We have some national clients that are relatively new to applying predictive analytics models to their marketing campaigns. It is not uncommon to observe them pre-deciding markets without considering regional cultures. Consequently, their marketing efforts often excel in some markets and don’t do so well in others. How do we handle this?
The starting point is to implement a national model. The data set should include some sort of market or regional indicator (e.g. Metropolitan Statistical Area, ‘MSA’). After the results come in, the data can be parsed by the market/regional indicator and the campaign results analyzed on that basis. What does this do for you?
This approach will allow you to compare and contrast performance by market. You may find some similarities among markets. In this case, you can combine those markets for targeted treatment – starting with a predictive model built specifically for that market. Unique markets that do not act like others get “teased” out. Strategic decisions can then be made as to the feasibility of giving those unique markets special treatment. For example, using this approach for a national client, we were able to tease out underperforming markets that included smaller MSAs in Texas such as Austin and Portland, OR.
By taking this approach you can use your marketing budget most efficiently while continuously refining your targeted marketing for maximum performance.