February 18, 2009

A Test ... Local Minima ... E-Commerce

Earlier this week, I executed a test.

You were my subjects!

This article, in my opinion, was well written, topical, and offers significant thought leadership. You were interested in it, you downloaded the white paper at an above-average rate. This is the "10" in the chart.

This article should have inspired thought, but was a complete snoozer. This is an "8" in the chart.

And then I wrote this one about Twitter --- almost no business value and not particularly well written. This one was the most popular article of the past two months, forwarded across the world in blogs and via Twitter. I received positive comments. I received hate mail. I received comments that were so inappropriate that they had to be deleted. The audience was "engaged". And I apologize if you were offended by the article --- I am speaking of Twitter from a marketing perspective, not from use of the tool for personal reasons. From a marketing perspective, we fail from time to time.

This article was a "4" in the chart.


I did this test, on purpose, to prove a point.

When we optimize on the basis of the wrong set of metrics, we evolve in a negative or mediocre direction. If my goal was to have the most subscribers and the highest level of engagement, I'd be writing satire about social media, because that riles-up the audience and gets people to read and interact.

Remember calculus? We used to talk about "local minimas". On this graph, the "catalog filter" article represents a local minima. This article is so much more important than pap about Twitter, but the metrics suggest otherwise.

Genius-level articles are few and far between. But the really good ones don't get much interaction, don't attract a lot of readers. The mediocre ones, those attract a huge audience.

This translates to e-commerce. We focus all the time on the average things, because the average things seem to work --- our metrics are calibrated to reward mediocrity. And when we try to do something excellent, the audience rebukes us. Try to innovate, try to do something wild and crazy with your homepage sometime --- watch your conversion rate plummet --- you won't make that mistake again! Our short-term metric, conversion rate, forces us toward mediocrity.

And yet, the local minima is where the excellence happens. We should all be so lucky to have a testing budget that allows us to detect strategy that can work on a long-term basis.

Now that I read this, I think I just re-invented Seth Godin's "The Dip". See, there's no original ideas left out there!