Data for execution: Unfiltered

We need mental maps to guide action in an uncertain world, but these maps are always imperfect. They reduce complexity to simplicity, omit critical variables, underestimate rivals’ ingenuity, and tend toward stability in a volatile world. The more turbulent the market, the faster maps grow outdated. Keeping a map fluid requires the right data-i.e., real time, unfiltered, and shared.  My last post discussed real time data, while this one describes the importance of unfiltered information.

Many organizations deluge employees with more data than they could possibly read, let alone digest. To stem the flood of information, raw statistics are packaged into standardized reports such as sales pipelines or monthly budget reports.  These reports include standardized categories that measure what matters, while filtering out information deemed irrelevant. Further filtering takes place when data are aggregated (from local to cumulative results, for example) and by packaging data into reports that impose a single interpretation of the facts.

Data filtering confers several benefits–It increases the speed and efficiency of data collection; mitigates information overload; provides a common language to interpret what is going on, and allows performance comparisons across units and over time.  For all its advantages, filtered data has one big drawback–it screens out anomalies that fall outside pre-defined categories or interpretations. Information that does not fit neatly into a pre-existing category is often dismissed, distorted, or ignored.

Ignoring incongruous information is efficient, dangerous when trying to keep a mental map fluid. Anomalies are surprising outcomes that differ from what we expect, and they often slip through the cracks of established reporting systems.  Common anomalies include products that shouldn’t sell but do, an inexplicable competitive move, or customers using products in unexpected ways. Anomalies highlight gaps between a mental map and the underlying terrain. Sometimes they signal a shift that renders an existing map outmoded.  Other times incongruities demonstrate that a map was wrong to begin with, as is often the case with a start-up’s business plan. Managers who rely exclusively on codified reports often miss clues that their map needs updating.

Anomalies can signal latent opportunities. Repairman for Chinese appliance maker Haier, for example, discovered that customers in one rural province used their washing machines not only to launder clothes, but also to clean vegetables. The repairmen relayed the information to the product manager who spotted an opportunity. She had company engineers install wider drain pipes and coarser filters that would not clog with vegetable peels. These modifications allowed the washing machine to clean cabbage as well as shirts. Haier then affixed large stickers emblazoned with pictures of local produce dancing with clothing on the modified washers. The stickers included instructions on washing vegetables safely. This new product, along with others (such as a washing machine that also made goat’s milk cheese), helped Haier win market share while avoiding cutthroat price wars.

To spot gaps in their map, leaders should balance standardized reports with regular exposure to raw data, unfiltered by reports, aggregation, or colleagues’ interpretation.  In Tesco, thousand of executives spend one week per year working in junior positions in one of the company’s stores, stocking shelves or working the tills. Marcel Telles, the former CEO of InBev, the brewer that acquired Anheuser Busch last year, typically spent two weeks each month in the field visiting distributors, observing competitors, and talking to customers.

Note raw information is a complement to real time quantitative data, not a substitute. Few retailers mine customer data as rigorously as Tesco, but they do not rely on codified data alone. Quantitative and qualitative information work best in combination. Standardized reports can surface anomalies in real time, and guide management on where to observe a situation first hand. In the summer of 2007, Zara introduced a line of slim fit items, including pencil skirts and tight jeans in bright colours like yellow and cobalt. Daily reports from the store revealed a sales far below expectations. The numbers remained mute on why the clothes weren’t selling. Zara marketing experts called store managers to discuss the sluggish sales, and found to their surprise the problem was not colour, fabric, or style. Instead, the slim fit styling required women to buy a size larger than usual–a blow to shoppers’ self-esteem that depressed their willingness to buy. Armed with this consumer insight, Zara collected the clothes back to the factory, and replaced the labels with the next size down. The relabeled collection was a hit.

Zara has designed its organization to smooth the flow of unfiltered information. While many fashion retailers franchise up to 90% of their stores to get big fast, Zara franchises only 10% of its outlets. Zara executives found that tensions and misunderstandings between franchisees and Zara employees garbled communication. Even as Zara has overtaken the Gap as the world’s largest fashion retailer, executives have avoided the temptation to add layers of management between the stores and headquarters. The marketing manager on any design team communicates directly with the stores in his or her region on a weekly basis. Indeed, Zara rotates marketing managers through store manager roles to ensure they know which questions to ask and how to make sense of the answers.

My next post will discuss the importance of shared data.

Leading in turbulent times

This blog is no longer active but it remains open as an archive.

Don Sull is professor of management practice in strategic and international management, and faculty director of executive education at London Business School. This blog is dedicated to helping entrepreneurs, managers, and outside directors to lead more effectively in a turbulent world.

Over the past decade, Prof Sull has studied volatile industries including telecommunications, airlines, fast fashion, and information technology, as well as turbulent countries including Brazil and China, and found specific behaviours that consistently differentiate more, and less, successful firms. His conclusion is that actions, not an individual’s traits, increase the odds of success in turbulent markets, and these actions can be learned.