Podcast | Retail Media at dunnhumby

With Neeraj Gehani, David Clements,
10 January '22

Dave Clements (00:00): Hello everyone. And welcome to our next episode of the dunnhumby Customer First Podcast. Today’s episode is all about retail media and how retailers are using data and science to make it more attractive to advertisers and more impactful for customers. My name’s Dave Clements. I’m the Retail Director at dunnhumby. And today I’m delighted to be joined by Neeraj Gehani, our Product Director for Media at dunnhumby, covering measurement insight and data science. Welcome, Neeraj.

Neeraj Gehani (00:46): Hi, Dave. Glad to be here.

Dave Clements (00:49): Great. Well, look, I’m really excited we’re talking about retail media today because it is in the retail news a lot at the moment with the likes of Walmart, Tesco, Target, Kroger, all building successful retail media capabilities and businesses. Now, we always start any conversations at dunnhumby with retailers about how we start with the data. So Neeraj, let’s kick off. And how do you go about building the right data asset to power retail media?


EVERYTHING starts with the data

Neeraj Gehani (01:22): That’s a great question, Dave. And I think one of the promise of retail media is to leverage customer level data to target customers across channels and measure the impact. You cannot do this without having a connected customer level identifier to unify a customer data across all channels. Now, it’s easier said than done. And we work with retailers where the data on the website, the loyalty card, the email campaigns, all of that is sitting in silos. But how do you bring all of them together to build a single clean, unique identifier is the key. And once you have that, that’s a prerequisite for building audiences and doing personalisation at scale. So yeah, everything starts with the data and a unique customer identifier in a team format is the starting point.

Dave Clements (02:11): Yeah. And things have certainly moved on significantly from just identifying customers from loyalty cards. What other ways are we starting to connect the dots with unique customer identifiers?

Neeraj Gehani (02:25): So if you think about the different kind of data assets, which a retailer are sitting on, they could have a customer data via loyalty card. They could have different login details via app or web logins on their website. So how do you bring all of them together is the key. But retailers, what they’re doing is they’re trying to bring all of them together to build a single unique, clean identifier, which they can then use for personalization purposes.

Dave Clements (02:52): Yeah. And we’ve talked before about when all that data is connected together, how that can really help the advertiser. And we’ve talked together how closed loop reporting being a really big advantage. Can you explain a little bit to the audience today what closed loop is and the advantages of it through retail media?

Neeraj Gehani (03:14): Yeah. Great question. If you think about advertisers, typically in the digital marketing landscape, they have been focused on vanity matters like likes and click through rates, which are great directionally, but they do not definitely show the actual sales impact. So, closing the loop on advertising spent with actual sales data is what it’s all about in terms of closed loop reporting.

Now, if you think about retail media, what it does is the retailers have a lot of access to the propriety first party data. So you are able to track the impact of advertising at different stages of the customer journey, starting from awareness to consideration, to sales conversion, and finally to loyalty. And this breadth end of the data is what CPGs are willing to pay if they’re able to see measurable ROI on their campaigns. Now, one of the beauty of retail media is that it allows you to track the impact of advertising at a fairly granular level. So you could look at the impact at a customer level, at a customer segment level. You can break it down by new versus existing customers. You could also look at the website data to understand which keywords are driving sales conversion. So all of this is very, very rich connected to the final purchase behavior of the customer. And that’s what the closed loop reporting is all about.

Dave Clements (04:37): Yeah. So I can really see how all of that connectedness and closing the loop can really help that retail media measurement. And I guess, give advertisers some real clarity of how effective all of that advertising is. Is it also showing that retail media is more impactful from say, some of the other traditional media channels that traditionally consumer goods companies have potentially advertised with. Is retail media proving to be much more effective in your mind? And why is that?


Retail media in real life

Neeraj Gehani (05:14): Yeah, I think beyond the closed loop reporting, which we spoke about before, it’s also important to understand the customer mindset. So if you look at traditional advertising channels, a customer is a passive receiver of the ad. They’re not in a purchase mindset necessarily. But with retail media, imagine you are sitting at home, you’re browsing the retailers’ app or the website sitting at home on a sofa. And then, you are already in a purchase mindset and that gets magnified when you get to the store, because you’re actually at the point of purchase. So that’s where retail media becomes much more powerful compared to other kinds of media. Now, if you think about all the possible channels to which a customer could be contacted in retail media, there is a whole variety of things that you could do.

