Data and analytics are invariably portrayed as something new and technologically powered. They aren’t. In 1885, the British government became alarmed by rumors of a Russian invasion of Afghanistan. An official in London disproved the scare stories by analyzing the Russian military’s annual contract for flour. With no more flour ordered – and no plans for bakeries en route to Afghanistan – he correctly concluded that an army could not invade without bread.
Today, powered by AI, machine learning and increasingly affordable software packages, data and analytics have become more mysterious – and yet more crucial. That is especially true in retail which has the unsought distinction of being transformed by three revolutions at once: geographic, demographic and technological.
In such a turbulent marketplace, making appropriate use of data and analytics could help executives answer three of the most profound questions facing retailers:
Getting the first question right could, in itself, be the key to survival – as the continuing demise of so many famous retail names demonstrates. In the US last year, according to Fung Global Retail Technology, there were around 9,452 store closures, 53 percent higher than in 2008, when the global recession was at its most devastating. In the UK in the same period, the Centre for Retail Research estimates that 5,800 shops shut their doors. In other countries, the likes of Carrefour (in France and elsewhere), Dagrofa (Denmark) and Sweden’s ICA (which has sold its Norwegian grocery business) have, to varying degrees, rationalized their stores.
At the same time, there is a counter narrative – thousands of new shops have also opened in the UK and US, with budget chain Dollar General recently celebrating its 15,000th store – but all this volatility makes it imperative that retailers make decisions on the basis of an accurate, up-to-date, picture of the marketplace.
The wave of store closures has prompted doom-laden speculation about a ‘retail apocalypse’ yet Paul Martin, Head of Retail, KPMG in the UK, says that in many areas such a correction was at least five years overdue. “In most developed markets, there is 35-50 percent more retail space than the industry needs. The old retail model – that you have a presence in every sizeable town – has been stood on its head by online shopping, especially in the UK where it accounts for around 20 percent of total spend (and 28 percent of non-food spend). The pressure on retail is relentless – and comes from all directions: platform companies, new regulations, new business rates, the need to invest in their digital offerings are merely the most obvious. Given that the retail market in the UK is essentially stagnant – it hasn’t grown significantly in real terms for decades – the over-supply of retail space has to be addressed, or it could lead to more business failures.”
The experience of retailers in the US and UK, Martin believes, will be a harbinger of future change in other markets – notably France and Germany, where e-commerce’s market share is much lower (though growing fast). As stark as that sounds, Martin says that although “Physical stores will remain the most important retail channel for years to come – they are pivotal in terms of providing a sensory experience for the consumer – their role will unquestionably have to change.”
To understand what that change might look like, he argues, retailers need to make far more effective use of data and analytics. If they don’t, they may struggle to compete with more forward-thinking rivals and, crucially, with platform companies which have data at the heart of their business.
To be fair, retailers have long recognized the power of data. The need to get information about shoppers instead of simply rewarding them persuaded Tesco’s marketing director – and future CEO – Terry Leahy to decide to replace Green Shield stamps with a Clubcard scheme back in 1993. Such programs have helped drive growth but, in an age when data is more critical and more voluminous, are they still fit for purpose?
There are, says Bill Nowacki, Managing Director of the Decision Science Initiative within KPMG Lighthouse, KPMG in the US, five questions that such schemes should be able to answer using quantifiable data: “Are customers enthusiastic about the program? Is it easy to use across multiple channels? How much of sales growth and customer retention is due to the program? How many customers are ‘points’ obsessed – and does the program treat them differently? And is the program delivering a satisfactory financial return?”
The more you know, the better placed you are to see disruption not just as a threat but as an opportunity
Retailers who cannot answer those five questions satisfactorily need to reinvent the program. Yet even a shiny new state-of-the art loyalty scheme will only tell retailers part of the story – it can reveal what a percentage of your customers are doing, but not necessarily why. Nor can it tell you what your customers are doing when they shop elsewhere. Unless you validate it with other authoritative data, you have no idea how accurate it is. In a marketplace where technology has made consumer behavior so unpredictable, it can’t tell you what customers who never shop with you are doing and why.
So when it comes to the critical decision to open, close or keep a store, such data – even when augmented by traditional market research and old school demographic profiles – might not be that useful.
Loyalty programs are a symptom of a sector where too many companies silo data, struggle to gain actionable insights from it and ignore it or misuse it. As Nowacki says: “Executives need to trust the data, and to do that they need to understand it. There is so much complexity in business now – which is a reflection of the complexity in the markets – that companies are using much more sophisticated methods, and the algorithm that helps them make the decision seems to emerge from a black box. Executives who get a little closer, who endeavor to understand, who ask ‘What data are we using?’, ‘Why do we use that?’, are much more comfortable with it.”
