Hello and welcome once again to The Pinsent Masons Podcast, where we try to keep you abreast of the most important developments in global business law, every second Tuesday.
I’m Matthew Magee and I’m a journalist here at Pinsent Masons, and this week we have the privilege of hearing from someone who pioneered the use of data in large consumer businesses and has some thoughts on the use of AI in retail. Edwina Dunn will tell us why retailers need to get their data in order before they start thinking about trying to benefit from AI.
But first, here is some business law news from around the world:
CJEU ruling reminds banks of transparency requirements in consumer credit agreements
UK Government sets out further sweeping reforms to toughen construction safety rules
And online platforms face child access assessment deadline this April
EU banks must meet their obligations to provide clear and sufficient information to consumers in credit agreements after a recent Court of Justice of the EU ruling said that failing to do so might mean they can't charge interest on loans or fees. The preliminary ruling by the CJEU concerned information included in a credit agreement between a consumer and a bank in Poland. The consumer, through Polish debt collection company Lexitor, claimed that the bank failed to fulfil its obligation to provide clear and transparent information to the consumer. It sought the recovery of the interest and fees paid. The court ruled that creditors must provide clear information about the annual percentage rate of charge and the total cost of credit, as well as set out transparently the reason for and method of any variation of these charges. Simply listing a certain number of circumstances justifying an increase in charges may not be enough to satisfy the obligations under the directive.
The construction industry in the UK can expect tougher regulation and a wide range of further reforms as the UK Government plans to adopt all 58 recommendations stemming from the public inquiry into the Grenfell Tower fire. The government's response to the Grenfell Tower inquiry's final report means that a new single construction regulator will be established to take on functions currently exercised by a variety of bodies, such as the building safety regulator, Office for Product Safety and Standards, and local authorities. The aim is to improve standards in the industry and ensure those responsible for building safety are held to account. The single regulator will be responsible for regulation of construction products, the regulation and oversight of building control and the licencing of contractors to work on higher risk buildings. But the testing and certification of construction products and issuing certificates of compliance of these products will continue to be carried out by conformity assessment bodies.
Online service providers can expect severe fines from a UK regulator if they fail to thoroughly assess the extent to which children can access their services and whether their service already has or is likely to attract a significant number of child users. A compliance deadline under the UKs Online Safety Act is due next month. Last month, Ofcom issued guidance in relation to the access assessments to assist providers of in scope services in assessing whether those services are likely to be accessed by children Ofcom will publish its guidance on child risk assessments and children's safety codes of practise in April. Services must then complete an assessment and will have until July 2025 to comply with their duties under the OSA in relation to content that's likely to be harmful to children.
Not many of us can say we changed an industry but this is exactly what Edwina Dunn did. With her business partner and husband Clive Humby she worked with UK retailer Tesco to develop the club card, the world’s first mass customisation loyalty card.
It truly changed the face of grocery retailing and had a dramatic impact on Tesco’s fortunes. So when she gave an interview a couple of weeks ago saying that AI wouldn’t do most retailers any good because the quality of their data was so poor it is fair to say the retail world listened. I wanted to know more. When she joined me we talked about what good data can do for retailers, what is needed to get good data practices going in a company and what the impact of AI might be for those who do have high quality data. But we started with her analysis of what data means to different kinds of retailer.
Edwina Dunn: I think it depends on the type of retailer. There are some retailers that have much more interesting data than others, so the real knack to data is that it is big, there are frequent transactions and that the transactions are varied and tell a story about consumer lifestyles. And that's less common than we would perhaps anticipate. I'll give you an example. You know, if you're a fashion retailer, it's quite probable that someone will only buy from you once a year or once every two years and if someone buys a coat or a jacket or a tie, you don't really know anything about them. So the data is there, but it's not that interesting. If you are Tesco and you're tracking what people put in their shopping basket every day or every week, and that food gives you an idea of how they live their life or the other things that they buy, that's really powerful and certainly people like Tesco, people like Amazon, who are in high volume, high range mass transactions, their data is interesting and powerful and it's the lifeblood of the organisation. So I think there's a massive difference between the retailers with high frequency and interesting data and those that really on the whole just use data to send out discounts and sale notices and that, I'm afraid, tends to be the vast majority.
Matthew Magee: And of course, if we're thinking about artificial intelligence and its use by retailers, that can't be very useful or sophisticated unless the information they hold is valuable and accurate enough.
Edwina: Well, exactly and I think everybody is sort of closing their eyes and hoping for the fact that the tools that are out there crawling all of the public data is going to give them the answer that has so far eluded them, which is they don't have very good data and it's not very interesting. And so they're hoping that this mass collection, it is going to be the answer for them. But there are a number of watch outs and of course you know the data that's being collected, scraped, crawled is historic data, so it's looking backwards. It's giving us what's happened. It's not predicting what's going to happen or what we are nudging at the moment and I think, again, the idea that the predictive AI is the solution for all of that may just be a little bit too optimistic.