You could reach out to customer via direct mail or an email. You could do instore point of sale advertising or digital screens within the store. You could also look at retailers website or app where you can do sponsored search, targeted placements, banners, etc. So given the richness of the number of channels through which you could contact the customer, if you put this in context, the customer journeys are increasingly becoming very, very complex and customer needs multiple reinforcements before they take the decision. So that’s where the power of multiple channels and a customer being in a purchase mindset come together to make retail media a very, very powerful platform for advertising.

Dave Clements (06:44): Yeah. So you’re really connecting with customers at all stages of the journey. When they’re in that intent mode and thinking at home about what they might want to purchase, whether they’re on the move, whether they’re actually in the physical store, you’re really connect with them with different channels. So you can run the appropriate campaign for a brand throughout all of those different touch points to really make the most impact. I can see how that’ll be really valuable and how that retail media collection of channels, that omnichannel offering of campaigns can be really powerful. Have you got any examples you can share of a retail media campaign, a brand or an advertiser’s done that sort of brings that to life a little bit more?

Neeraj Gehani (07:30): Yeah, definitely. And in our experience at dunnhumby, we’ve realised that omnichannel campaigns, given they cover all stages of a customer contact, they’re very, very powerful. So a recent example was where we were working with a global dairy products brand in the UK. They were trying to increase the sales for their butter brand. And they ran an omnichannel campaign, which spread across instore, onsite and social media channels on Facebook and Instagram. So the campaign ran for around period of four weeks where we touched around six million customers, who were exposed to around 60 million impressions across different channels over the period of the campaign. The campaign was really, really successful in terms of meeting their objective of increasing sales. So the return on ad spend of that campaign was around eight to one. So for every $1 you invested the campaign, you got $8 back. And the campaign also brought 90,000 new customers to the brand. So it was very, very successful and brings to life the power of an omnichannel campaign than using retail media.

Dave Clements (08:40): Yeah. That’s a great example. And you started talking about how you start to target the right customers there and find the right audiences. And I’m sure that discipline’s developed quite a lot from early days of, customers are in a certain segment, or whatever that might be. Give us a flavour of the sophistication of some of the science involved in targeting and audience selection these days?


Sophisticated science

Neeraj Gehani (09:11): So let’s start with the customer context, right? Because in the world where we are living now, personalisation is table stakes, right? It’s not optional. Customers almost always expect it. And it’s not always about the discount. It’s also about the richness of the content and how neatly that fits into the customer decision journey. So specifically on the science perspective, over the last couple of years, we have seen that the market has evolved significantly from basic rule based and segment level personalisation to one-on-one customer level personalisation. And it’s getting to a stage where you can predict using the retail media data, what is the customer going to buy next? And it’s not only about the things they have bought in the past, it’s about what they could potentially buy based on the behaviour of customers similar to them. So the data science which sits behind this is really high quality data science, which predicts what a customer is going to do next. And that’s the level where we are into this world.

Dave Clements (10:20): Yeah. And I can imagine in an industry like grocery or pharmacy industry, where there are so many products and skews in a store, the richness of that data can really help drive that personalisation and those connections between products. What are good substitutes for each other? If you buy these products, would you be interested in other products? What are the next entry categories? All of those things. So I can really see how that can benefit the targeting and really make sure you select the right customers. Because at the end of the day, I guess, when you’re selecting the right customers for your right messages and the more targeted you are, the more relevant the messages and the better results you get. When retailers are running these big campaigns, how are you making sure and how are we making sure that we have the right impact on customers? So what techniques can you employ to make sure that customer experience is right and the measurements of these campaigns is really effective?

