Sometimes, even the most mundane things can extend some profound perspectives.
There was this retailer who noted a simple habit of one of his regular morning shoppers while running through video footage: she would walk into the store, pick up a pint of milk, follow that up with bananas, and check out. Imagine her surprise when one such morning she found bottles of banana milkshake stacked innocently along her shopping path. The retailer in the story, UK-based Tesco, started on the data analytics journey several years ago and hasn’t looked back – it is famous for incorporating customer science into almost every aspect of business.
Globally, some QSRs (quick service restaurants) like McDonald’s employ quantitative video ethnography (trained cameras) on drive-through lanes to determine what items are to be displayed on its digital menu-board dotting the driveway. When the lines are longer, the menu features items that can be dished out speedily; during low traffic, the menu is promptly changed to reflect items that take longer to prepare.
In fact, McDonald’s use of analytics is pervasive – it uses point-of-sale data fed into a global data warehouse from the 34,000-plus outlets worldwide, in addition to data unique to each restaurant-such as demand, customer arrival patterns, in-store and drive-through configurations, product mix, staffing, layout, menu etc. As part of its simulation modelling, the QSR uses a variety of technologies, including eye tracking to study how customers move through a restaurant and video analytics to track time spent in the store and drive-through.
In the US, retail chain Macy’s adjusts pricing in near-real time for 73 million items on the basis of demand and inventory, using SAS Institute technology. Why is all this interesting? Because these companies aren’t the digital-born-and-bred variety. They don’t possess the advantages of, say, e-commerce giants, for whom data analytics is pretty much a hygiene factor. Increasingly, offline retailers – and brands whose core revenues aren’t from digital means – are trying to bring themselves up to speed on all things big data, to push their business goals and keep up with competition.
For many such companies, research has moved from data based on survey to data based on consumption. With greater access to transaction data, rather than asking customers what is working for them, conclusions are being drawn on a proactive basis. Disciplines like sales analytics/channel analytics have now become de rigueur. “This particularly helps in promotions and discounts; you can immediately tweak them on the basis of market feedback/trends,” says Saurabh Mittal, vice-president, client services, consumer packaged goods and retail, Fractal Analytics.
Big data in retail is doubling in India every six-eight months. “With single digit margins, big data is most important for retail,” says D Shivakumar, CEO, PepsiCo India. “With over 100 million active social media users today, location is slated to become very important, and hence a focus on smartphones is essential.”
Clearly, technology is a key enabler of a brand’s go-to-market strategy. Three companies in particular stand out when it comes to data readiness: Future Group, which is using its loyalty data for increasing consumption and launching apt packaging; PepsiCo, for better operational efficiency at the ground level, and Nivea, for its operations and product innovations.
At the operational level:
Every day, there’s a huge amount of data generated pertaining to sales, products, customer feedback etc. Based on the business model and data requirements, you have to figure out the best data mix. At Nivea, the sales team gets daily, weekly and monthly dashboards customised for each individual, depending on which territory the sales officer is in, so that he is not overloaded with data. He gets 10-12 critical data points/parameters relevant to him every 24 hours so that he can manage his work better. There is a hierarchy approach where his senior gets a view of a larger territory, and in turn, his senior gets a view of the complete zone and so on. Giving executives actionable data they can actually use is better than overwhelming everyone with a full sales report. “It isn’t just about the availability of data; it is about focus on the right data points,” explains Rakshit Hargave, MD, Nivea India.
Nivea uses certain customised ERP (enterprise resource planning) packages for conducting historical trending on the basis of inputs given by managers. Whether it is for demand or supply planning, dispatch or inventory planning, this approach helps in managing a larger system with multiple SKUs and locations. Elements like stock forecast accuracy have become much better compared to 8-10 years ago where manual systems of estimates would get collated, leaving in its trail a higher possibility of error.
“Operations is anyway a left-brain job: process-led and scientific,” says Sandip Tarkas, president, consumer strategy, Future Group. The company uses data to look at product adjacencies (a combination of products sold best as a package), product display and the depreciation rate of inventory. Carpets can either be stacked horizontally on the ground in a pile or dangled from hangers vertically, for instance. The Future Group uses algorithms to figure out at what rate they are picked up under each kind of display, thereby arriving at the right kind of display for every category.
Now consider PepsiCo, which has a four-fold objective for its sales automation system: standardise, simplify, automate and eliminate. The system enables frontline sales teams to streamline processes and manage time better. “We introduced hand-held devices in urban markets three years ago that capture market data that is integrated with our back-office operation software at the distributor point to manage inventories and billing,” says Sudipto Mozumdar, senior director, customer development, PepsiCo India. “This has helped us to move all the distributor claims from manual to 100 per cent automated.”
Let’s start with the millions of retailers that stock PepsiCo’s products. Through hand-held devices, the company tracks what each store is buying – and selling – over a period of time. With this knowledge, PepsiCo can tailor products to specific stores based on their needs and consumption. For example, if a big store caters to an affluent clientele in the vicinity, PepsiCo can target/tailor its new premium launch to the segment that frequents this store instead of launching the offering simultaneously across all stores. This sort of launch is sharper and targeted, and the company can garner quick feedback before going the whole hog.
