Adrian walks into a retail store only to find that his favourite brand of ice cream is sold out; other brands are overstocked. Party planner Deb picks paper plates from one online store, but soft drinks and soda from another. An irate Pam confronts the store manager saying: “Two heads of broccoli is all I find in the produce tray, just not enough to see me through an entire week.” Are these really life or death issues? Well, not exactly. Even so, retail experts think it is pertinent to retouch these small, but undesirable, operational blemishes and furtherwing retail’s customer relationship. And by so doing, make the shopping experience better: highly personalised, anytime-anywhere, and affordable.
Let us begin by drawing out scenarios that help extract insights to improve decision-making and operational efficiency. We can do this by leveraging the ‘four internets’, which are separately identifiable as virtual internets of people, of things, of data and of ideas.
The Internet of people
The Internet of People is represented by the set of interconnected information about individuals, including their social and collective activities and interests, attitudes, and images, audio and video. This can offer segmented and holistic views on human behaviour, perceptions and interactions in space and time. At its core, it is about customer centricity. The Internet of People allows organisations to expand business processes beyond the borders of the enterprise. This further enables the fashion industry, for instance, to find the next craze before it occurs by analyzing what people talk about in social media.
The Internet of things
Especially in large retail stores, there is an entire lineup of devices made ‘smart’ with implanted sensors and actuators. Retailers must come to terms with data originating from these devices as well as from the RFID tags applied to high-value goods to track their sales journey.[] Keeping this aspect in mind, teh retail industry is arming itself more and more with an advanced big data arsenal, so it can ‘tame’ the variety of large and difficult-to-process data streaming in, including from their online grocery business, loyalty cards, and social media posts. When there is a possibility to get information about a physical object or a process by instrumenting it with sensors, RFID tags, transmitters, GPSs, logs and other means of sending information via wired or wireless networks, the opportunities to analyse the data and find new patterns are endless.
The Internet of data
The Internet of Data is about bridging information to understand physical, societal and business environments by connecting data at scale, both inside and outside the enterprise. The most obvious characteristic of the Internet of Data is variety: images, video, geo-location, etc. combined into a data fabric, to hold the information that organisations wanted to have all along.
The Internet of ideas
The Internet of Ideas is all about the power of connected minds. It involves humans at scale and aggregates individual ideas about societal, business and physical environments through crowdsourcing, crowdfunding, leveraging open-source products and integrating ideas from outside the enterprise. It also provides solutions for generating new ideas from outside sources or that need multiple perspectives or statistically significant representation of participants to enhance the business benefits.
For some strange and inexplicable reason, almost everyone thinks their olden days were better than their present. So is the case with retail when it comes to coping with data. In a less competitive past, retail companies had enough stocks of structured information, typically stored in tabular formats, to get comfortably by. At least that is what they liked to believed. A lot of retailers did not feel the need to process and draw meaningful patterns, like hidden market trends and customer behaviour, in their unstructured data dumps. But as newer and nimbler competitors began snapping at the heels, the game soon became supercharged for a host of retailers. They soon realised that mining heaps of operational data would help them segment customers better, create new product lines, recapture lost ground with price adjustment, convert non-buyers and, above all, keep tab on customer preferences as they change over time.
If data shows party crowds are buying only paper plates and plastic forks, it can be reasonably inferred that the competitor is making a buck selling them soft drinks at a lower price. Well, this should ring the warning bells and maybe it is time to compare and adjust the price or throw in a discount coupon and bring back lost business. Or, say, baby wipes are fast-moving, but not diapers; it probably makes sense for the retailer to launch an own-label range of baby laundry. Shoppers who buy diapers for the first time at some stores receive coupons not only for baby wipes and toys but also for beer, according to a Wall Street Journal report. Aggregation of customer data collected via personalised loyalty cards, scanned at the checkout, reveals new fathers tend to buy more beer at the retailer because, with a kid at home, they cannot spend as much time at the pub.
The foreseeable future of retail lies in examining large amounts of granular customer data of different types in an effort to unlock hidden trends and connect the dots. Analytics teams can crunch this data to develop thoughtful business models that can tell retailers what they ought to be selling in the future. But without the right kind of data mining gear, approaches, and imagination, all this is pretty much like needle in a haystack At a minimum, retailers must be able to paint a more accurate picture of demand, so they can steer clear of two extremes – running out of stock and holding too much stock. The sight of a lone drumstick leaning coolly against the produce tray can be annoying for customers looking for a week’s stock. Neither do customers want to be left with just two heads of broccoli in the produce tray. Predictive analysis[ http://blogs.sap.com/innovation/analytics/what-is-predictive-analytics-027317] is a handy tool, which can, to an extent, spare the retailer this exasperation. At the same time, it ensures retailers get as much inventory required into stores and customers have fresh inventory to pick from.
