Think about the danger of not having access to your customer data – what if the histograms that you produced in the early part of the year, prove to be fine for a time, but you’d like to compare the effectiveness of this model against one that your team put forth.  You’d have to wait 6+ months to aggregate enough data for the second model to show itself worthy, but not if you’d have simply kept access to the data. 

 

For smaller companies, however, this may be a good option – but only if there’s a high hacking risk and lack of technology support internally.  It’s better to reduce risk at that point than to save data for a rainy day – if you don’t have a current plan of using your customer data for marketing activities in the next 6 months, then it will not likely be of use anyway.