Advice for Data Analytics Implementations

A recent article by Mitchell (2013) discussed the top twelve mistakes organizations make when implementing a predictive analyics project.  The mistakes are as follows:

1. Begin without the end in mind
2. Define the project around a foundation that your data cannot support
3. Don't proceed until your data is the best it can be
4. When reviewing data quality, don't bother to take out the garbage
5. Use data from the future to predict the future
6. Don't just proceed, but rush the process because you know your data is perfect
7. Start big, with a high-profile project
8. Ignore the subject matter experts when building your model
9. Just assume that keepers of the data will be fully on board and cooperative
10. If you build it they will come: don't worry about how to serve it up
11. If the results look obvious, throw out the model
12. Don't define clearly and precisely within the business context what the models are supposed to be doing

Reading the article and looking at these twelve mistakes, it is clear the same considerations we give most information system implementations apply to analytics projects too.  We need to have a plan, involve the end users, practice iterative implementations, ensure we are using and applying good data, use risk-mitigating staged or pilot implementations, and evaluate our results.

I still think this is a very good article but we need to keep in mind that this advice is the same type of advice we should heed for all of our system projects.  If we apply the same mature IT implementation practices to our analytics projects, we will have an increased rate of success.  The article provides us with a good reminder of what we should do in all of our projects.