Big Data, Big Deal?

Tech Event speaker says "kinda big data" important to convenience stores

Published in CSP Daily News

By
Angel Abcede, Senior Editor/Content Development Coordinator

Jim Manzi

DALLAS -- For all the hype retailers have been hearing about so-called "big data" or the ability to collect customer transaction data and derive enough information to personalize interactions or predict purchase behavior, convenience store retailers may only be interested in a portion of the potential, said a speaker at this year's NACS technology conference, The Tech Event.

Speaking before about 400 attendees, Jim Manzi, chairman of Applied Predictive Technologies (APT), Arlington, Va., said too much hype surrounds the idea of extracting action-driving information from mountains of data that ultimately has lead to disappointment.

For the industry, predictive sales trends based on changing elements like product price or weather, or the effects change has on the overall basket are within retailers' reach, using a fraction of what big data can offer. Raw elements like complete electronic logs of customer transactions going back five years, loyalty database information and even Twitter "sentiment" data (used to see how favorably Twitter followers respond to a product or initiative) are all well within deliverable and affordable scopes, allowing retailers to influence sales and drive earnings.

Retailers will need to access "cloud" or Internet-hosted service providers as a way to make the analytics and data handling financially manageable, but these resources exist, as do price-optimization tools that can aid in developing strategies.

What will separate competitors, however, will be those who take the next step: In-field experiments.

"Using controlled, real-world experiments will make the difference," Manzi said.

But to be successful, certain elements need to be in place:

  • The proper number of stores to make for a "control" group.
  • The right calibration to tie results back to true root causes of success.
  • Low-cost tests so that the mere act of experimentation isn't prohibitive.
  • A lot of testing. Manzi said retailers have to make experiments a regular part of business, simply because most experiments (80% to 90%) will fail.

In one case study, he said the c-store retailer APT worked with added labor to its foodservice program in a test. It found that doing so in urban locations turned a profit, but in rural sites, led to losses. The chain ended up putting the added labor into only a portion of their stores, resulting in increased profit.

In another case involving the raising of coffee prices, the c-store retailer found that in some stores, gross profit on the coffee itself went up but the stores got killed on ancillary sales. Conversely, when they lowered coffee prices at other stores, gross profit fell, but additional sales rose. Again, the retailer applied specific strategies to different stores in the same chain.

The day's general sessions went on to address topics of web standards, the new HTML 5 version of the computer code, data breaches and developing inter-departmental cooperation on key IT projects. NACS's The Tech Event and coinciding Petroleum Convenience Alliance for Technical Standards (PCATS) meeting is scheduled to run through May 10.

For more reporting from the conference, visit Angel Abcede's blog, Mobile 2 Go, on mobile topics on CSPnet.com.

Keywords: 
back office
By Angel Abcede, Senior Editor/Content Development Coordinator
View More Articles By Angel Abcede