Predictive analysis takes into account current and historical facts to foretell the future using advanced algorithms. When retailers order groceries, say, for the next day or next week or next three weeks, one of the things they need to factor in is the impact of weather or seasonal events. Predictive analysis lets them sculpt a sales model that combines data from weather and seasonal events with sales data from across stores and products to predict future demand, based on seasonal forecasts. The historical data captured for the previous day, or week, or month, also needs to be taken into account. Retailers must, on a continuous basis, improve the quality of their algorithms and tweak the demand forecasting model, so it gets better than the best. This way retailers can make shopping trips better for shoppers by ensuring products are always at hand when needed. Taking the predictive analytics route, a multinational grocery and general merchandise retailer saved as much as £100 million a year by reducing wasted stock.[ http://www.retail-week.com/topics/analysis-how-tesco-and-otto-are-using-data-to-forecast-demand/5053784.article] Many retailers from more developed markets have set up centres in emerging destinations like India to help ‘crunch’ their tidal wave of big data and deliver key insights and foresight on customer behaviour.[ http://www.tescohsc.com/media-centre/media-centre/article/big-data-analytics-taking-a-deeper-dive-financial-express]
To steer ahead of the rat race, retailers need extra enhancements that can ensure 100 per cent availability (high availability) of stocks regardless of time or day. Here, ‘video analytics’ come in handy. Automated video cameras looking down at the produce department shoot millions of images, which are analysed to detect if a certain item of grocery is out of stock, near-end of stock, full stock or fresh stock; the results of the analysis are automatically played back into the retailer’s ordering systems. A world-famous bakery replaced calipers and colour cards with high-speed image analytics to scrutinise thousands of buns per minute for colour, size and even sesame seed distribution – instantly adjusting oven and other process controls to create uniform buns and reduce wastage. Another food products company similarly photo analyzes and sorts each and every French fry produced to optimise quality.
Loyalty cards provided retail with an unprecedented level of detail about ‘who their shoppers are and what they buy or do not care to buy’. It has helped retailers keep abreast of shopper trends and transformed them into firms of endearment that harnessed emotional connect with the shopper. Having significantly cut the degrees of separation with their customers, retailers are now taking the next logical step – that of targeting online customers with banner ads based on their age, gender, education, socioeconomic status, interests, lifestyle, browsing behaviour and so on.
Carrying forth this conversation into the physical world, it is now possible to show an ad to a retail customer at the cash register custom-made to their demographic. The concept makes use of face-scanning technology, a game-changer for retailers, to identify the shopper’s approximate age and the length of the hair is used to work out the gender.
The advanced facial recognition system also spots known perpetrators entering one’s locations so one can stop them before they cost money.
Through advanced facial scanning, the following can be achieved:
- Receive descriptive alerts when pre-identified shoplifters walk through any door at any store
- Get alerts when known litigious individuals enter any of the locations
- Build a database of good customers, recognise them when they come through the door, and make them feel more welcome
- Keep repeat customers coming back and spending more at the store
Modern retail traces its genealogy to the market stall, an immobile temporary structure like the ones in London’s East End, where merchants would display their wares and call them out to passers-by. In retail’s infancy, the customer was at best a sketchy concept. Computerised checkouts in the ’80s captured and filled in more details around the customer. In the mid-1990s,[ http://en.wikipedia.org/wiki/Clubcard] by capturing details of individual customer behaviour, loyalty cards added flesh to the stick picture of the customer. Retail data analytics goes further along this road and attempts a photo-realistic representation of the customer, with as much detailing as possible. Needless to say, all of these systems need to be ethically deployed, so the individual has more control over personal data, while being able to take advantage of new technologies and shop for a better deal.[ http://ec.europa.eu/justice/data-protection/document/review2012/factsheets/1_en.pdf] A Forbes Insights report, sponsored by Turn, a provider of data-driven marketing services, says ‘both businesses and their customers generally agree there is value in sharing personal information.’[ http://www.forbes.com/sites/forbesinsights/2013/11/13/big-data-tesco-probes-the-frontiers-of-privacy/]
Ultimately, with the emergence of cutting-edge technologies like face scanning, augmented reality, virtual mirrors, interactive walls, bluetooth beacons, connected fitting rooms and virtual queuing systems, the potential to better comprehend shopper behaviour enhances, thereby modernising and esteeming the shopping experience for the customers.
About the author: Vinod Bidarkoppa, Director (Group IT) and Chief Information Officer, Tesco HSC & Member of the Board (Tesco HSC)
