Podcast | Customer First Personalisation at Scale

Dave Clements (00:00): Well, hello everyone. And welcome to our next episode of the dunnhumby Customer First podcast. Today’s episode is all about Customer First Personalisation at Scale, and how retailers are using data, science, and machine learning to drive better experiences for customers across as many digital touchpoints as they are able. My name’s Dave Clements, I’m the retail director at dunnhumby. And today I’m delighted to be joined by Emily Turner, one of our leading consulting directors for Customer Engagement and Media. Welcome Emily.

Emily Turner (00:45): Thank you, Dave. I’m excited to be here.

Dave Clements (00:47): Yeah, and I’m excited. We’re talking about personalisation today because both of us have been working with our clients on this for many years, but it’s an often-used word. So let me start by asking, what does personalisation mean to you? And how do we look at it at dunnhumby as we support our clients?

Personalisation at its core

Emily Turner (01:06): So for me, I think personalisation at its core, is really about an experience or an interaction. And that could be a message, an offer, a recommendation that’s really relevant, tailored, and targeted to an individual customer. And from a dunnhumby perspective, it’s really got to involve customer data signs. So, everything starts with the customers. Utilising that great first party data that retailers have and combining it with that zero-party data. So that’s the contact information so that you can execute on the personalisation that you deliver.

Dave Clements (01:39): I really like the way you’ve expressed that. And personalisation being about relevance, it’s about individuals and it’s also about doing it at scale as well because you can’t just do it, one customer once and then not be able to carry it on. Are there any things you think it’s not that people talk about? Because it’s a buzzword that’s out there quite a lot. People often use it. Do you sometimes see people use the term personalisation when you think they’re not really talking about it in the right way?

Emily Turner (02:05): Yes. I think that’s quite a frequent challenge now and often you see it when people are referring to segments. So certainly that is targeted. It’s specific to a group, but it’s not personalised. That’s still a mass communication to a segment of customers. The other area where we tend to see it is in the use of customer personal information within the communication. So think of that as using it… Your name in an email. That’s definitely personal, but it’s not necessarily personalised because the rest of the content would be absolutely irrelevant.

Dave Clements (02:34): Absolutely. So it’s linking it back to your previous experiences, your likes, your needs, your interests, and not just recognising you as an individual by your name or an item like that.

Emily Turner (02:46): Yeah. So it’s really knowing what’s important to you, what are your needs? How are you shopping? What have you done before and using that great information to then predict what are you likely to be doing? What would be of interest to you?

Dave Clements (02:59): I remember personally, being involved in some of the first personalisation that Tesco did with Clubcard when they launched it back in the mid-90s. I think I’m showing my age there. Things have certainly moved on a lot since then. And especially, really in the last few years with COVID-19 happening. What are some of the biggest evolutions you’ve been seeing out in the marketplace and which clients do you think are doing personalisation well at the moment?

The channel conundrum

Emily Turner (03:22): I think so much has changed in this area. There’s been quite dramatic changes. So, if we think just from a targeting perspective, that’s gone from using rules-based targeting to highly predictive algorithms, focusing on customer propensity in the future. It’s gone from just coupons and offers to making recommendations to customers. You’ve touched on COVID, which we’ve seen, has had a dramatic impact on customers moving to digital channels and the need for retailers to really quickly turn on capability in those channels. And that again has really come down to how can you help customers when they’re now shopping within the e-commerce environment? And I think we’ve also seen, in the past few years, an explosion of channel. So I mentioned digital, but beyond just digital, think of the variety that includes… It’s not just apps, it’s not just about email, there’s messages now. So how can a customer have a personalised experience in those additional channels? And it’s proving incredibly challenging for retailers as I’m sure you see too.

Dave Clements (04:20): Yeah, I know definitely… As you say, there’s been a lot of evolutions and we’re seeing a lot more of it now. All the communication I tend to get from all the different retailers I personally interact with is becoming more and more personalised. But you also see pitfalls of where some people are not being that successful with it and doing it. Who do you think is doing it quite well at the moment?

Emily Turner (04:41): So I think there’s a large number of retailers that are starting to do it well. A lot of these areas come around the personalised offer programs. That a lot of retailers offer to their loyalty or reward members as a fixed benefit. So that’s an area that we have a lot of experience in as dunnhumby. And we see that the retailers that are successful have large offer pools of a thousand potential offers that they could be giving to customers, which allows for great breadth and depth to deliver highly relevant communications to customers. So one example that springs to mind is Meijer, a Midwestern, regional grocery chain in north America. They have their core mailer is called HPO, Hand Picked Offers and it’s really their flagship communication. And a huge amount of effort goes in, across multiple teams at Meijer to deliver this great communication. And customers just love it. Meijer’s received handwritten letters from customers. They’ve had customers taking to social media to talk about how much they love this personalised communication. And they feel that Meijer really understands them as a result of it.