Matthew: AI hype is everywhere. Whether it’s driven by excitement, fear or investor exuberance it is easy to overestimate what AI can do today and maybe even what it will ever be capable of.
What any serious thinker on AI knows, though, is that the systems themselves can only operate well if the information in them is good. Without high quality, relevant, accurate data no amount of AI processing will get you a useful result.
This goes to the heart of the issues Edwina thinks retailers face – that not enough of them have data that is good enough. But it goes further than that, she says, it also comes down to how organisations fundamentally think about and use data.
Edwina: I think the difference when data becomes a strategic tool is so fundamental it's hard to explain to those that don't do that. I think historically data has been used by most organisations to back up a plan that's already been hatched. There's a kind of sense that market research is there often to reinforce an idea, a product development or even just my gut instinct as a leader of 20-30 years.
The companies that are using data who are really using it deeply across all elements of planning down to which products they stock, which new products they hunt for, where they stock them, what kind of store format, even introducing new elements to their brand, the way that they do it, they will say data tells us what we don't yet know, and that is something that I think divides organisations at this point, because we often pride ourselves on knowing the questions to ask, but in reality we don't know what we don't know. And so, a really true data protagonist, someone who puts it at the heart of their organisation will often be surprised that results are counterintuitive, or reveal things that they hadn't yet thought of or understood and that is where the advantage lies.
Matthew: London-based retail expert Florian Traub of Pinsent Masons emphasised that of course companies need to gather data from consumers in line with data protection law and should store it securely, make sure it's accurate. But they need to do more to be careful to act in line with consumer expectations.
Florian Traub: Obviously, when you collect data, you have certain obligations to disclose to the consumer how the data is going to be used, where it is stored for what purposes is it is going to be used and you need to remain within the frame of that, not only from a legal perspective but also from a perspective of actually protecting the trust that the consumer places into the brand and into the retailer business. In terms of making sure that consumers don't really see that data used in a way that they don't expect to be seen. So for example, they are, consumers are particularly sensitive about is if the data is then not only used for commercial purposes of the data collector of the retailer, but which is then sold on to other providers and then used in a way that you would not be expected the data to be used.
Matthew: So what are the potential rewards here? If a retailer can gather good data, then store it and use it properly, what are the returns on that investment? Well it doesn’t even have to be all about AI said Edwina, thinking back to the impact Tesco’s Clubcard had on the UK grocery market. It's hard to overstate the impact it had. It wasn't just a combination of membership scheme and a discount card that had existed before it was personalised to each customer and sought to learn from their behaviour and it transformed Tesco's fortunes.
Edwina: When club card was launched in 1995, when it went national, Tesco were number two, way behind Sainsbury's in the market and you know, they were nearly third in the market and Clubcard was launched, it was received brilliantly well right from the outset, there was huge marketing effort, huge sign up of the card behind it. In less than three years, Tesco had doubled its market share, doubled and so they overtook Sainsbury's and were way in the lead within three years and this was simply by knowing which products a customer was buying. So previously retailers and this is still the case for many today, most retailers know how much of a product they sold. So, fashion retelling, they'd know how many size 12 dresses of blue particular line they'd sold. What Tesco knew was that Mrs. Brown had brought £200 worth of clothing or product in a year and Mrs Green hadn't bought anything that year. That was the difference and so suddenly you could tell that some customers behave very differently to others and that you could know them and you could know how often they shopped, how much they spent, whether they liked fresh food or frozen food, convenience food, baby food. Whether they drank alcohol, whether they ate meat, whether they use the fish counter, you would know all of those things suddenly. And that was the radical difference.
You could then take that data and understand were there particular shopping habits based on particular stores. So then you start relating it to the geography of stores and this of course was before we massively went online. So you'd begin to see that different stores provided metro solutions versus out of town solutions. And in fact, the data was very powerful in setting up online business, because what we were able to do is actually say this is your shopping in store, let's provide you that shopping list so that when you come online we help you and pre populate what you buy online with what you have bought in store. So it kind of made it easier and faster. And you know, lots of things have progressed since then. But those are the kind of simple steps.
There's one other that's really powerful, which is at the time we were working with Tesco towards the beginning, Marks and Spencer’s were the heroes of the High Street because they sold expensive ready meals that people loved and they were the treat for the weekend or when you were entertaining. They were the cheat meal for entertaining. Tesco studied this and then launched Tesco finest and Tesco finest was launched using club card data and it allowed them us to see who was spending a little bit more in luxury type products, who was interested in grading up for a special occasion, and that's what data helps you do, and you end up doing it better because you eliminate all the waste, all the failed learning and it just leads you to the right place quickly and really effectively.
