Averages Are Evil

Mar 13, 2015

BY JIM SAWYER – CHIEF SCIENTIST

I’m a serious NFL fan. Seriously serious. I’m the type of guy that schedules his entire life around NFL Sundays. I have season tickets and never miss an Atlanta Falcons home game.[1]To be totally transparent, I missed the Falcons-Bears game this year due to my sister-in-law’s wedding. Why she wouldn’t do me a simple little favor and reschedule it is beyond me. I took a vacation to London last year, not because it’s the shining crown jewel of a society that has had an incredibly significant historical and cultural impact on the entire world, but because the Falcons were playing there. Yeah, I’m that guy.

So it was late in the season last year, and as usual, I was settled in on the couch at home on a Sunday night, this time watching to see if the Broncos could clinch a playoff spot against the Bengals. And as the second quarter started, a graphic popped up on the screen: The Bengals’ running back was averaging a whopping 17.2 yards a carry. “Wow,” pontificated the generally insightful Cris Collinsworth, “Jeremy Hill is having a heck of a game.”

Heck of a game? If Jeremy Hill can carry it 17.2 yards every time he touches the ball, he would undoubtedly be the Greatest Of All Time! Step aside, Walter Payton. Move over, Barry Sanders. Vaya con Dios, LaDainian Tomlinson.[2]Yes, I put L.T. there. We can have that debate sometime. Open the doors to the Hall of Fame right now — Jeremy Hill is your new Numero Uno!

Of course, that’s an average.

And averages are evil.

Here’s what Hill’s play-by-play looked like in the 1st quarter:

  • J. Hill right tackle to CIN 20 for no gain. (0 yards total)
  • J. Hill left guard to CIN 23 for 3 yards. (3 yards total)
  • J. Hill right tackle for 85 yards and a TD. (88 yards total)
  • J. Hill right guard to CIN 13 for -4 yards. (84 yards total)
  • J. Hill left tackle to DEN 43 for 2 yards. (86 yards total)

That’s a bunch of little numbers and one really big number. When you put them together, the average is a medium-sized number. The average isn’t telling you the whole story.

Now let’s compare that to what we too often see in retail.

“Did you know that our average customer only shops with us 4 times a year?”

…observed the inspiring retail CEO during his inaugural town hall meeting to a throng of enthralled employees, who were waiting with bated breath to hear his vision for growth in the coming year.

The idea, you see, was that if you could simply find a way to encourage those customers to shop a couple more times a year, revenue and profit would skyrocket. Adding 2 more trips on top of the original 4 would mean a 50% increase in overall revenue, right?[3]In reality, the stated vision was to move those customers from 4 trips to 12 trips a year, which would inevitably triple the company’s revenue. However, I felt that would obscure the more important point by making the story completely unbelievable for you.

Well, sort of. On average.

But averages are evil.

Here’s what the data really looked like (this is a representative approximation, not actual data):

averages_are_evil_image

See what’s happening here? The reality is that a tiny percentage of customers shop exactly 4 times a year. In fact, the vast majority of customers are shopping 1 or 2 times a year. But this is offset by a small group of highly valuable, highly engaged customers that shop many, many times a year.

The best customers pull up the average, painting an artificial picture of customer behavior. There is no such thing as an average customer. And yet, we all too often make critical business decisions on the basis of averages.

Are you still relying on averages? Do you know who your best customers are?

Footnotes   [ + ]