Celebrating a data revolutionary
When Tesco launched its Clubcard 30 years ago – on 13 February 1995 – loyalty within the retail sector was largely limited to Green Shield Stamps (for those people old enough to remember) and the grocer’s initiative was likened to these very same paper stamps by its sceptical rivals.
All those years back I took a trip down to Ealing in West London after the launch and it turned out to be revelatory for me. It was clearly a ground-breaking organisation and when interviewing Edwina Dunn, co-founder of Dunnhumby, it quickly became apparent that Clubcard was so much more than a replica of the one-dimensional collectible stamps that were simply a margin-eating mechanic to give discounts on future purchases.
Little known independent
At the time of my first encounter Dunnhumby was a little-known independent company even though it was the power behind the Clubcard loyalty programme. Its pioneering data-driven approach not only quickly propelled Tesco to become the UK’s leading grocery business, overtaking the Clubcard naysayers in the process, but it also revolutionised loyalty and personalisation in the retail industry and beyond.
Thirty years on and its impact is still very much being felt, as evidenced by the recent comments from Ken Murphy, CEO of Tesco, that the grocery company could harness AI alongside data from Clubcard in order to prompt recommendations to shoppers about healthier choices and reducing waste.
“I can see it nudging you over time, saying: ‘I’ve noticed over time in your shopping basket that your sodium salt content is 250% of your daily allowance. I would recommend you substitute this, this and this,” he stated at a FT retail conference.
The inside story
It might be an explosive statement but it very much highlights how Dunnhumby has continued to remain relevant. So how did this game-changing data-crunching operation come about? To find out the real inside story I revisited West London to catch up with Dunn who today is no longer involved with Dunnhumby but is still pushing boundaries – this time through her charity The Female Lead.
The origins begin at CACI where Clive Humby ran the UK division of the US-based business and Dunn was recruited as a graduate to help with its work on spatial targeting systems that led to a site location modelling tool. This involved applying CACI’s Acorn (a geo-demographic segmentation of residential neighbourhoods in the UK) with Census data to determine that neighbourhoods were different and so the people in them would purchase different things.
This tool helped retailers determine store locations but Dunn and Humby were convinced there was a lot more potential. “The lightbulb moment came when we considered that if this is good on census data, which is 10 years old, then just imagine what we could learn from customer data. Our idea was to analyse the customer data (from the tills) and build it onto Acorn,” recalls Dunn.
The idea was taken to CACI management but they were uninterested in funding it so in 1989 Dunnhumby was created as an independent operation. It was a problematic start as CACI took the start-up to court, according to Simon Hay, former CEO of Dunnhumby, who agreed to meet me for the first time in many years and recall the early days. Over coffee in a Soho café he tells me the case was won by Dunnhumby after which he jumped ship from CACI to join his old boss Dunn.
Tough sell
He recalls that it was a tough sell because at that time data was very expensive to store and process at that time. Early customers of Dunnhumby included Booker Cash & Carry, Mercury Communications and Lotus Software. But not Tesco: “At the end of 1993 we saw them after they had just created the Clubcard concept and we said we can analyse customers. But they told us to go away.”
The breakthrough came through Grant Harrison in Tesco marketing who heard Humby speak at a conference and reference the tool Dunnhumby had built to analyse big data. Dunn says: “It was fortuitous and we were given some data from Clubcard and told to let him know what we find in it. They’d created the idea of Clubcard in nine stores. We didn’t invent it, we made it work.”
Making it work was a big deal as it gave Tesco genuine insight into its customer’s shopping behaviour. The issue they had was that they were giving a 1% reward to loyalty members and they wanted to know what happened when they did this. Hays recalls them saying: “We’re about to spend £100 million on rewards and we need to know that it brings in new business and that it works.”
