August 02, 2016

Loyalty Lizard Logic

I've measured this phenomenon more times than I care to mention.

"Brand X" is upset that customer loyalty isn't great ... so instead of doing the hard work to offer merchandise that customers have to have, "Brand X" decides that it must instead create a loyalty program, and then offer discounts/promotions and other nonsense in an effort to increase customer loyalty. The marketing analyst points out the following:
  • 100,000 12-month buyers ... they spent $240 last year.

Then, an analyst mucks up the measurement process. She quickly determines the following:
  • 10,000 12-month buyers are in the loyalty program / they spent $600 last year.
  • 90,000 12-month buyers are not in the program / they spent $200 last year.
  • Average customer spend last year = $240.
Powerpoint slides are quickly assembled ... and the analyst proudly proclaims that the loyalty program is creating customers who are worth 3x as much as non-loyalty customers. Everybody offers a round of applause ... they loyalty vendor, sitting in the room, winks at the CMO with an "I told you so" level of arrogance that makes the CFO want to vomit.

Why does the CFO want to vomit?

Because the CFO spent money on the loyalty program ... but still has a housefile of 100,000 customers spending $240 a year.

Nothing changed.

Well, two things changed.
  • A loyalty vendor is getting paid.
  • A marketing analyst is writing terribly biased queries.
But the organization doesn't care ... they put down the gas pedal. A year later, the results look something like this.
  • 17,500 12-month buyers are in the loyalty program / they spent $580 last year.
  • 82,500 12-month buyers are not in the program / they spent $170 last year.
  • Average customer spend last year = $241.75.
Once again, everybody but the CFO misinterprets the data ... they see the $580 / $170 ratio, and they proudly proclaim that the loyalty program is causing a 3.41x increase in spend. #Wow.

But the CFO is doing slightly different math.
  • 100,000 customers spending $241.75 = $24,175,000.
  • 100,000 customers used to spend $240 = $24,000,000.
  • Incremental value of the program = $24,175,000 - $24,000,000 = $175,000.
  • Profit Factor = 40%.
  • Cost of the Program = $100,000.
  • Program Profit = $175,000 * 0.40 - $100,000 = ($30,000).
Technically, the CFO's math is also wrong. The CFO is assuming that all of the gain in 12-month buyer spend is attributed to the loyalty program, and that's not a fair assumption. More/less demand may have been generated, after accounting for changes in marketing strategy, discount/promo strategy, and merchandise productivity.

Regardless, you get the picture.

It is quite likely that the loyalty program actually caused Lifetime Value (LTV) to decrease, because the sales gain did not offset program costs.

And it is quite likely that the marketing executive will speak at a conference next year, and will tell the audience that they achieved a 3.41x increase in spend in their loyalty program. The audience will enthusiastically cheer, knowing that they've proven once again that loyalty programs work. They'll enthusiastically endorse lizard logic, won't they?

P.S.:  The marketing analyst and the CFO analyzed the data incorrectly. I could share how I'd analyze it, but you'd criticize me, suggesting I was measuring things incorrectly as well. Then, you'd share with me how you'd analyze the data, and I'd tell you that you were wrong. We'd all go in circles yelling at each other, all of us wrong. The key is to realize that as we triangulate toward an accurate answer, we all notice that sales are essentially not increasing, and that is the only metric that matters.