Execution starts with when a team or organization forms a shared understanding of the overall market situation. A start-up’s business plan and an established company’s strategy are both examples of what I call mental maps, shared models that represent reality and serve to guide action. Mental maps can range from detailed plans in thick binders to a simple insight sketched on a cocktail napkin. Differences in form should not obscure similarity in role–all mental maps represent the environment, highlight important variables, and suggest a way forward.
Even the best mental map is an imperfect representation. Mental maps can only incorporate current knowledge, and exclude new insights that will only emerge in the future. Maps simplify a complex reality, thereby ignoring potentially important variables and interactions. Competitors will go out of their way to exploit blind spots in any map. We know only one thing about our maps with certainty–they are flawed.
How can leaders update imperfect mental maps as circumstances shift? An important part of the answer lies in collecting the right type of information. Along with my co-author Stefano Turconi, I have been studying the type of data required to map a situation in flux, and identified three critical attributes: The best information is real time, unfiltered, and shared. My next few posts will discuss each in turn, beginning with real time.
Many companies make sense of their environment in an episodic manner, when teams meet each month, quarter or year to discuss aspects of the market. This episodic approach has several disadvantages. First, it leaves managers less time to react. If a product managers needs a week or two to collect and manually calculate sales data by product line, she loses precious time to respond to shifting trends.
Second, relying on stale information prevents managers from anticipating future opportunities or even understanding the present. Instead analyses based on historical data do little more than predict the past. Like light from a distant star, old data reveals little about what is happening right now, instead illuminating past events.
Finally, episodic data collection and interpretation prevents managers from building an intuitive feel for the flow of events. When data are discussed on an irregular basis, much of each meeting is spent repainting the picture of what is happening in the market. When employees view markets through a series of static pictures glimpsed at intervals of a month, quarter, or year they cannot develop deep intuition about how the market is evolving. As a result, they find themselves surprised by events time and time again.
Managers can take concrete actions to ensure real time data. At Zara, store managers receive detailed, hourly sales and replenishment data on a handheld device. At the retailer’s headquarters in La Coruna, Spain, cross-functional teams of designers, marketing managers, and buyers pore over daily sales and inventory data from the stores to continuously update their view of the market, and spot opportunities to introduce new products.
The recognition that real time data matters is not new. From the early days of his steel company, Lakshmi Mittal and his team ran the firm’s global network of steel mills using the partha system, a set of practices developed in the nineteenth century by the Marwaris, a tightly knit group, renowned for their commercial savvy. Originally based in northern India, the Marwaris migrated throughout the Indian sub-continent when the Mongols invaded their territory. Faced with the need to track performance through out their far-flung operations, Marwari merchants honed a system whereby they calculated revenues and cash flows from operation at the close of the day.
New technology enhances our ability to capture real time data. Engineers at Google found that flu-related search terms correlated with actual flu outbreaks. Google flu trends tracks flu-related search queries, and uses the aggregate data to estimate the incidence of flu outbreaks within geographic regions. Google flu’s monitoring system reports results within a day. Traditional flu monitoring systems, such as the US Center for Disease Control, require up to two weeks to collect, synthesize, and publish data.
My next post will discuss the importance of supplementing codified data with qualitative information that is not filtered through standardized reports.