Scary findings
Dunn says: “The project was in 1994 for three months and we showed them conclusively that people returned to stores more often and bought more.” After this the real big breakthrough came when they presented their findings to the Tesco board. Then chairman, Lord MacLaurin, declared: “What scares me is that you know more about my customers after three months than I know after 30 years.”
With that ringing endorsement it was all about rolling out Clubcard to the Tesco stores as fast as possible as they wanted first mover advantage. “It was – do it national and do it fast. The risk taking of this is now lost. The commitment to do something so big is lost today. For the 13 February 1995 launch they printed 16 million Clubcards, closed every store the weekend before to dress it up with Clubcard [promotional materials], put people in all the stores to sign-up members and there was national TV advertising,” she recalls.
Such was the strength of the promotional activity that the impact of the launch was “absolutely huge straight off”. Apart from the odd naysayer in the senior management team at Tesco Dunn says there was widespread support and few doubts about the potential impact of the programme with CEO Terry Leahy, McLaurin and marketing director Tim Mason fully behind the project.
Monumental impact
Their conviction was justified because the effect on the Tesco business was monumental as Dunn says in the first year after Clubcard’s launch the business overtook Sainsbury’s and in less than three years the loyalty programme had helped double the grocery market share of Tesco. “It was all incremental business and we calculated that over the next 10 years Clubcard and Dunnhumby made an extra £60 billion of sales for Tesco,” she says.
This incredible success came on the back of what today would be regarded as only very modest amounts of data. “It was just [analysing] some of the data some of the time. We’d get an overnight stream by polling all of the stores’ [transaction data]. We could only analyse groups of products weekly and on 10% of the data. We could look at frozen versus fresh and soft drinks versus carbonated. We could not analyse to the detail of Coke versus Pepsi. It was only modest but some knowledge is an advantage. It had not been done before,” explains Dunn.
The technology was a seriously limiting factor in the early days, according to Hay, who says Dunnhumby was doing work on machines that people did not think was possible. “At that time data was thrown away as it was expensive to store and IT was monolithic. The Tesco team at Dunnhumby was initially three people and this got to 40/50 at the peak,” he says.
Data revolution
As well as the technology Hay suggests the idea of working with data was almost revolutionary. “No one was doing data work. You needed to have the right mindset and to be curious and collaborative. This was the late-90s and they worked with real data. It’s a mindset and discipline today but it was new back then,” he says.
An integral part of Clubcard were the money-off vouchers and promotional offers posted to members. By 1997 these personalised coupons were down to a one-to-one level because of the in-depth segmentation that Dunnhumby undertook. Customers would be in various segments and Venn diagrams across these would determine the specific vouchers and promotions they received
This activity had the involvement of the brand owners who were benefiting greatly from the insights although Dunn says they were mainly obsessed with driving brand switching from their rivals: “Coke wanted us to coupon Pepsi buyers. We said they could only send promotions to their existing customers. If they typically bought six then we’d say buy 12 and get two for free. We had a big fight over this and today it’s still in CPGs (Consumer Packaged Goods) DNA to brand switch.”
Driving strategic change
Along with massively assisting Tesco’s marketing activity the Clubcard data was also instrumental in driving strategic decision-making across the whole business. The analysis of customer data led to the launch of the Finest own-label range that celebrated its 25th anniversary in 2023 and is today worth £2 billion in annual sales. It also helped with determining store locations and prompted the creation of the Metro convenience format as well as Tesco Bank, Tesco Mobile and the online business.
“The assumption was that online would cannibalise store sales but we showed it was incremental sales and that it attracted new customers. Our IP showed it as a growth strategy,” recalls Hay.
As the Dunnhumby business had grown Dunn and Humby had reinvested profits in it (rather than pay themselves) and at the turn of the Millennium they sought to address this with sourcing outside investment. An initial dalliance and possible IPO with a South African media business that was artificially buoyed by the dotcom exuberance at the time did not come off as Dunnhumby turned sour on the arrangement. This near-miss was a lucky break and Tesco then asked if it could be an investor rather than simply being a client. It took a 53% stake in 2001.
