You might know Best Buy as your neighborhood go-to source for TVs and DVD players, but to a marketer, it’s a dream case study in customer analytics. As a global retailer with a portfolio of brands, the challenges of extracting competitive value from its data are as big as the opportunities. Scott Friesen is Sr. Director of Analytics in the Consumer Insights division at Best Buy – and while he points out that his organization reports up into marketing, he’s quick to clarify that his team’s analytics practices touch many other departments across the company. Here, he makes a few recommendations about customer analytics and marketing.
- Be helpful, not annoying. Imagine you have a friend, and this friend asks you for money every time you see him. Besides being annoying, this behavior is not likely to contribute to a long-term relationship. Similarly, Friesen says, if every communication to a customer is asking that customer to buy something from Best Buy, the company may do more harm than good.Best Buy’s Geek Squad services give the company an opportunity to engage with customers in a supportive, helpful context.And Best Buy has also initiated a series of lifecycle emails called “Happy 24,” named for the typical 24-month duration of a cell phone contract. Emails establish a “contact rhythm,” and start at the moment of purchase with getting started tips, to later offers of accessories and end-stage guidance on how to replace or recycle an old model at the end of a contract.
- Stay current on your constraints. In e-mail marketing, for example, cost used to be a constraint. But sending email has gotten so cheap, that marketers at Best Buy suddenly found themselves sending millions of them. “At that point,” Friesen says, “the constraint becomes the customer’s tolerance. Over past year, we’ve seen email volumes come down, and ROI go up, because we’re focusing on sending the right ones, not just more of them.”
- Make your results actionable. Friesen says his boss borders on obsessive about insisting that analytic results are actionable. “We use SAS, we’re big fans of the product,” Friesen says. “But having the right answer is not good enough if the business doesn’t put it into action.”And putting that data into action takes more effort than leadership often sees. In addition to addressing the recruitment and training of analytic talent and translating business problems into an analytic approach that his business colleagues can grasp—the discussions that his boss often sees him in—Friesen and his team have a slew of technical issues to manage like hardware, data architecture and access, and assessing analytical tools.
- There is no golden key. The age-old debate of how you define a customer within a database is still alive and kicking. Friesen is adamant in his assertion that there is no golden key, no one single identifier that can reliably link all the variations of a customer’s name together into an integrated view. For example, a loyalty number might not always represent the individual that the account is tied to, because families often share those accounts. And that means marketers might be making inferences about that consumer’s behavior—which, through the pursuit of analytics, marketers are trying to avoid.
- Decide how fast is fast enough. A colleague of Friesen’s at Best Buy has a rule of threes to deconstruct the term “real-time” and how the concept applies to customer expectations: 3 milliseconds to render a Web page; 3 seconds to pull up a customer’s information when they call on the phone and 3 minutes to trigger a campaign offer. The modeling team develops models that are appropriate for the speed expectation.
Get more great tips from Friesen in this post by Beth Schultz, editor of AllAnalytics.com: Best Buy analytics exec shops for customer insight.