Prospecting With Database Technology
What is marginal list enrichment? MLE refers to the application of statistical modeling to individual prospect lists. Why rely on one or two variables (e.g., recency and/or dollar select) when you can use a model that will statistically blend all of the relevant data and generate a score to predict who will respond to your catalog? The model will create a more powerful select that usually will beat traditional selects in head-to-head testing.
Broadly speaking, MLE refers to the application of any statistical tool intended to improve a select on a marginal list. However, two specific techniques are most common:
· Enriched (optimized): selects from cooperative databases.
· Custom: mirror-image modeling of a specific file.
Five cooperative databases offer selects that use enrichment tools:
· Experian Z24, enhanced list rental.
· Prefer, cross-member optimization.
· Abacus, cross-member modeling, list optimization.
· NextAction, optimization.
· I-Behavior, optimization (limited application).
Most providers that offer custom modeling services can develop mirror-image models using the list owner's transaction data. They often overlay compiled data to let other variables enter the models. Generally, the richness of the transaction data from the cooperative database provides a larger lift over traditional selects than a mirror-image model.
The majority of catalogers' new names comes from two sources: outside list rental and cooperative databases. MLE is in essence a hybrid combining the best of both sources.
The economics of MLE. Pricing for an enriched select using the cooperative databases varies based on the database. Prices range from $20/M (Z24) to $60/M (Abacus). Considering that the mailer is typically attacking universes that have been unreachable, the income to the list owner will be incremental. As such, the broker often can negotiate a discount that will cover all or part of the enrichment costs from the database. The broker/mailer's goal should be to pay no more than you typically would for a six-month select with two selection charges.
For example, assume the following for a straight select:
Base list price $100/M
2 selects 3 mos/$100 AOV $20/M
Now the enriched select with a negotiated lower incremental rate:
Base price $80/M
Enrichment select charge $40/M
The cost of the enriched select equals a straight select but typically performs 10 percent to 30 percent better. In addition, selects taken from a database will be delivered net of your house file, thus the effective cost can be considerably less. Better response coupled with lower net cost will significantly lower the cost per name. Previously marginal segments now will rise above your cutoff.
The pricing structure for mirror-image modeling varies among modeling providers. Some require a fixed fee, but others are willing to develop the model for free and charge a cost per thousand on the back end. Generally, avoiding fixed fees and keeping the risk on the provider is the preferred approach. But if chances of success are great, you may choose to pay a fixed modeling charge and minimize ongoing run charges.
The MLE process. Your list broker should have the knowledge and expertise to make enriched recommendations and execute the approved tests. The broker should follow three basic steps:
1. Identify marginal list/segment candidates for enrichment. Take a detailed look at the past two years of prospecting results and identify candidates. Plotting response graphically often illuminates enrichment candidates. There are three main applications for enriched selects:
· Marginal list. Previous tests were close but below cutoff. Retest the list with an enriched select.
· Deeper select. Currently taking a select (e.g., three month or hotlines) but unsuccessful in deeper selects. Test an enriched select of deeper segments.
· Seasonal list. A list works only in holiday. Try an enriched select in spring.
As a rule, for modeling to be effective, the list chosen for enrichment should be 150,000 or larger. Also, do not choose lists that previously failed miserably. Enrichment is a great tool and can improve response, but it's not magic. If a list needs more than a 50 percent increase to be mailable, it is unlikely enrichment can make it a winner.
2. Select an enrichment tool. To use a particular database to enrich a list, both the mailer and list owner need to participate in the database. Pick a cooperative database in which both you and the owner participate. Your broker can guide the selection by researching what has worked in similar situations and will consider the pricing of the various tools.
3. Define specific select criteria. Target a desired universe population and engineer the specific criteria to hit that target. Unlike normal selects, where counts are fixed, the model scoring is more dynamic and depth can be adjusted to match a desired universe size. Keep targeted universe size relatively small on initial tests to increase chances of success. If the test is a winner, test deeper into the model at the next opportunity. Your list broker should have the expertise to work with the database provider to define the criteria of the select and place the order.
MLE as a source of incremental list income. The focus so far has been using MLE to open new prospect universes. It is also an excellent tool for generating additional list income. Your list manager can identify opportunities for additional use of your file by offering enriched selects to non-users and existing users. The price/M of the additional volume may be lower than your list has averaged, but it is incremental. It is new money to add to your bottom line.
In summary, applying marginal list enrichment techniques is an excellent way to:
· Expand your effective prospect universe.
· Improve performance of prospecting and reactivation.
· Generate incremental list income.