Profiling

Customer profiling provides insight on a customer’s identity through demographics and lifestyle information.  To conduct a profiling analysis, a match of a customer file to Telematch’s database is required.  The customer data is then tabulated to define relative distributions of demographic and lifestyle attributes among the profiled population.  This analysis can be conducted against an entire database, or by key customer segments.

Profiling is the first step in the usage of demographic and lifestyle information to augment the marketing process.  With a customer profile in hand, more informed decisions regarding the use of such information in customer communications are possible.

Customer File Segmentation 

There are many reasons to segment a customer file, the reasons range from: differential treatments based on corporate policy,  identifying affinity groups, defining custom communication groups for CRM.  Segmentation utilizes customer behavior data and appended demographic attributes. Segment definitions are defined via multivariate analysis techniques: cluster, factor, or discriminant analysis. Once defined segments represent a homogeneous customer population.  Models developed for homogeneous segments have less variation due to the differences between customers and are therefore more predictive.

Response Modeling 

Thorough knowledge of customer data is the most important element in the creation of powerful models.  This in depth understanding of the relationships in your data which can be leveraged in predictive models.  Our unique research techniques require seeing your business through your customer data.  We employ sound statistical methods to ensure all models will discriminate in roll-out applications. In addition, we develop models that will remain predictive over time.

Acquisition Response Models  Response models require information on responders and non-responders to your direct mail or telephone offer in addition to a "snap shot" of their attributes at the point-in-time the offer was extended. These models provide a means of discriminating between prospects that will and will not accept your offer. The benefit is better targeting of your best prospects.  The drawback is a longer lead time to read results and apply in a roll-out; however, a response model will always outperform a clone.

Clone or Look-a-Like Models 

These models work on large compiled files.  By supplying your ideal customers, a clone model defines their attributes and identifies replicates in the compiled file.  A key benefit of a clone model is the speed with which you can roll-out and gauge its success.  One drawback is the possibility of not identifying new markets for your product since you are basing the analysis on current customers. In addition, care must be taken when defining the ideal customers for cloning.  If proper names are not selected, the strength of the model may weaken in a roll-out.

Performance Models 

Performance models are utilized to predict a customer’s payment behavior.  Performance models may be employed to target payers or to identify non-payers depending on the need. For direct marketers utilizing a soft offer, performance models can keep bad debt to a minimum.  Telematch’s database of demographic attributes can be used alone in acquisition efforts, or in concert with customer behavior data in cross sell promotions to target offers and limit bad debt.

Assessment of compiled list and enhancement data  Knowing which enhancement data to bring on file is a complex proposition.  We have the experience to evaluate outside demographic/psychographic data based on its ROI.  We use proven techniques to assess how such data can improve your customer segmentation or models by running analyses with and without such data.  The difference in these two analyses, projected over the life of the contract, forms the basis of the ROI calculation.  To evaluate the true strength of such data, we employ various unique multivariate statistical techniques.