A version of this article originally appeared in Postings, the member newsletter of the Direct Marketing Club of New York.
One of the biggest changes in the direct mail world during the past decade has been large mailers’ expanded use of cooperative databases, both for mining house names and for outside prospecting. Major coops include Alliant, Epsilon, iBehavior, Wiland and Dataline.
Cooperatives offer a great way to zero-in on multi-buyers at an attractive CPM, without paying modeling costs.
Here are some secrets for getting the most value from these coops, which are aimed at making your mail more effective. I am organizing these into three categories.
Mining House Names
After selecting the house names for your campaign, allow one of the coops to score the unselected house names. Test-mail the top-scoring segments and then expand deeper, as warranted. If you have different kinds of buyers (in Boardroom’s case, we have publication subscribers and book buyers), have the coops score the universes of unselected names separately.
Then prioritize the house names scored by the coops and that you select to mail above all other outside names that you are mailing.
Provide the coops with as much historical data about your customers as possible. Don’t be afraid to allow the coops to score very old house names. Coops have recent information about many of your old customers. We have been successful in identifying 20-year-old expires that are profitable to mail using coop data.
Finally, consider giving the coops names that you would generally not mail, such as bad debts, or TV-sold buyers, or Web-sold names. We have been able to identify many more incremental house names to mail from these previously “unmailable” universes, with the help of the coops.
Use multiple cooperatives in the same mailing. Different coops often pull different names from their models, and the duplication rate between the coops is not always as high as you would expect. We regularly use two different coops to select outside prospect names in the same mailing, and we have even used three different coops on occasion, when the product is very strong.
Fine-tune the selections on outside prospect models that are built by the coops. Gender selects might allow you to go deeper by selecting male or female names only. You can also ask the coops to optimize their models after they are built, and select only the top 50 percent or 75 percent of a segment.
If there is an outside response list that has worked for you in the past — or that has come close to working — and this list also participates in the same coop that you are using, you can ask the list owner to allow the coop to use a cross-member model for your mailing. Make sure the economics work, though. Paying a CPM to both the list owner and the coop can often price you out of the market — but this tactic is particularly effective if you are getting the names from the list owner on exchange.
Before the coops make any selections on the names that they score, make sure they suppress all of the house names that you are planning to place into the merge — both input names and suppression names. This will allow you to net out at the highest possible rate. You may have to build in an extra day into your merge-purge schedule, but it’s worth it. The biggest mistake that mailers often make is over-suppressing with the coops (telling them not to select any house names, regardless of whether or not they are being input into the merge) or under-suppressing (not having the coops suppress house names at all, in which case you will net out very low and end up paying for house names you are already mailing).
Contract with one of the coops to score your files after your merge is complete, to help identify unprofitable names that you can drop from your mailing.
Separate house and outside names when doing post-merge scoring, as you will likely drop more segments on the outside universe than you will on your house universe. Also, you have the option of dropping greater or fewer names by individual list.
Select house names that you would ordinarily not mail, and place them into your merge. Then, take the top 25 percent or 50 percent of the names after they are scored post-merge. You can also apply this strategy to outside lists. If you are already making a last 3-month hotline work on a particular list, try selecting a 4- to 6-month hotline on the same list, but only mail the top portion of the list once the names are scored after the merge. Make sure the economics work, as you will end up netting out very low. You can apply this method to mailing incremental model segments of the outside prospect names you are mailing from the coops, too. Try mailing only higher portions of a segment by dropping the lower portions after the merge.
Re-key and mail small portions of the segments of names that you are dropping, to prove that it really pays to drop these names. If the rekeyed names in segments 19 and 20 prove to be unprofitable, re-key segment 18 in the following mailing — and continue this strategy until you reach the point where it makes sense to stop. Always re-key house and outside names separately, as they will likely perform differently.