For a business, listening to customers can come in a lot of forms–from handling complaints in a call center to conducting focus groups. But customers communicate in another way that can be just as valuable: Through data. You just have to know how to interpret what they’re saying.
Michael Haydock, an IBM math-whiz PhD who also happens to be a former CEO of supercomputer maker Cray, develops sophisticated systems for interpreting data. Today, he opened the kimono on his work, releasing a forecast of retail electronics and appliance sector growth in the U.S. this summer. Haydock’s prediction: Sales at physical retail outlets will rise a healthy 4% in June, July, and August from last year’s totals.
There’s no shortage of business data analysis and forecasting in the world, but what makes Haydock’s approach different is he’s bringing a level of rigor to the process that until now was primarily employed in academic studies. He uses statistical and analytical software to evaluate both the long term sales trends and seasonal ups and downs. His “secret sauce,” he says, is applying a methodology to the data called Holt-Winters Triple Exponential Smoothing. He charts 64 potential paths for sales going out 18 months, then chooses the most likely path. Checked against real sales performance, his technique is correct 98.8% of the time. He’s very happy to report that he was able to correctly predict the path of electronics retailing sales in the aftermath of the global financial meltdown in 2008.
Why is this level of accuracy important to retailers? With super-accurate forecasts they can make better decisions about inventory and staffing levels. And, in a tight-margin business like retailing, every basis point counts for a lot. “The moral of the story is you can’t study your customer too much. That data is the valuable corporate asset,” says Haydock.
A lot of decisions in retailing are made based on the gut instincts of smart people with a lot of experience to draw from. These days, though, retailers are listening to the data, as well.