Dave Clements (05:41): And I’m sure some pretty high engagement and redemption rates alongside those letters and those comments as well.

Emily Turner (05:47): Exactly. And so the results come through, not only from a redemption rates, you can see high coupon redemption rates of well over 20%. Participation rates of over 50%. But think also about the customer satisfaction from delivering a personalised communication like that in much improved sentiment and really helps drive engagement with the retailer overall.

Who is getting personalisation right and delivering the best experiences?

Dave Clements (06:08): I was going to call out a couple of clients as well. One, Coop Norge, a client that’s really transformed from originally doing everything from paper flyer and mass promotions, to a truly now completely digital program. That’s delivering personal own recommendations, using some of the similar techniques you just mentioned with Meijer but really driving industry leading redemption and engagement rates. But also Tesco, I’ll have to mention them, because I think they’ve really made big strides, personalising their web experience. Especially, with things like relevant offers, compliments and substitutes. Have you forgotten real time recommendations as people are adding to their baskets? So it’s doing it in the moment, being really relevant and really seamlessly improving that shopping experience online. So a couple there that I’ve really noticed

Emily Turner (06:54): And coming back to that point about the recommenders and digital and everyone going online to shopping e-commerce during the COVID time. Being able to provide the personalised experience when a customer shopping in line has been incredibly important and Tesco has been doing that for a long time. And a lot of the successful retailers we see are also doing that. We see from our RPI study that customers are expecting speed. They’re expecting convenience, they’re expecting rewards. And those overall recommenders are where you provide a personal recommendation, should that be, “Have you forgotten this item? You might also like…” When you provide those to a customer during the e-com shopping experience. It really helps to drive not only bigger baskets, but repeat purchase as well. And again, it just makes that shopping trip much more seamless, much easier for a customer. And particularly, if you’re starting to shop for the first time online.

Dave Clements (07:45): Yeah. As you say, it’s driving a much better experience, but it really is driving big outcomes as well. Higher basket sizes, real growth in sales and engagement of customers as well.

Emily Turner (07:54): Yeah. So “Have you forgotten?” Is one of my favorite recommenders so this comes right at the end of the checkout walk when you’re online and it’s a recommendation for an item that based on the customer data science, we would expect you to buy, and for some reason isn’t pairing you in your basket. So makes that recommendation to the customer before they complete the checkout. It’s such a fantastic proposition because it really feels of benefit to the customer. So there’s no emergency trips to the grocery store because you’ve forgot something. And at least a great result for the retailer because they’re gaining an extra item in that basket.

Dave Clements (08:24): Yeah. And if I’m right, I think retailers using that type of algorithm are seeing around one in four customers actually directly interacting whenever they see that. And that’s a really massive impact on nudging customers and helping customers at the checkout online. So one of the characteristics of these retailers that are doing it well, is they’re doing it at scale I think Emily. And what’s needed, do you think to make that shift from being hand cranked semi-automated to really doing it at scale with multiple touch points and that sense of always on personalisation.

The right science, algorithms and systems

Emily Turner (08:55): At its core… It all starts with the data. You have to have the data, you have to have that data available. And that’s product data, customer data, transactional data and exposure data. So without that, that’s really going to limit your ability to even get started. But I’d say beyond that, having the fundamentals in place, it comes into insights of really being able to understand your customers and do that in a way that’s really efficient for your business. As you said, it’s not running bespoke queries. It’s having algorithms that are going to really help you identify what it is that your customers need and how they’re shopping.

And from that then comes that predictive customer data science. It’s going to enable you to target your customers and identify what is the right message, recommendation or offer for that customer. And supporting all of that, you’re going to need the right tech ecosystem to be able to deliver against it. So as you said, you don’t want to be hand cranking, so you need to make sure that you’re really mindful about which of the partners that you need and which of the channels that you want to focus on delivering. And then you need to be able to get that data back. Once you execute on those campaigns, to be able to do that in a really precise test and control measurement, to see the effectiveness and optimise ongoing.

Dave Clements (09:58): In the old days, building some of those tech and products and ecosystems was quite daunting, I think. But I feel that it’s really changing, that actually you can inject the right science and algorithms into the right systems that you might have the right platform and actually having a customer management platform isn’t as complex as you might think. Would you agree with that?

