a + b + c = c + b + a

My takedown of Wikipedia’s “gimme gimme gimme” ad copy drew this response from subscriber Matt:

+++

Funny you should mention wikipedia. I’ve actually sat in a presentation by their data scientists and they have probably done more a/b testing on that banner than 99% of websites….. and they make bank… millions.

+++

Awesome—didn’t know that.

That made me wonder if I was too harsh on Ol’ Sad Eyes Jimmy.

To find out I jumped on teh googlz to see what I could find out about their testing.

Which lead me to this:

https://meta.wikimedia.org/wiki/Fundraising_2010/Banner_testing

It’s a HUGE wiki page (natch) with all the results from their testing from a few years back.

I took a look at a sampling of their ads, and here’s what I discovered…

Yes, they did a BOATLOAD of testing.

They tested different pictures of Jimmy.

They tested using different Wikipedia authors to deliver the message.

They tested different background colors for the banner.

But you know what they didn’t test?

THEY DIDN’T TEST DIFFERENT MESSAGES.

They’re all variations on the same “guilt and obligation” theme:

“You get free info from us. It costs us money. Donate to us or you’re a poor excuse for a human being.”

And so every single ad they tested SUCKED.

Big time.

This highlights a really common fallacy with “conversion rate optimization” and “split testing”…

You can easily end up testing things that don’t make a big difference.

As marketing pro Glenn Livingston says:

“Most people think they’ve tested 20 ads. They’ve really tested 1 ad 20 times.”

If Wikipedia was smart about this, they’d hire a handful of skilled direct response copywriters to come in and churn out 50 different ads…

Taking COMPLETELY different angles.

Story-based ads.

Benefit-based ads.

“Make ’em cry” ads.

“Make ’em laugh” ads.

As the copywriting saying goes, “Test shouts, not whispers.”

If you test 10 ads like this, 1 or 2 will beat the others by a factor of 10X.

Now, coming up with 10, 20 or 50 completely different angles on your product can be daunting.

I know my brain sometimes starts to scream after 2-3.

There’s a great tool for this, though…

The Copywriting Swiss Army Knife

This tool makes it easy and automatic to come up with great new ways to pitch your product.

You can use it for headlines, emails, sales pages…

I use it when I’m stuck and need a fresh idea.

Details are here:

http://joshuaearl.com/sak

Related posts:

  1. An Open Letter to Wikipedia Dear Wikipedia, Today I visited your site to look up the titles of the books in the Harry Potter series (which I have not yet, to my undying shame, read). The following message greeted me (dressed in full funereal black): +++ Dear readers, We’ll get right to it: This week...
  2. Who else wants a bigger email list? My #1 goal for 2016 is growing my email list to 10,000 subscribers. That’s a 7X increase—pretty ambitious. And my new “weapon of choice” that’s going to help me get there is a suite email marketing powertools from a startup called Thrive. Here’s the problem I’ve always had in the...
  3. Marketing smarts from a 6-month-old My youngest son is a beast. He’s just hitting 7 months old now, and he’s already wearing 18-month clothes. (No idea where he gets it—my build is… let’s just call it slight and leave it at that.) A few weeks ago he figured out how to haul his huge frame...
  4. What’s the best time of day to send your marketing emails? Does it matter what time of day you send your emails? And what about days of the week—are some better than others? I have 2 answers for this. One is the tried and true response of all consultants: “It depends.” And truthfully it does. Different markets have very different habits...