Idle: a customer that was recently active and unless we do something will soon churn

We had one box for black marbles and another one for white marbles. And everything was great until we found a grey marble. It was then that we realised that not all marbles in our collection had the exact same black or white colour. So, we added a third box to better classify our marbles.

Active customers. This is perhaps the most commonly adopted and reported metric and yet there is no single definition for it. In banking an active customer may be defined as any account holder with at least one completed transaction in the last 2 or 3 months. In telcos and insurance companies on the other hand, an active customer can be defined as anyone with a contract that has not reached its stated time yet or did so in the last 30 days. In SaaS companies, it is usually any account with a paid subscription that has not yet expired whereas most PaaS companies need to employ tailor made definitions.

Moving to churned customers now which is also a very common and important metric. In this case however, it is much easier to come up with a general definition, one that could look like the following.

churned: a customer that used to be active but is no more

Some really important decisions are made based on how figures for active and churned customers evolve over time. This is why I incorporate a new status of customer that lies between active and churned customers; one that can improve decision making, especially when it comes to customer retention. This is the idle customer.

idle: a customer that was recently active and unless we do something will soon churn

The idea is simple. When a customer is no longer active, we wait for a short period before declaring him churned. In the meanwhile, he is considered idle. In order to determine how short (or how long) this period should be, we need to combine some business parameters like:

  • The segment of the customer - in example a loyal customer can be given some “time off” more easily.
  • The probability to reactivate and the typical time to do so - although this can also vary by segment, it has to do with how a segment behaves after churn rather than before that. So, if a significant portion of churned customers gets reactivated after a few days, it is a good indiction that we should wait these few days before declaring them churned.
  • How frequently and with what means we have the intention and the capacity to contact our users - in example, if we do not want to send messages to our users less that 3 days apart, then the idle period could be at least 3 days long. If on the other hand, we estimate that it takes at least 2 weeks for one of our users to fully move into using a competitor’s services, then it is also clear that the idle period should be less than 2 weeks.
  • The time it takes to resolve known payment issues - basically, if it takes up to 1 week for the majority of financial transactions to be validated, then we should give our users the benefit of doubt for at least that long.
There are 3 direct benefits by introducing the notion of idle customer:
1 You get to understand your customers better
2 You get to worry about the real deal of churn
3 You get to allocate your budget in a more targeted and therefore more efficient manner

Let us elaborate a bit.

  1. We had always had customers that were not really active but should not be considered churned - in example a SaaS customer that purchased an annual subscription 6 months ago but has not had any account activity in the last 3 months. We had also had customers that were not really churned although not exactly active - a very loyal PaaS user that for various reasons did not bring in any revenue in the last week or month. For these cases we either compromise or define exceptions. The idle customer status breaks the active/churned duality and allows us to define active and churned customers based on what they should be and on that alone.
  2. An active customer goes idle and then either goes back to active or goes churn. This leads to lower and more realistic reported churn figures. Don’t take my word for it. Have a look at your data: how does the average time to reactivate compares to the churn definition you use?
  3. An increasing number of idle customer it is still not a good sign. However, idle customers can be targeted with lower or even zero cost actions which means that we can keep the more expensive means for those that actually “make it to churn”.

So, would you add idle status in your customer view?