Emily Turner (10:18): Absolutely. I think it’s much more common now for retailers to have customer management platforms in their ecosystem. The tech is readily available. There’s a wealth of providers in the market for this, that makes it so much easier for retailers today to adopt this. And as you said, not be creating offer by offer, within an old-fashioned legacy tool that they’ve got.

Dave Clements (10:41): And presumably within this, as well as part of these management platforms, measurement and understanding what’s working and the self-learning of the science is really important. Do you want to talk a little bit about how we go about measuring the science and the impact?

Getting the measurement spot on

Emily Turner (10:54): Yeah. So at dunnhumby, our preferred method is test and control. So this is about creating look like customers who look like the customers that you want to expose that personalised communication to. And holding back a group so that you can have a really robust sense of “Did this communication drive it different. Did this change the behavior that I was seeking? Has it delivered the results to my business that I was looking for?” So some of the measures that we look at for these communications could be redemption rate, participation rate, but we like to look as well as sales uplift. “Did it drive repeat purchase? Did it drive category uplift as well?” And the measures that we look at obviously would vary depending on the objective of the campaign and whether it was an offer or a recommendation, but really are all around. “Did this make a difference? How has it changed the customer behavior and how has it really driving my bottom line is a business?

Dave Clements (11:43): Yeah. And all available through the right type of platform. So it’s really easy to access that measurement and assess it and see the recommendations.

Emily Turner (11:51): Exactly. So coming back to one of the things that… We talked about challenges, obviously you need the right exposure data. So you need to be able to… Once you execute and push out to those channels, being able to get that data back from those channels, to know who was exposed? Who interacted? If that’s an email, what’s the click through rate? To be able to measure really accurately and create detail.

Dave Clements (12:10): So last but not least, let’s talk a little bit about the science. We love our data science at dunnhumby. We love developing lots of different algorithms to drive those relevant recommendations. As you were talking about Emily, what are some of the new innovations that have come out recently from dunnhumby or are coming out in this space that’s exciting you?

Future innovations in science

Emily Turner (12:29): Yes. I’ve already told you about how much I love our Have You Forgotten proposition. But some of the newer things that we are working on is science around new products. So CPT spend a lot of money, obviously in developing these new products. It’s expensive to launch them. It’s expensive for retails to launch them. And we know that there is often… Can be quite a high failure rate for these products. And that’s sometimes just as simple as a lack of an awareness with the customer that product’s launched. So we are developing some science to help. Basically, when you launch a new product, identify who is the right cohort of customers to let know that this product is now available in their store or online for them to purchase.

Another science that we have that I think is really exciting is, looking at customers who are showing a propensity to decline to lapse. So this isn’t wait until a customer has stopped shopping with the retailer. We’ve already lost them. Because then it’s so much harder and so much more expensive to win them back. It’s really starting to look at what are some of the identifiers in their behavior that we can see that they’re showing, there is a likelihood decline? That could be, small decreases in their visit pattern that could be decreases in their spend in certain categories. And being able to identify that and develop that churn model, to help a retailer win those customers back and really retain them.

Dave Clements (13:40): Yeah. And one of the favorite ones I’ve been seeing that we’ve had in recent months is, the idea of The Next Best Product. Especially, with a lot of the clients we work with, where they’ve got 30,000 products on their shelves sometimes. How do you really identify with accuracy the most relevant next product in that particular category or that particular mission for each customer and then encouraging them at the right moments and through the right touch points with the right recommendations on those ideas. I think that’s another acquisition science, if you like, that I think will get a lot of traction with clients in the future.

Emily Turner (14:15): Absolutely. I think that’s another really exciting area that we are working on. It’s “Yeah. What is that next best product? Is that a compliment? Is that a similar recommendation for something out of stocks as well.” Really helping to drive acquisition to new brands to new products into additional categories, really is a win-win for retailers and customers.

Dave Clements (14:35): Well, I really look forward to seeing some of these innovations in action, Emily, with our clients and live with their customers. So it’s been fascinating to talk to you today on the topic of personalisation. It’s clearly still evolving a lot and we’ll have to get together soon to hear some of the latest updates. So thank you.

Emily Turner (14:53): Well, thanks for having me.

Dave Clements (14:54): And thanks to everyone for listening today. I hope you found the discussion useful. Whether you are a retailer who’s just starting out on the road of personalisation or are one of the leaders doing personalisation at scale. We’d love to hear your thoughts on this subject and you can always do that by emailing myself, David Clements or Emily Turner at dunnhumby.com. And join us again soon for our next Customer First Podcast. But remember you can access all our podcasts on a variety of different subjects that are impacting retail @customerfirstradio on Spotify or on our dunnhumby.com website.