Neeraj Gehani (11:23): Yep. That’s a very, very important question on retail media, Dave. So if you think about the short term measurement, so the best in class measurement solutions are always based on the test and control methodology. So what we typically do is before the starting of the campaign, we hold out a group of customers or a group of stores, which are exactly similar to the customers or the stores you’re going to target via retail media and see the difference in customer behavior based on this campaign at the end of the campaign. So there’s a lot of data science which sits behind this to make sure that the test and controls customers and scores are exactly matching and the results which you are getting are significant or not. And we at dunnhumby take this very, very seriously, because there is so many actual pound or dollar investments, which are sitting behind all these campaigns. So it’s a very, very important point.

Now, the other point which you touched upon is the long term impact. So it’s very easy for a retailer to move to a model where they could maximise their revenues by increasing the number of advertisements on the website or in the store, but that’s not exactly customer first. So we know that even if a campaign is showing or a channel is showing small uplift, incremental uplift during the campaign window, how do we prove that whether this channel or the channel combinations are working for the customers over a longer period of time?

So what we have done is we have come up with a unique methodology called dunnhumby Retail Media Index, which looks at the long term impact of retail media. So what we do is we hold out customers who do not get exposed to retail media across channels or across specific channels. And at the end of the 12-month period, we look at their difference compared to the people who were targeted. And the difference in span of customers across these experimental groups relative to control gives us a proxy of whether this is adding value to the customers or not. And that’s very, very important for retailers to consider.

Dave Clements (13:35): Yeah, it’s such an important point that you made that to make sure you get the right short and long term control groups so you can really keep on top of the impact on customers and experience. When I’ve discussed retail media with a lot of clients, one of the concerns is always, am I impacting the experience of customers? How do I make sure I get that right balance between native messages I want customers to get from just personalisation, and how do I mix that correctly with campaigns that advertisers want to run on my site or through my app or in my stores?


The future of retail media

And I think, as you’ve described there, by having the right measurement, both short term and long term measurement, you can keep a really good handle on, what’s the impact on customers? How are they behaving? Is the advertising ever intrusive? Making sure you get that measurement right and those results properly reported. So you can really keep your finger on the pulse of having the right balance between personalisation and advertising campaigns.

So just listening there, some of those questions we’ve discussed around measurement and omnichannel and the different value of retail media, it’s really clear that, that whole area of retail media is evolving fast. And it’s certainly being a lot more data science driven. What’s next for dunnhumby in this space? What are some of the plans around retail media that dunnhumby have?

Neeraj Gehani (15:04): Yeah. This is exciting time to be in the area of retail media. And we are working towards launching the world’s first omnichannel retail media platform over the next few months. What it would allow the CPGs to do is to target multiple channels, including instore or offsite channels. So increasingly, we are going to bring in more and more channels into the retail media platform. And the advertisers can activate campaigns. They can do their booking, billing, forecasting. They can get access to insights and get access to measurement within the same platform. As we may collect more and more data, we would be bringing in multichannel measurement as a part of the platform. And given that we would be sitting on a lot of insights in terms of how the campaigns are performing, we also plan to build benchmarking data at the base of that data.

Dave Clements (16:09): Well, I really look forward to seeing how that evolves. And it strikes me that whilst the science and the targeting is becoming more sophisticated, it’s just as important to have intuitive, simple, self-serve ways that the advertisers and media owners can book and plan and execute their campaign. So I’m sure that’s going to help because, as you’ve described there, if you really want to build a retail media business and successful business, you need to have the right tools. It’s not just about the science, the right tools and platform to make sure it’s really easy for those advertisers and agencies to access and plan those campaigns.

Well, look, thank you, Neeraj, for coming along today. Such an interesting subject, retail media. And I’m sure we’re going to have some more conversations. Hopefully, you can come back after the launch of the retail media platform and let us know how it’s getting on. And maybe we can get some of our clients, whether it’s the retailers and some of the advertisers to talk about how they’re finding the retail media business evolve for them. So thanks very much for coming.

Neeraj Gehani (17:18): Thanks. Dave.

Dave Clements (17:20): And thanks to everyone for listening today. I hope you found the discussion useful. And whether you are an advertiser, agency or retailer in building your own retail media business. So join us again soon for our next Customer First Podcast. Remember if you want to access all of our podcasts on the various subjects that we cover around how customer first can really impact retail, you can follow us on Customer First Radio on Spotify or on our dunnhumby.com website. Thanks very much and see you next time.