Do you know what is really driving your business?

Does the flapping of a butterfly’s wings in China really affect your sales in Tulsa?

Recently I was having breakfast with a client who runs IT for a national mid-sized retailer.  I asked him if the company was successful at testing a high labor model, adding more store associates when Best Customers were scheduled to come in.   His answer to me was very illustrative of one of the risks of “gut-level” business decision-making, and a great lesson to us all.

He told me, “That may very well make sense, but probably not right now in our company.  You see, one of our competitors is failing, and they have way too much staff in their stores.  So any discussion about labor is off the tables for now.”

Let’s think about that for a moment, shall we?

What is the three greatest factors influencing whether a retailer is successful?  Why “location, location, location” of course. (the old retailer adage).

What factor might be next in determining success or failure?  Let’s go in order:

• Square footage (are your stores simply too big?)
• Product assortment (do you keep in stock what customers need/want?)
• Pricing (are you priced competitively with the marketplace)
• Customer experience (how easy are you to do business with?)

Then, and only then, can you get to staffing levels as a cause of success or problems.  What is more, higher staffing levels has potential to improve the customer experience, not hurt it (not always true, but true enough).

So what happened at my client?  How did such a conclusion reach such a broad acceptance without rigorous analysis?

I think the answer is easy — the confusion between correlation and causality.  While this may sound like a statistical argument, it is actually anything but.  Many times when management has a simple answer to a complex problem, they tend to use any available data to prove out that answer, even if the results become somewhat far fetched, given what we know about the industry or business in general.

For example, can I make the statement that high staffing levels occurred in this failing retailer?  Absolutely.  But can I then go on to state that high labor levels CAUSED the failure of this retailer?  No way.  Identifying actual causes is one of the most important roles of marketing, and helping your company sift from correlations to causations will markedly improve decision-making in your company as well.

How do you do this?  We use two approaches to help clients sift from correlations to causes:

• Study history. It is absolutely true in business decision-making that those who do not study history are destined to repeat it.  Track the key metrics from your competitors (easy if they are publicly traded) and chart them vs. their business performance, particularly comp store sales if you are in retail.
• Build control groups. If you plan carefully, the control should receive all the other market effects — competitive, advertising, distribution, etc., but NOT receive the marketing effort.  Planning out your control groups is critical, since, in the end, you will be evaluated by incremental lift of your program group.

And finally, build a healthy skepticism about overly simple answers to complex questions.  Especially when those answers are provided without a great deal of back-up info.  Experience is important, but when we let our experience create a screen that rules out certain answers or conclusions, we are not doing our companys and our careers, any good at all.

Related articles by Zemanta

• Causality and Statistical Learning (stat.columbia.edu)
• Keep in touch with your customers (mba-geek.com)
• Customer Service: Get & Keep Clients (womensblog.score.org)
• Correlation doesn’t mean causation: yawn (scienceblogs.com)

Tagged as: business analysis, Causality, correlation, Customer experience, Database marketing, Decision making, key drivers, marketing analysis, Retailing, statistical analysis, statistics