Learn how mailers are successfully improving their acquisition results by dropping names that are unlikely to respond - and filling the void with more ideal, higher scoring prospects.
Most cooperative databases (and many data providers) offer Merge Optimization as a tool to help eliminate less responsive names from their mail campaigns. This is done by applying a predictive model to acquisition lists after the merge/purge process. Let’s break the process down into 4 steps:
1) Send the file: After the merge is complete, the mailer sends the entire mail file to the database company.
2) Apply the model: A model is applied to the mail file and the names are ranked from most to least likely to respond. The model used is typically a response model, built on prior mail campaigns (a Mail Match Model).
3) Select the names to drop: The mailer decides which of the segments to drop from the campaign and how many names, depending on the budget and goals for that campaign. The provider typically charges a per thousand fee for the names dropped.
4) Add more names: In order to maintain the budgeted mail quantity, the mailer must either order more names from their outside list sources upfront, or fill in the balance with additional names from the data provider after the fact.
In the second case – a model is built on the “balance” of the database, after suppressing the mailer’s house names, the mail file, and any other acquisition names that have been supplied to the mailer. Hence, the name: Balance Model.
The Balance Model typically doesn’t perform as well as an acquisition model, but should perform better than the names that were dropped. The goal is for the overall campaign to outperform what would have been mailed.
Keep these key points in mind when doing this kind of modeling:
While many mailers are content to simply eliminate the dregs of the campaign, some are having so much success with Balance Modeling that it has become an integral part of the campaign. If it works for you and you decide to continue, make sure your models are continuously refreshed because they can fatigue rather quickly.
What to Know Before You Model: The 4 Most Commonly Used Models for Acquisition