The importance of challenging conventional wisdom and getting beyond the hype of scale
By Jim Manzi
Any experienced businessperson has seen this movie before with earlier technologies ranging from the World Wide Web to CRM to enterprise data warehouses. It’s the plot in which a technology goes from promise to hype to true application. Big data is now deep into the hype phase of this cycle. All the classic signs are there: You can eat buffet dinners all 52 weeks a year at big data conferences. Big data tag lines are now common in emails from industry analysts, and even investment bankers are tossing around the phrase. But as with these other innovations, there is real substance at the root of the hype. And – like CRM, the web, and data warehouses – big data is a big part of running any large corporation in the future.
Profitably exploiting the emerging opportunity for big data will require using some of the key learnings from companies that have already gone beyond the hype: first, an unwillingness to be snowed by conventional wisdom and technical jargon; second, the ability to act quickly at low cost, learn what works from trial-and-error experience, and then reinforce strengths; and third, a ruthless focus on profits as the success criteria for proposed investments of time or money. Those three characteristics will be necessary as data moves beyond conventional storage capacity and into the cloud. And those characteristics will be critical as retail executives balance the immensity of scale with the practicality of business applications. In short: Big data consists of small data. The challenge is to take the right data and make it drive decisions.
In the end, the transaction data remains the most important. It will show the retailer what makes consumers trade their money for our goods and services. Smart consumer businesses ignore these external data feeds or rebuild their infrastructure around them. Instead they use abstractions to extract most of the analytical value, while only needing a tiny fraction of the data volume.
The thing that is clear is that today – right now – large consumer companies can begin taking advantage of many of these data streams by capturing them at an abstracted level, incorporating them in data schemas, and using them to improve decisions.
The Humility of Test & Learn
Retailers and other executives are aware that data needs to drive decisions rather than drive them crazy. Through careful experimentation to test new programs and approaches, the most serious industry analysts have started to recognize that Test & Learn is central to making big data create value.
A Test & Learn capability for a major marketer requires a specialized analytical platform, but also has several process and organizational components. The starting point is executive commitment. The person or small group with ultimate operational responsibility for shareholder value creation, typically the CEO or president, must legitimately desire reliable analytical knowledge of the business. Second, a distinct organizational entity, normally quite small, must be created to design experiments and then provide their canonical interpretation. Third, a repeatable process must be put in place to institutionalize experimentation as a part of how the business makes decisions.
Here are two examples. First, let’s look at Wawa, a 645-location convenience store chain. Its marketing team had developed a new flatbread breakfast offering that had performed well in spot-testing. Management wanted more robust measurement of what happened to other products when the flatbread was introduced. Wawa used APT’s Test & Learn software to design a scientific test and measure the impact of the flatbread introduction across all key performance metrics and product categories. The flatbread performed well. Unfortunately, the new item was so enticing that it cannibalized sales of existing menu items. Wawa decided not to roll out the flatbread.
Number two: At Subway the debate surrounded whether they should launch a low-cost $5 product. A set of Subway franchisees had implemented a promotion selling their famous footlong sub at $5. While some franchisees were convinced that the promotion was driving incremental sales, others were skeptical. APT’s Test & Learn software compared the performance of franchisees that had implemented the $5 Footlong versus a scientifically matched group of restaurants that had not implemented the promotion. The software showed that the $5 Footlong was profitable, and executives decided to launch the $5 Footlong nationally.
These two examples show companies that have access to a staggering amount of information. They also show two companies that incorporated a systematic method of feeding that data with practical experiments that expand (or manage) key operational and marketing decisions. The systematic focus produced the right data. It cut through the clutter and produced clear results.
Test & Learn, at its heart, is a simple concept for a business promise that is begging for simplicity. The orientation should not be toward big, one-time “moon shot” tests, but instead toward many fast, cheap tests in rapid succession whenever this is feasible. The goal is to build a mountain of pebbles.