Is Service Level a dead metric?
I recently spotted an excellent article on Call Centre Helper magazine asking ‘What are the industry standards for call centre metrics?’ https://www.callcentrehelper.com/industry-standards-metrics-125584.htm. I logged in and wasn’t at all surprised to see the first metric mentioned was the 80/20 rule for Service Level. A metric which bugs me, as from a planning perspective it is often close to impossible, before we even start.
All call centres are different – they have different staff, organised in different structures. They have different products to support and they have a different customer base. Why then do we keep trying to silo them into one bucket with respect to how quick we answer the call? Let’s step back and think about our experiences of queuing?
We go to the supermarket to do the weekly shop. When we come to pay, do we expect to be having our first item scanned through the till within 20 seconds? Or do we expect that there will be an element of queuing? We take the kids to a theme park. Does the fact we have to queue, and don’t get access to every ride reduce our enjoyment of the day? We go into the bank or Post Office. Again, we see a queue – but do we expect to be interacting with a teller within 20 seconds. In all these cases we expect that there will always, or often, be a queue. The key is whether it is what we see as manageable?
Queueing is a way of life. Yes, it would be great to have that personal assistant at our beck and call – ready to answer our every query in life within seconds. Dealing with a company by phone is a little different than in a physical person to person environment as we often have no idea how the queue is moving and have no way of estimating the wait time. But, as a society we expect to queue. Companies know this, and balance efficiency with customer service hygiene factors. It’s about knowing your customer.
Think of an extreme example. Mortgage application type calls in the UK can easily last 60 – 90 minutes. So in essence, once the front-line advisor takes the call they could easily be there for many multiples of the Service Level target. This puts a lot of risk of potential customers ringing up and very quickly going out of service. It’s not necessarily a bad service if someone waits 1-2 minutes for a call they know will take most of the morning – but 80/20 suggests it is. Even with calls that last 8 – 10 minutes or more, there is significant cost to the business in ensuring 80/20 is met (if that is the Service Level) and potentially lots of staff sitting around idle awaiting their next customer interaction. In today’s increasingly competitive world that is not sustainable.
Before returning to the original question, let me pose another question. If you were to design a report which would tell management that yesterday was a good or bad day, what would you include in it? I regularly ask this to classes and the answers are often the same – AHT, Service Level, abandon rate, customer satisfaction, Average Speed of Answer, Answer rate, speed of abandon and so on…. But think about it – some of those say the same thing! Service Level, Abandon rate, average speed of answer all tell us if the customer was able to get to the frontline advisor in a timely manner. In fact some tell us more than Service Level. It is possible for Service Level to be very low, yet abandon rate is very low and customer satisfaction still high. That is because the customer expects a reasonable wait and values the interaction more.
Some of the other metrics tell us more.You could answer every call at exactly 21 seconds, and subsequently have 0% Service Level. Would you have unhappy customers? I doubt it and most, if not all wouldn’t realise they were out of target with respect to how quickly they were answered. In fact I would suggest the customer satisfaction would be identical to answering every call at exactly 19 seconds. Therefore missing Service Level doesn’t easily tell us the whole story – abandon rates and speed of answer probably gives more insight.
Arguably the only reason Service Level is still around is to help planners with configuration of WorkForce Management software. Erlang needs it as a input to work effectively and most legacy systems still work this way. But customers expect to queue, so why not relax the 80/20 a little. It will allow us to take the pressure off and maybe even get to a point where the Service Level results tell us a story and allow us to take action in the way review of other analytics does.
Don’t get hung up on 80/20 – provide the service your customers want and expect. Find the MI that supports that and use it to maximise the service to your customers.