January 17, 2016

Weighted Preference Pairs

That's a satisfying cup of supplemental nacho cheese.

When a customer purchases something at Taco Bell, the customer often purchases multiple items. Nachos and a Chalupa. A MexiMelt and a Taco.

From a targeting standpoint, you weight your historical transactions ... and then you create what I call "Weighted Preference Pairs", or "WPP" for short. You discount dollars spent on older transactions (i.e. spending $0.50 on extra nacho cheese two years ago means that, in today's terms, you spent $0.10). Then, you sum up all weighted historical dollars, and segment your customer base via primary merchandise preference and secondary merchandise preference.

In the future, the "Chalupas" / "Burrito" customer will behave different than the "Burrito" / "Taco" customer.

In my work, I create the Weighted Preference Pairs twelve-months back. Then, I measure how each Weighted Preference Pair responded to each merchandise category in the next year. This allows me to understand the customers who are aligned with each product category, so that your marketing efforts are appropriately merchandised.

Weighted Preference Pairs. It's such an easy analysis, and it's powerful. That's a good combo, don't you think?! More on this topic tomorrow.