PepsiCo is currently working on developing modules where it can help the salesman push particular SKUs that a store finds easy to move off the shelves. “So it will be like a ‘prompt’ that will remind the sales executive that the store that he is visiting sells ‘x’ products, and since he hasn’t sold ‘x’ to them for a while, he needs to push this product,” says Mozumdar. “In a sense, we will partner with retailers, because we don’t want to end up selling things to him which he doesn’t sell further on, thereby blocking his inventory.” From just using data to create targeted programmes, this will help the company get into proactive selling.
PepsiCo also equips its sales force with tablets that enable them to get role-based reports on their KPIs. That apart, PepsiCo aids distributors when it comes to their working capital. Software tracks how much he is selling, how much of credit is going into the market, which are the stores he caters to, etc. Now, as he has visibility on what his stocking norms are, the orders that get generated and driven to him can actually be tailored to replenish what he has sold. So, the distributor becomes more efficient.
PepsiCo funnels all this information into dashboards that gives the company a good sense of how it is doing on a particular initiative. The company’s advertising can be played on these tablets to showcase before the retailer the company’s upcoming products, making it a coaching tool. This way, the whole interaction moves a notch higher than the standard product brochure familiarisation process of the pre-digital world.
The product is the hero:
A conventional consumer company works in a different manner than digital-led companies. “From a personal care consumer company’s perspective, I would define certain technical variables in terms of numbers based on past records or what competition does,” says Hargave of Nivea, talking about the relationship between product development and big data.
Data also helps Nivea do a lot of claim testing. Says Hargave, “For example, if you are looking at certain levels of moisturisation or dark spot correction, there is a guidance data system that the product has to do ‘x’ correction in ‘y’ time. So numbers play an important role because unless you quantify things, you can’t develop products.”
The genre of whitening deodorants is a result of data analytics, says Nivea. While the segment is prevalent in the West, feedback from deodorant users in India on social media led Nivea to conceptualise and launch the whitening deo range in the country in 2012.
Data from loyalty marketing:
Even until a decade ago, retailers would fall back on focus groups for data because transactional data was never mapped the way it currently is. “Now we know which customer comes at what frequency, purchases what kind of items, her average ticket size, her family size based on what is purchased, and her geographical footprint – which outlets she prefers to shop at. All this information can be extracted from loyalty cards. “The customer drops a lot of cookies on you,” quips Tarkas of Future Group.
While its customer loyalty card Payback maps shopping occasions and profiles its customers accordingly, one of the group’s loyalty programmes goes beyond that. In 2010, Future Group launched its telecom brand T24 in partnership with Tata Teleservices to provide additional loyalty benefits to its customers. With T24 the company got into a customer’s life. “We know who she banks with and where she fills her car’s petrol… without violating privacy, of course,” says Tarkas. “We can get to know which customer is roaming a lot out of her city based on her mobile usage, and hence what sort of travel products should be targeted at her.”
Which brings us to distance mapping: how far a customer is willing to travel to reach, say, a Big Bazaar outlet. “A lot of our stores reflect a very high long-distance customer profile. We call these our feeder stores,” explains Tarkas. For example, the LIC Road store in Kolkata, close to both Howrah Station and Sealdah Station, is one such feeder store. Shoppers at the store are found to commute from nearby towns to Kolkata for work – when they share their permanent address for loyalty membership, it is often a long-distance one. These customers have larger ticket sizes, but their frequency of visits is lower. Many such customers aren’t quite sure if they can get a T24 recharge in their own town -which they can – so they end up recharging for Rs 3,000-5,000 at one go just in case they can’t come back for a while.
Market demand projections:
For any marketer, predicting demand or forecasting the size of a potential market is a tough task. But with ammunition like big data, PepsiCo considers itself market-ready. “Depending on why you want to enter a particular state in India, certain numbers (of stores) will be thrown at you from our data system,” says Mozumdar. There are models to predict how many stores a product needs to get into. For instance, the model will tell you if you get into 200,000 stores, you would end up selling to the tune of Rs ‘x’.
Future Group too is gung-ho about predictive analytics. For one, it does life-stage segmentation of customers to get a sense of her future buying patterns. Say, when a customer starts buying diapers, you get to know that there is an infant at home. There on, you can predict the child’s development and the kind of things the mother will buy along the way. “A lot of people buy apparels for their kids from hypermarkets because kids grow very fast, and parents may not wish to spend big money on top labels,” observes Tarkas.
Marketing mailers at Future Group are based on the RFM (reach, frequency and monetary value) model. The company evaluates the merchandise that is picked up and comes up with the ‘next logical offering’. “We have done this in a lot of categories and gained traction with several customers and often we get the ‘aha’ moment, where customers feel we understand them,” says Tarkas. This is similar to how online players like travel sites use cookies to track you while you are browsing other sites, to remind you of unfinished transactions. “In fact, there is more information available offline. The potential is vastly unexplored,” Tarkas sums up.