07:28:51 am on November 3, 2008 |
The Evil Lies in Properly Calculated Customer Maintenance Costs
(Part 3 of 3)
OK, on the final part of this three part series – how to calculate your customer maintenance costs. Yes, this is the longest one as far as the number of metrics to track, but if you have been following the series you know where this is going. Check out the previous two entries before you continue on this one, trust me – it makes more sense that way.
As you know, I advocate an individual approach to calculating customer acquisition costs, and customer maintenance costs. You already read that I believe that true business decisions cannot be made in a bundle. You cannot decide to cut costs for customer maintenance across the board, hurting those customers that need service but cannot get it since someone else abused it. You cannot decide to implement a specific channel or program only for a segment of your customers just because you may think that, for example, business users will benefit more. You have to make decisions based on the true costs of getting each customer, retaining and servicing them.
When it comes to customer maintenance costs, most companies just take the total they spend on customer service (by their definition) each year and divide it by the total number of customers. Simple, and ineffective. You can harbor expensive customers (constantly demanding service) as well as not properly reward inexpensive ones (they never, or rarely, require service). You can ignore budding problems, such as the excessive cost of a specific segment, tool, or channel, by not properly focusing. What is the solution? There are two things you have to consider for properly measuring customer maintenance costs.
First, the overhead costs you have for maintaining a customer service operation. This is where you can see real savings when outsourcing, for example, or by decreasing specific sunk maintenance costs (such as what happens when you optimize your KM initiative). This is the base cost per customer — not per interaction — for infrastructure deployment.
Second, I am assuming you already know your transaction costs per channel, broken down into different type of transactions (if not, as they say on TV – get them!). Monitor interactions and add those costs to each personalized customer maintenance cost as they happen – yes, that also means web self-service or similar automated transaction tools (e.g. IVR) which tend to be ignored by most organizations since they are either cheaper than regular channels, or there is no prescribed way to track users through those channels (you may even have to ask users to log-in to get service).
OK, so this gives you a basic cost per customer for maintenance. Now, the final step is to compare each personalized costs against the average per segment, channel, or transaction – and assign bonus points (or deduct from the total cost of maintenance) to those that are below the average. This will reward those customers that don’t abuse the system with increased access to different tools, technologies, or even a higher status with the organization – depending on your business decision on who gets access to what. Of course, you need to constantly (once a month would be fine, no longer) update the scores, the averages, the bonus points, and the status of each customer to ensure they receive all their benefits.
Chances are that you will have to make some small to medium changes in the way you measure and monitor your data and interactions; you will be surprised at how much certain segments or channels will cost you to maintain, and change your opinion on who is your best customer. A simple per-customer benefit calculation (revenue minus cost) will allow you to create literally segments-on-the-fly as needed, based on all new different metrics, benchmarks, and results.
Are you ready to try a new approach to rewarding good customers and bechmarking your true costs? Who knows, maybe you get to debunk the original entry for this series “a new customer is X times more expensive to acquire than to retain a current one”.
Ready to prove the myth wrong? Let me know how this goes for you…