Email Frequency

You test email frequency all the time, right?

RIGHT?

I know, you don't.

Here's the results from a test, post-attribution (after averaging the results of five different attribution vendors).

• 0 Contacts / Week = $0.00 total.
• 1 Contact / Week = $0.10 total.
• 2 Contacts / Week = $0.17 total.
• 3 Contacts / Week = $0.19 total.
• 4 Contacts / Week = $0.20 total.
• 5 Contacts / Week = $0.205 total.

I'll convert the metrics to fractions of the totals.

• 0.000 contacts = 0.000 demand.
• 0.200 contacts = 0.488 demand.
• 0.400 contacts = 0.829 demand.
• 0.600 contacts = 0.927 demand.
• 0.800 contacts = 0.976 demand.
• 1.000 contacts = 1.000 demand.

Guess what? We've just established another a*(x^b) relationship!!

• a = 1.071.
• b = 0.408.

Now, it turns out that a*(x^b) doesn't work great here ... the r-squared is north of 90% but it isn't reflecting reality great ... and that's because after 3 contacts a week you don't generate any additional demand.

This is why you execute email frequency tests. If you learn that only three contacts a week generate incremental volume, then you either kill two campaigns a week (which you'll never do) or you experiment like there is no tomorrow with two campaigns because there's absolutely no downside to doing that.

Test different creative treatments ... test new merchandise ... test no discounting ... test humor ... test content ... TEST SOMETHING!!!! If your test results look like these, then you have just been given the freedom to have a TON OF FUN. What would stop you from having a TON OF FUN?????