From telecoms to finance, e-commerce to government, a predictive model is being utilized across various sectors to tackle all kinds of business problems. Companies that have yet to benefit from this practice need to examine the ways in which they can do so.
For example, AccuData worked with an ad agency to help its client, a nationally known health and fitness franchisee, increase response rates by over 100%. Their challenge: demographics alone revealed little insight.
With more than 700 stores, they wanted to partner with a direct marketing agency that could provide actual marketing intelligence – not just postcards and lists. The project began with a simple customer profile request as part of an effort to increase response rates. The client chose to work with a savvy, web-to-print direct marketing agency since they had a reputation for excellent creative strategy and back-end analysis.
The direct marketing agency recommended that the client create a descriptive clone model for each of its fitness stores. However, while this would help paint a picture of the various markets, demographic elements and relational penetration, it would not solve the problem of “how do we increase response?”
Matters became even more confusing when they got the results of initial tests between saturation data and scored data from the customer profile. It appeared that scored data actually offered no measurable lift in response – on average, profiled records responded virtually the same as saturation records. The reseller wondered, that perhaps the fitness stores were placed in market areas where immediate surrounding geography was populated by demographics very similar to those on the modeled client file.
That’s where AccuData came in. We provided an actual solution: Using a predictive model to uncover their prime prospects. While demographic elements looked very similar between the profiled records and those saturating the geography of the store sites, we wanted to identify key differentiators.
By implementing a predictive model process, which looked at over 200 individual demographic elements and compared responders verses non-responders, the resulting scored data became a tremendous resource for the client. They learned exactly which demographic traits – such as credit card usage, age, presence of children in the household, and interest in travel – were positive influences on the probability of an individual responding to an offer.
To further improve results, they also applied prospect suppression details – removing from the list non-responders and responders with low propensity to buy. By using a predictive model, the client was able to double their response rates and reduce the costs of their prospect acquisition marketing campaign.
A predictive model will find the statistical differences between responders and non-responders whereas a profile will only give you insight into your customers without regard to those that are not and will not be a customer. So you end up marketing to people that meet a general profile but statistically have a low probability to respond.