The way most contact centres measure performance is wrong. I’ve met a lot of managers who argue that they use KPIs that benefit the customer, or that they measure customer satisfaction, while neither is actually the case.
Things have changed in the past 10 years in customer service, and how we treat our customers has now become the main differentiator in most industries’ competitive landscapes. To excel from a business standpoint, you must be customer-centric. In order to do that, it has become necessary for a lot of contact centres to revisit which KPIs you measure and a why. That’s not to say that most KPIs are useless, in fact the well-known ones have their uses, but I do think that you’ll most likely improve your customer experience if you rethink them.
When deciding which KPIs to measure, many are decided upon from the point-of-view of the company and don’t reflect actual customer satisfaction. We all need to be more conscious and conscientious about our use of KPIs; which ones do you measure and how do you use them in communication and reporting to the rest of your organisation? Which ones are the best indicators of productivity and performance and which ones are the best indicators of customer satisfaction?
In the following article, I’ll go through each of the commonly measured KPIs and have a closer look at some pitfalls and fallacies, specificaly in regards to their relationship with customer satisfaction.
Is answering 80% of calls within 30 seconds good customer service? Not necessarily. Is it a reflection of customer satisfaction? In my experience, only very rarely will it correlate well with customer satisfaction. What is it then? It’s mainly an internal KPI to measure the performance of your contact centre. For that purpose, and that purpose alone, measuring and following up on your service level is definitely a good thing, but don't make the mistake of thinking you're doing so for the customer’s sake.
I’ve seen many contact centres make strategic, tactical and operational dispositions based almost solely on service level results in conjunction with number of contacts and cost per contact. This paradigm is exactly what now leads to bad customer service - something that most businesses can no longer afford as customers become less and less brand loyal.
Like I said, most KPIs have their fair uses, and service level does as well. One case is when service level drops dramatically. When that happens either one of two things (or both) has happened:
- You had fewer agents available than needed - due to workforce planning or employees not showing up for work, whatever their reason.
- You had more customers contact you than expected.
If it's the latter, you will be the first in line to collect valuable feedback. Taking that part of your job very seriously is a great way to improve not just customer service (and the cost of it) but also other, crucial parts of your business and product. Exactly how to go about this task could easily take up as much space as this article, so in the interest of rethinking KPIs, I’ll move on with a promise to return to this subject later.
Average Waiting Time
When looking at Average Waiting Time (AWT), firstly, you won’t have an adequate measure of AWT unless you ask your customers if they experienced the waiting time as 'short', 'adequate', 'long' or 'too long'. Most companies do not ask their customers this question and, sadly, without customer feedback, AWT doesn’t reflect any customer value in and of itself.
One of the reasons for this is that local culture and the type of product or service you sell create very different expectations. In the Nordics - I’m from Denmark - there’s a big difference between expectations for an acceptable waiting time amongst customers in Denmark and Sweden, in spite of the fact that we’re neighbours and so alike in many ways that most other nationalities can’t tell us apart.
I once did a customer survey in which we asked our customers how they experienced the waiting time on the phone to get through to customers service. In Denmark, there was an average waiting time of 30 seconds and the majority of the respondents felt the waiting time was long or too long. In Sweden, there was an average waiting time of two minutes and the majority of respondents felt the waiting time was adequate. Similarly, different products carry different expectations for an acceptable waiting time: as a rule of thumb the more complex the product and the higher the perceived value, the longer wait time people will find acceptable.
Personally, I have almost always worked with a goal of answering 80% of all offered calls within 30 seconds (for telephony). The rationale behind that goal being that a short time to answer equals higher customer satisfaction. I have since learned that this is not always true: in one case, my department went from an AWT of 30 seconds to an AWT of 2 minutes without any noticeable change in customer satisfaction.
As it turns out there are factors at play when reducing wait time that are much more valuable than just a short AWT. One of them is routing customers to a person that is able to solve their issue with little effort as quickly as possible. If given the choice, most customers would rather wait longer and get their issue resolved at their first contact than get through quickly to an agent who is unable to resolve their issue. And that’s very well reflected in customer satisfaction. I guess the lessons here are:
- Never compromise your first-contact-resolution rate on behalf of a lower AWT, at least not if you value customer satisfaction.
- Don’t standardise your AWT across different cultures and products/services.
Answering a call from a customer directly creates customer value. An abandoned call means that a customer gave up and that’s always a bad experience for the customer. It’s obviously also bad for the company, since it can be a missed opportunity to make a sale or retain said customer.
Most contact centres measure lost calls, abandoned calls or abandon rate. While not necessarily wrong, this places the focus on something negative and for that very reason I’ll advise you to do the opposite: if your goal is to have a maximum of 5% abandoned calls, it’s fundamentally easier to celebrate that you have answered 95.42% of offered calls than it is to celebrate that you’ve only lost 4.58% of offered calls.
A thought: have you ever called a customer service and hung up before you got through? Did they call you back and apologise and ask you what you needed help with? No? Would you have been positively surprised if they did? I have never experienced this and would definitely be surprised by this kind of proactive service.
Therefore, I recommend you follow up on abandoned calls and call the customers back as soon as possible. However, many customers who hang up call back at a later time. If your technology supports it, you should be able to see if the customer successfully called you back later before making an outbound call to her or him. An abandoned call is essentially a missed opportunity and a great way to show respect for your customers' time by you taking the time to call them back and asking them what they need help with.
To summarise: measure percent answered calls - not abandoned calls - and start calling back customers who abandoned their call as soon as you can.
Handling long wait times: time limit, callback and voicemail
Many contact centres have periods in which they receive a lot of calls at the same time. This usually means longer queues and higher waiting time.
To give your customers a better experience, you can set a time limit for how long your customers are allowed to wait. When that limit is exceeded, you should offer the customer the option for: a) a callback where they don't lose their place in the queue; b) leave a voicemail; c) forward the call to a different queue or an external number; or d) play an announcement saying you are unavailable and ask them to call you back later. Option ‘d’ is of course horrible, but it can often be preferable to the individual if it’s coupled with information about when you’re easier to reach. In essence: let the customer make an informed decision.
A quick note on how to handle voicemail: don’t wait until tomorrow morning to listen to it. If at all possible, get technology that supports delivering the voicemail to the next available agent and then call back that customer as soon as you are able.
Part 2: FCR, NPS and more
Since this turned out a bit longer than I expected, I'll divide this topic into two parts. In part two, which is coming in a few days, I’ll discuss the rest of the important KPIs and round out it off with my final conclusions.
Meanwhile, you can always find me here or on Twitter for any kind of discussion about customer service & customer experience. Also, since this is my first blog post on MyCustomer, go easy on me, but don't let it stop you from joining the discussion!
About Tue Søttrup
Tue Søttrup has been delivering excellent customer service for more than 20 years, the past 12 years as a manager for several contact centers. Tue is currently VP of Product & Customer Success at Dixa where he's carefully nurturing a small bonsai tree that he calls GenZen.