Making data transparent can engender trust. As Nowacki says: “No matter how sophisticated analytics might seem, they’re always easy to understand. You can take the most complicated analytics in the world, explain it to a 7-year-old child, and they get it.” That trust becomes more critical if retailers are to use new kinds of data to become truly customer centric. At the moment, too many companies still don’t vary their product range from store to store. This doesn’t have to be the case. Using 6,500 different indicators, KPMG member firms can create a detailed picture of a retail environment, predict customer behavior and advise on product selection.
KPMG can draw on everything from neighborhood crime rates to the quality of public transport links and the proximity of ATMs to understand a specific retail environment. In many cases, the cliché about the devil being in the detail applies. “Small differences – such as locating a store 300 yards further along the street – can have a really significant impact,” says Nowacki. Research by retail consultancy TCC Global suggests, for example, that only three percent of British shoppers would take more time to reach their preferred grocery store. And why would they? The study found that the average British shopper has a choice of 11 supermarkets, 10 of which are easy to reach.1
There is nothing new about location-based data. Yet, according to KPMG International’s 2018 Global Consumer Executive Top of Mind Survey, it has not yet become the norm among traditional retailers and manufacturers2. Only half of the 530 companies surveyed said they were using real-time location-based data (physical customer tracking using GPS or sensors) to understand consumers, while two out of three said they were crunching data about the local area (demographic, competitors, crime statistics, weather, traffic). In a striking contrast, 88 percent of the platform companies surveyed already rely on location-based data.
The great strength of technology, he argues, is the vast number of permutations it can analyze. Neuroscience suggests that, as human beings, we are only capable of considering a specific problem in six or seven different ways. That tends to encourage a kind of confirmation bias in which we draw on the last analysis we conducted, or our gut instinct, instead of being guided by data.
If retailers understand a retail environment in granular detail, they are more likely to make informed decisions about where to locate a store, whether to keep it open and what selection of stock would generate the most revenue. KPMG recently helped a leading international food retailer expand its footprint in the UK. Drawing on those 6,500 indicators, it was able to create accurate forecasts of customer demand by location, by the product category and for specific parts of the day. “In this way, they were able to rigorously test new locations before investing significant capital,” says Nowacki. “Ultimately, they were able to get a better return on their investment in new stores.”
One way retailers could make their stores more effective is to improve the experience. KPMG International’s The truth about online consumers survey found that more than half of consumers prefer to shop in stores to see an item or to try it on, yet only 23 percent say they actually enjoy going to a store.3 Although the concept of ‘retailtainment’ was first mooted by American sociologist George Ritzer in 1999, it is clear that many shops haven’t got the message.
Some grocers are simplifying their stores to make it easier to shop. In the UK, Tesco has trialed a program to combat shopper fatigue in 50 stores, trimmed its number of SKUs by 20,000 to 70,000 – at one point it offered 64 variants of dishwasher tablet – and rearranged product lines to fit customer expectations by, for example, putting all its home-baking products together. The program has helped the UK supermarket to consistently grow like-for-like sales since 2016. In the US, a national supermarket chain monitors when shoppers enter its stores, uses predictive analytics to estimate when they are likely to reach the checkout and make sure enough tills are open. The program cut the average waiting time from four minutes to 26 seconds, giving shoppers a much more satisfying experience.
Data and analytics can be used to make product selection, pricing and promotions more effective. “Many retailers talk a big game about differentiating their activities between stores,” says Nowacki. “But many of them still sell the same range in stores that are in completely different retail environments.”
To acquire a granular understanding of their environments – and their customers – retailers need to change their mindset. “The retail sector’s capability for mining data is not that great,” says Martin. “They have a lot of data, a lot of it is unstructured and data analysis is not part of their DNA.” Although retailers can do a lot internally to improve their ability to analyze data – as a start, by ensuring that it is shared across functions – Martin suggests they need to think more radically, and consider whether they want to own or rent the expertise. “It might make more sense for retailers to partner with a business that has been built to deliver data.” A small but significant number of retailers are partnering with the very platform companies they are competing with.
As retailers analyze their bricks-and-mortar stores using data, they are much better placed to understand which catchment areas they should exit, which areas they should stay in (and how to make them generate more revenue) and what areas they should move in to. “The whole point of data analytics is to enhance your return on investment,” says Nowacki. “The more you know the better placed you are to see disruption not just as a threat but as an opportunity.”