermarket and it’s half price on nappies through [member] pricing … at one time that was great for me. Now … I no longer have a child in nappies – praise be to God – so all of that messaging … if it’s still on nappies – I see zero value.

“Whereas through [gamified personal promotions] we all get the same message … but the way to get there is totally relevant to me.”

‘The people pleaser’ and ‘the diviner’: the data science that personalises promotions

Without going too deep into the technical side of things, Chéreau explained how EagleAI works to create individually personalised promotions that power loyalty schemes like Clubcard Challenges.

“We built different kinds of algorithms, and each algorithm has a specific mission – it’s designed for something very specific,” he says. “For example, [one] algorithm we call ‘the people pleaser’.

“The purpose of this algorithm is to identify the products or the brands or the categories that are really relevant to each customer – indexed versus the average. … What are the products that [each customer is] buying more than anybody else?

“For example, I have a very special shredded cheese that I’m buying on a very regular basis – so if you really want to please me, that’s the product that you should offer me.

“We have another one that is called ‘the diviner’ that is designed to identify products that I’m not buying, but I should be interested in those products – because very similar customers are buying those products.”

In a way, these types of offers seem like they should be table stakes for loyalty programmes – offers that use the knowledge a retailer has about a customer to discount and suggest only the most relevant products. However, while the principle might be straightforward enough, putting it into practice is much more complex – at least without AI.

“I worked for about eight years in a data science business in the loyalty space,” recalls Jarvis. “With the advance of AI, things have changed quite dramatically – we used to deploy teams of very smart data scientists … and they would, by hand, select the offers for customers to try and maximise ROI for the retailer.

“Using humans, you can only go so far … what [EagleAI] have managed to do is take it out of the human skillset and put it into these algorithms that are maximising hundreds of different decision points – all at the same time – [in a way that] the human brain, even with great tools, could never manage.

“You can … optimise the products they’re going to like … based on all the other customers who shop like them, based on the cadence of how often they’re visiting the store, based on how price-sensitive they are … There’s a lot of brainpower that’s now been outsourced to machines so that [this personalisation can be accomplished] much better, but also much more quickly and easily, than you could have done a few years ago.”

How to measure success

While it’s too early to say exactly what success will look like with Clubcard Challenges, Chéreau explained the metrics that have been key to measuring success with EagleAI’s other retail clients. One is customer engagement: “Are customers participating in challenges?”

This is a key indicator in combination with the second metric: ROI, or incremental retail sales. “It’s the mass of incremental sales that is important, because for retailers and brands, what is important is to see more trucks of products going out of the inventory,” Chéreau says.

Based on past initiatives with other retailers, he states that the scheme can generate £7 of incremental sales for every £1 that is distributed as a reward.

This sounds like a solid return – but Chéreau notes that the ROI could be higher. However, this ultimately comes at the cost of engagement, as some customers will drop away if challenge targets become too difficult. “7 is the best level of ROI … it’s asking the customer to do a little bit more at Tesco than what they would spend naturally. Instead of £10, I might be incentivised to spend £15 – and that’s not £5 more, it’s £5 that I would have spent at another retailer.

“In the past, we managed to achieve much better ROI – 12, 15, 20 – but it means that the challenges are more difficult to win. … If you do this, you lose customer engagement. ROI alone is interesting, but … what we’re trying to achieve is the combination of customer engagement and profitability. It’s a balancing act.

“We need to have something that is reaching most, if not all, customers, and we need that initiative to be profitable.”

Balancing innovation with simplicity

No matter how innovative or effective, is there a risk that introducing new loyalty initiatives like Challenges can over-complicate the proposition behind a loyalty scheme, putting customers off instead of enticing them to interact?

“It’s all about explaining the challenges in the right way,” says Chéreau. “The proposition is quite easy to get – the customer feedback we’ve had on most of our campaigns is that ‘Challenges are easy’, ‘Challenges are very generous’. We’ve never had something like, ‘This is hard to understand’, or, ‘It took me a little bit of time to get it’.

“What you’re saying is – we’re adding more and more stuff to a loyalty programme, and in the end, it looks like a very strange beast … but what we’ve seen in the past is that Challenges are a great add-on, first, and then a replacement for other stuff.

“…The main advantage for retailers is that they can … [later] remove other things from the loyalty programme that are less effective.”

Jarvis adds, “In the same way that Tesco, Sainsbury’s, Morrisons all came out with member pricing – that wasn’t part of their scheme a few years ago.

“There hasn’t been confusion from the consumer base, because it’s positioned as … another way to earn more.”