Wednesday mornings in the early 90s were an important day to me. After hours of keying green bar reports into Lotus 1-2-3, I finally had updates on all my direct mail campaigns going back about five full years.

I’d learned direct response by building up campaigns in a granular fashion during the merge/purge. Assigning names to lists (really, list/segment combinations) and tests in such a way that I could look at hundreds of thousands of responses week by week, test by test, and list by list. And not only the level of response, but the “take” of those customers over time and their projected LTV as it accumulated over time.

On that Wednesday, looking at the results for a particular campaign–which was performing well in aggregate–I noticed a list/segment that I’d mailed many times and which had very predictable performance. But in this campaign the results were out of line. I knew where that catalog company prospected, I could recite the size and composition of the list from memory and who mailed that list on a regular basis.

How’d I know that? I was the geek that, when somebody told me “you need to know X”, decided that it might be helpful to also know A through Y, as well as AA through ZZ about that thing. So I’d basically memorized the data cards of the over 400 lists I regularly mailed and used to even key their data cards into my spreadsheets to check their math on the size and growth of new customers and overall file size. Just in case.

Something was wrong. I pulled one of the binders off my shelf and looked at the last year or so of data cards for that list, as well as the samples of the catalog mailer’s controls. After putting that all together, I called the list broker with the question:

“They changed their control, right?”

After all, it was the only reasonable thing that could explain my aberrant performance. I followed it up with my second hypothesis:

“The changed to a slim jim format? I hope not.”

Remember the geek above? I also knew how the control mailings for the catalog mailers whose names I rented drove my performance. And I knew that slim jim acquisition (or even current customer catalogs) degraded my performance, when mailing names acquired with that format.

A couple of hours later my broker called to confirm. Indeed, they’d changed their control to a slim jim and the changeover happened right around the time which would have caused their hotline name performance to weaken for me.

Why would this thirty year old tale matter to anybody who cared enough to get this far?

It probably doesn’t. Nobody keys in three foot stacks of data into worksheets anymore. But if you know how to organize your campaigns properly and chop things up in the right way, if you understand where the names came from, if things go sour, you might be able to find the answer. No data science, AI, or fancy SaaS platforms required!

More here: https://markpilip.com/2020/12/04/who-pulled-the-names/

Also posted on Linkedin here.