Going international
The downside to the deal was the reluctance of Tesco’s rivals to work with Dunnhumby. Among them was Amazon that was a client but then fired them, which Dunn says was not helped by Tesco management having stated that it was going after Amazon’s market share. “Nobody else in the UK would work with us so it was either just work with Tesco or go international,” she says, adding that the latter option was taken.
Helping this move was the strategic decision by Dunnhumby to sell its data (initially just from Tesco) to the CPGs. They had enjoyed using it for the Clubcard promotions but when Dunnhumby created an analytics platform (The Shop) for them to analyse the data themselves – with the customers’ personal details hidden – then this relevant data could actually be sold to them.
This became 80% of the company’s revenues as Dunnhumby tapped into the global reach of the top CPGs in key markets. Only 12 countries represent the majority – in terms of revenues and profits – of the market share in products. It was an objective to obtain their data. This desire came alongside the recognition that CPG companied are global players while retailers are local.
Major US breakthrough
A major early win in this international move was the signing of a joint-venture with US grocery giant Kroger although Dunnhumby was unaware of its scale at the time. “In 2003 they came hunting for us and kept saying they were interested as we were the only people doing this [data analysis] and they wanted to ‘do a Tesco’. We were saying who is Kroger?”
A powerful piece of the Kroger deal was that they invited on-board some big CPG players such as Kraft, General Foods and Coke who very much liked the fact that the Dunnhumby data was reflecting their activities in numerous international markets. They liked the unique global insights it could provide. Dunn says things really kicked off when the company had retail clients in all 12 key markets supplying customer data and it could charge serious money for The Shop.
Heading up the operation in the US with Kroger was Hay who suggests: “We went to the US to show it was not a fluke. And we also knew what we hadn’t done at Tesco, which included change management. We could do these things with Kroger. People are the biggest barriers to change and at Kroger we set the strategy and bonuses for all employees including the CEO based on outcomes from the data. There was a fearlessness on our part.”
Improving technology
Over time as the technology improved the move was made to crunching 100% of the data and the frequency of the reports was increased to every day. “Marketing could be long-term but which sales were hurt by competitors’ pricing meant there was a need for daily analysis. The tech galloped ahead. We could also deal with all the data down to the EAN (European Article Number) for each individual product.”
Tesco went on to increase its stake in Dunnhumby to 84% in 2006 and in 2011 the Dunnhumby founders stepped back. This combination led to a fall-off in the disruptive nature of the business. Hay became CEO of the overall organisation and recognises there was great value in Tesco owning the business but acknowledges “it’s harder to be a disruptor when you are in the family”.
Dunn is a little more forthright in describing way things have panned out. “It lost its voice when it was absorbed into Tesco. Sometimes it is good to have a counter voice. Clive and I owned part of the equity and were treated with respect and I warned Simon when he becomes paid help then it’s harder. You are just a division of the company. Dunnhumby was about disrupting and innovating. What it’s doing now is absolutely not. I wish it had not been absorbed but been kept independent. It makes me sad that it’s just business as normal,” she suggests.
Leveraging data for AI
Despite some of the reservations she has about the teeth Dunnhumby now has to bear she is bullish on its prospects for potentially using AI as described by Tesco CEO Murphy because she days Tesco has some of the best data in the market. “I’d never bet against Tesco to deliver with AI,” she says.
But more widely she expresses the view that despite the work undertaken by Dunnhumby in delivering incredible value from data many companies still have awful data today and they continue to miss the point of what can be achieved with data. This places them in a weak position in these days of emerging AI capabilities.
“All the AI players will rush their products to market and we will find the users [retailers] will be bitten as their data is no good. This will be a reflection of what has been the case for years. There’s not any intelligence [in the market] and they’ve forgotten the values of Dunnhumby,” suggests Dunn
