I’ve been having great fun with our software this week - so much so, in fact, that I have officially termed it "tasty dog food" within the company. Our latest project was a direct mail campaign to a new list of a bit over 1000 names. Among other things, we want to test what kind of response we get for each of three different formats: poster, postcard, and foldover notecard. Now, there are certainly several ways to do this, but using our software to segment the list based on formulas was the easiest way for me.

My boss had the following requirements:

• At least 200 (but not too many more) names had to receive posters.
• VPs and above at companies of over 150 employees should not receive posters (less likely to stick them on their walls at large companies).
• Likewise, VPs and above at larger companies should not get postcards.

Now, in all honesty, I’m a bit lazy, so any time I can set up technology to do most of the work, I’m happy. My solution here was to create a bunch of TRUE/FALSE formulas and use a few spiffy features of our software to slice the lists, review, adjust, and send.

First, I created a formula to test the titles of the folks in the list and one to test the size of the company. Because I like to make my formula names meaningful, I called them RECIPIENT_IS_MANAGER and COMPANY_IS_SMALL. In order to make the selection somewhat random, I sorted the list by street address and then assigned each record an index number.

The formula for POSTER_RECIPIENTS started out looking like this:

(RECIPIENT_IS_MANAGER or COMPANY_IS_SMALL) and INDEX_NUM % 5 = 0

That last bit means that the index number has a remainder of 0 when divided by 5. This is our implementation of the Modulus mathematical function. I figured that was a good place to start when looking for 200 names out of 1000. Well, that only gave me ~150 names because the VPs at large companes got weeded out of the pool. Using 4 instead of 5 gave me ~180 names so I ended up going with:

(RECIPIENT_IS_MANAGER or COMPANY_IS_SMALL) and INDEX_NUM % 3 = 0

which gave me just over 200 names with a value of TRUE for this formula.

Next, I created my list for the foldover mailer. I had to segment out the POSTER_RECIPIENTS and include the VPs and above at large companies so they wouldn’t get postcards. The FOLDOVER_RECIPIENTS formula looked like:

((INDEX_NUM % 3 = 0) and (NOT POSTER_CANDIDATE)) or (NOT(COMPANY_IS_SMALL or RECIPIENT_IS_MANAGER))

Basically, I’m counting off every third name, tossing them out if they got a poster, and then adding in the folks who are VPs and above at large companies. That gave me ~350 names with a TRUE value and we’ll have to adjust our response statistics for the fact that this list is VP and CXO-heavy.

The POSTCARD_RECIPIENT formula, then, is really easy and gave me ~500 names:

(NOT POSTER_RECIPIENT) and (NOT FOLDOVER_RECIPIENT)

Then, I did a quick sanity check to make sure that no person had a value of TRUE for more than one of these formulas (i.e. they would be on more than one list segment). Happily, I had made no logical errors and each person was TRUE for exactly one of the three formulas.

Next, I used our List View feature to create filtered versions of my master list based on my three formulas and place the print orders for the posters, foldover cards and postcards. Approval emails were sent to my boss and to our VP of Sales to make sure that they liked the artwork. Once they clicked OK, the print files got FTPed to our vendor.

On each of the direct mail pieces going out, there’s a personal URL leading back to a custom web page for each of the people we sent to. Whenever one of these URLs is visited, my master list is going to get updated. A little simple math and I’ll be able to give percentages of visits and clicks for each of the formats. And I’ll be able to easily extract the opt-ins to put into our CRM system with another list view.

This marketing stuff can be awfully fun!