In fact, when they launched the bank Tesco Bank their acquisition rate using data from transactions meant their acquisition rate was, I think was five times less than traditional banks acquisition rate.
Matthew: So the rewards can be enormous for retailers of getting on top of the data and using it to spot geographical trends, manage product stock and placement and even identify product gaps and move into whole new areas of business. So if you're a big retailer and you do manage to acquire good data and handle it properly, how would AI itself work with that information?
Edwina: I think some of the obvious things about AI is that it can do massive processing of multiple data points, incredibly fast, systematically and that's a huge advantage. Being able to analyse what has and hasn't worked in the past immediately and across all learnings is incredibly powerful and I think that's what AI is completely brilliant at. But I have a horror that there are going to be some, some rather bad rollouts. We're going to have all our data. We're going to segment it, but we're only going to segment it by who buys at 10% discount who buys at 30% discount and who waits until 50%. The other interesting things are who are fashion leaders who buy early on in the season, who are trendsetters, who buy collections versus individual items. That often gets forgotten and the discounting becomes the driving theme and the kind of slightly lazy theme of let's just shift stock quickly.
Matthew: So, if they use data better to trace those people's particular needs, they could discount less and profit higher.
Edwina: Yes, yes and I've had conversations like that with many retailers and they just don't have the strength in their I suppose they would call it their CRM department to achieve that. And so they do what they've always done and it remains sale discounting and then it's the rush, the race to the lowest price which takes all the margin out of the business and leaves them with less and less to invest on smart approaches and smart positioning.
Matthew: Florian of Pinsent Masons says that if your data set is good, there are clear benefits to combining this with powerful AI systems.
Florian: Think they're probably still at a relatively early stage when it comes to the use of AI in terms of data collection for retailers and consumer facing businesses. I can see at least two areas where AI can play a role. So the first area is around the data collection, the cleaning, the analysis, the whole process around that can be more automated and can be made more efficient, but it also then ensures that the quality of the data outputs of the guidance to businesses is better and then has the competitive advantage that businesses are looking for. But the other area is really to be more predictive. AI can predict trends and customer behaviour and then helping really businesses to be a little bit more ahead in terms of their marketing approach and their strategies in terms of promoting certain product and service lines.
Matthew: We are, it seems, some way from a point where lots of retailers are completely on top of how they gather and use data and how this will be able to feed their AI systems.
So I asked Edwina what steps they need to be taking now to prepare themselves to take greater advantage of data and technology, and she was pretty blunt: the entire organisation needs to be fundamentally oriented to doing this as a top priority. This isn’t small change we’re talking about.
Edwina: I think it goes back to a quite an old fashioned concept which is what is it that the customer wants from them and a sort of going back to the world of you know, why do my customers buy from me and what can I give them that would make their life better, easier, more exciting, more functional. whatever it is that they're trying to achieve. And I think the answer isn't always technology. And I think there is a real danger that AI is just another layer of dispassionate financial ambition that forgets that at the end of the loop is the customer who just wants a nice experience and to be valued and managing data and getting the best out of it is not easy and it needs real commitment from the top down. It's not something that you outsource to someone in middle to junior rankings within your organisation where you as the board director have no appreciation of what they are wrangling, and that happens a lot that that the boards that that have an appreciation of their data and an understanding of it are wildly ahead of other organisations and I think for me the biggest discriminator between success or not is something that someone said to me recently. I think they captured it brilliantly. Was a head a lead data scientist and she said to me, how do I ask a better question? And that to me summarises everything that’s important, which is the best question gets the most powerful answer. And asking a great question can open up an avenue of growth and originality that just wasn't there before, so you can have a hundred data scientists and they can be very busy analysing and collecting data and answering none of the important questions. Or you can have one really good senior lead data scientist who asks a brilliant question and then goes and gets the data that they need and that is what will make the biggest difference.
AI is an exciting era and you know I've always been part of the technology explosion, but I've always found that the winners are the people who never forget that their success is in the hands of each and every consumer that they support.
Well, that's it for this week. I hope you found that useful. Thank you so much for spending time with us. I know lots of people are competing for your attention and we really appreciate it. Don't forget, you don't have to wait another two weeks to hear from us. You can get daily business law news and analysis updates at pinsentmasons.com from our dedicated team of journalists. Or you could go to pinsentmasons.com/newsletter and sign up for a weekly e-mail tailored exactly to your interests. But please do stick with us here, tell your friends and colleagues about it, review us, rate us, help us spread the word and hope to catch you next time. That's all for now. Thank you and goodbye. The Pinsent Mason’s Podcast was produced and presented by Matthew Magee for International professional services firm Pinsent Masons.
