Database Marketing Questions and Responses

¶ … dropping" is actually occurring. This process will involve reviewing 24 months of historical data. A trend analysis will confirm if the downturn in profit is abnormal. The second step is to break down the sources of profit in order to understand which subgroups are increasing and which are decreasing. When the subgroups that have shown decreasing profits have been identified, an examination will be done to determine the causal factors.

Some possible causal factors that will be explored are overall quantity, margins, cycle time, total balance per account, and the acquisition cost. The most likely factor swill be explored first. If the evidence is inconclusive with respect to the most likely factors, then other factors will be investigated. These secondary potential factors include delinquency rates, default rates, and finally penalties or other unusual expenses, such as those pertaining to regulatory issues. Scenario #2. A reconciliation process between the two reports can be used to try to determine the differences between them.

The first step is to check the logic of the two reports, to ensure that the databases both have the same filtering conditions applied. It must also be determined whether the two databases have the same definition of "new checking." Differences in either of these areas will naturally lead to deviations between the two reports. It must also be determined if there are differences between the source data of the two reports.

It is possible, for example, that one report uses real time data while the other has a time lag. In order to prevent this problem from recurring, we can make a note of the underlying logic, which will help to shift the customers' expectations. Another solution might be to merge the two reports, especially if they are used by the same audience.

This can be done either by adding and combining all columns into one template, or by sourcing the data from a single source, and ensure that the data is refreshed at the same date/time. Scenario #3. The first step in determining which customers to target is to define a list of segments that are meeting the budget and goal. First, a sufficient set of reference data should be gathered to run a back-test. This can then be used to set the breakdown of the target market.

A certain percentage of the offer should be made to both the average income and the affluent families with a given demographic variable (such as zip code, income to debit ratio or age bucket). If there are multiple options from which to choose, risks for each should be identified. Choosing between the different options should be a group decision based on the risk and potential return characteristics associated with each target segment. Scenario #4. The first consideration is how many customers have signed up to receive email from the bank.

Within that pool of email addresses, how many of those are legitimate, functioning, current emails. The second step is to use that data to set up the objectives for success. The benchmark can be a traditional campaign. We should use the responses per 1000 for a traditional campaign as a benchmark for the email campaign. For example, if a mail campaign yielded a 10% response rate.