A major problem in the health club industry is customer retention – it may well be the industry’s single largest issue. Hence the constant aggressive push to get members signed up and in the front door, at a rate faster than they are exiting out the back door. I have seen figures showing that as many as 40% of members churn in the average health club, regardless of the exact numbers, it is a known fact in the industry that it is a higher number than any health club manager wants it to be; and obviously any reduction adds directly to the club’s bottom line.
Equally plenty of members renew their memberships year in, year out. Accordingly, any member retention strategy should involve two key components: 1) identifying those members at risk of leaving and 2) targeting those at risk with appropriate interventions.
It is beyond the scope of this article to go into intervention methods. However, I will address the identification of members at risk of terminating their memberships (‘at risk’ members) – and how predictive analytics can be applied to help with this.
Like all businesses health clubs have limited resources, and it is absolutely pointless for a club to invest resources to try and retain each and every member, when a good deal of them are not at risk in the first place. If a member is identified as ‘at risk’ there is a strong business case to be built around investing resources in trying to retain that specific member (theoretically you could afford to invest up to $1 less than the cost of acquiring a new member, and still be ahead of the game), conversely if they are not ‘at risk’ and are going to re-sign anyway, you may just as well burn the money as hand it over to that specific individual in the form of an incentive or time invested.
The other consideration is, it is far easier to pro-actively try to retain 2,000 members than 4,000 member, so by segmenting, and making the size of the task more manageable, it increases the likelihood that a health club will do something – and if we know nothing else, we know that doing something is usually better than doing nothing.
So we have a clear business case for identifying which members are most at risk of churning. Our next mission then, would be to take our database of current members and identify which ones specifically are ‘at risk’ and which ones are ‘loyal’. Ideally we would take it one step further than this, and be able to rank our whole customer database in rank order from those statistically ‘most at risk’ to those ‘least at risk’. The benefit of doing this, is that it provides our sales/retention staff with a sequenced work list, which they would start at the top of and work their way down sequentially. This simple act in itself would give us comfort that our resources are being focused on those that most require them – a form of retention triage if you will. This can even be taken one step further, and we can – again using statistical methods – determine the statistically optimal place in the list to stop.
Though we have a business case, and a reasonably clear vision of what would be useful, the problem is that for the managers of most health clubs, the scenario outlined above is closer to science fiction, than something they perceive they can practically deploy within their club. So the status quo prevails: 1) do nothing, 2) treat all customers as equally at risk, or 3) perform some random haphazard interventions with no real science behind who is targeted and who is not.
So to get to the point of execution, and movement from theory to reality, let’s discuss how we would take this utopian vision and turn it into an actionable reality. Ironically for many health clubs this vision can be actualized faster than it took me to write this article – literally.
Most health clubs have a reasonable amount of data on their members. Let’s imagine that we have all the data about every member of our club for the last five years, lined up in an Excel spreadsheet. Every row is a unique member, every column is the information we know about that member. The columns we call input columns as they are the inputs that help us make our prediction about that persons future behaviour, these would contain things such as: her age, her marital status, change of marital status, # of visits in January 2010, number of visits in January 2009, etc. payment method, # of address changes, average time she spends in health club, etc, etc it would be no problem to have 100 or even 500 columns, and in the very last column (our target column) we add a label ‘loyal’ or ‘at risk’. Anybody that terminated their membership previously is labeled ‘at risk’ and ‘anybody’ who re-signed is labeled as ‘loyal’. We would eliminate from the spreadsheet anyone who had not had been with us a year yet, as we don’t have any conclusive information about their behaviours.
Now I will skip over the math here, which nobody would want to try at home, but you can take it on good authority that there are patterns within all the input columns that can help to predict the customers propensity to churn. This is as you would well expect, for example prior to terminating a membership, a member may start coming in less frequently, and if this data is recorded this would show up, or a change in marital status may impact an individuals propensity to re-sign, and most likely it is an aggregation of many factors. Typically a human cannot detect these patterns, but there are software applications that can, and once the patterns are defined, the software can look at the patterns in an unseen group of members and make a prediction as to each individuals propensity to churn, and then output these members in a sequenced list as described previously, complete with the optimal point in the list to stop making interventions.
To explain it a slightly different way, we are: 1) consolidating historical data about behaviours that we think may be correlated to an individual churning from historical members 2) we are letting software examine that data for patterns and how they relate to how a member churned or did not 3) that relationship is frozen in a ‘predictive model’, and finally 4) the model is applied to unseen members to statistically predict their behaviour (vis a vis churning or not).
I would encourage anybody interested to visit http://www.11AntsAnalytics.com and watch the 11Ants Model Builder QuickStart tutorial video, which will better show the process (the data is different, but it won’t require much imagination for it all to make perfect sense). Feel free to email me if you have questions about this – doing this sort of thing is ten times easier than most people imagine.