Jan
19
2011

Models Have a Shelf Life

Marketers sometimes forget that analytical models have a shelf life. Generally, marketing teams build and deploy models as needs arise. But as internal and external influencing factors change, the effectiveness of a model changes — and it may no longer drive the results that were initially intended.

So if model effectiveness changes, what can be done about it? The answer is simple; review each model on an ongoing basis, including periodic checkups or more in-depth audits as needed. 

 Review each model on an ongoing basis, including periodic checkups or more in-depth audits as needed.

At minimum, the following 4 questions must be answered for each model:

  1. How is the model holding up? 
  2. Is the underlying data changing?
  3. Is the objective / purpose of the model still a priority?
  4. What other priorities are emerging?

These audits, checkups and general maintenance routines could be done in-house or by working with an external partner. Regardless of the approach chosen the effort requires focus, discipline and effort.

That addresses existing models, but what about creating new models so that they last longer? Building the right model is both a science and an art, and building for the long haul requires you to keep the long term view in mind. Doing so will extend the shelf life of the model and ensure freshness post-deployment.

Have questions about analytical models? Feel free to ask the author in the comments section below.


Comment are closed.