Approaching the customer journey through more efficient statistical models

From marketing and origination to customer management and collections it is now more critical than ever to provide the best quality of service to your clients. Statistical models have been proven to help organisations make decisions based on predictions across the customer life cycle. Building these models, however, can be very challenging from both a resource and project management perspective.

If companies cannot adapt their decisions promptly to take into account changes in their target market or business processes then there can be a negative financial and regulatory impact.

Getting the best data available

All standout statistical models are built on strong and relevant data so before you start to think about modelling, it is necessary to extract and verify the relevant data. This can be a difficult and lengthy process in itself, depending on how the historic data is stored and the infrastructure of the organisation. It is also worth considering if information from external sources could add benefit to the model development, or help make it more predictive.

Designing the modelling sample

Once the modelling team has acquired the data, they need to fully understand it and design a development sample. This is typically the most challenging part of the whole model development process. The analysis to determine the best sample needs to take into account several factors. The data needs to be robust, representative of the current or future target portfolio, and be as predictive as possible.

Building the model

The ideal scenario is where there is a lot of data available to build linear or logistic regression models. But it is possible to adopt alternative approaches on a case per case basis. Regardless of the modelling methodology employed, the key here is to build the most predictive, accurate models whilst ensuring that the results do not disregard business logic or objectives.

Experian’s modelling teams have access to high quality data and the experience and skills to use several different modelling methodologies. They can ensure that your models are as predictive as possible to help you meet your business objectives.

To find out more on how statistical models can help you offer the best customer journey, please contact our analytics team on 0844 481 5638 or email decisionanalyticsUK@uk.experian.com.