Implementing appropriate Contact Centre Wait Time Metrics

In this short article I talk about implementing appropriate Contact Centre Wait Time Metrics and strategies.

It’s not about the know-how – it’s about implementing that know-how

I’ve been teaching advanced Contact Centre operations for nearly 20 years.  And there’s a lot of knowledge to pick up.

So I thought it might be helpful to share some implementation tips to help Contact Centres take their know-how and bring it to life.  Because at the end of the day it’s not the know-how – it’s the implementation.

In this short article I’ll address Contact Centre Wait Time metrics and strategies.

In future articles I’ll address other aspects of operational implementation including KPI selection, forecasting and quality.  I can envision a mini-series here.

One other note – this article is written for those who’ve been through a rigorous operations course.  So I’m speaking directly to those equipped with the know-how.  If you’ve not been through a rigorous operations course I hope you find this article helpful to some degree.

And if you find yourself saying you don’t agree, or you don’t understand – that’s usually because of a lack of know-how, not a lack of intelligence, passion or desire.

In this industry we simply don’t know what we don’t know if we don’t make a concerted effort to fill the gaps in our know-how.

Now on we go.

Wait Time Metrics

For Customer contacts that you’ll handle in 60 minutes or less you’ll use Service Level.

For Customer contacts that you’ll handle after 60 minutes you’ll use Response Time.

Because of the confusion around Service Level and related Wait Time metrics let’s hold off on Response Time measurements for a future article.

As operations experts know – the way you define, measure and plan for performance between Service Level & Response Time contacts is completely different and it doesn’t do either justice to ‘mix them together’.


Service Level

1. Ok first things first – set your Service Level

You’ve got to set a Service Level for your Centre.  No Service Level?  Then it’s time to set one.  Because you can’t plan and staff to a moving or non-existent target.

If you’re using an ACD you’ve got to be using Service Level – it’s that simple.

Remember there is not an industry standard for Service Level.

What works for Organization A is not going to work for Organization B.

Even within a single organization you may find up to a dozen different Service Level objectives depending on the nature of that Customer queue and the types of Contacts handled.

Because Service Level is a major driver of your labor budget, it is typically reviewed annually – at the normal annual budget cycle.  Avoid changing your Service Level frequently – that will make life difficult for everyone.

Annual reviews – with the right mix of senior Participants – works well.


2. Ok – now decide what interval you’re going to use for measurement

Service Level performance is always measured on an interval basis.

Typically at 30 minute intervals but if your contacts are long then at the 60 minute interval basis.

If you have a very small Centre or Queue – with just a few Agents – an hourly or even shift basis may be enough for you.

But if you have a larger Centre or Queue  you should be measuring to the 30 minute interval. Very large Centres and those that pursue a significant cost efficient strategy measure down to the 15 minute interval.

If you’re only reporting a daily or weekly average – for example to give to the bosses – you’re going to have to add an additional set of internal measurements & reporting around intervals.

Because the Customer Wait Time experience, the Agent Occupancy experience and your best bet for cost efficiency live within interval management – not with daily or weekly (or heaven forbid) monthly averages.

Some Centres use a green/amber/red system to indicate interval performance across a day.

For example, you have run a 24 hour operation and you measure down to the half hourly interval you have 48 intervals to ‘get right’.

So define what a green interval looks like.  For example Service Level performance of 90% or above = green.  And carry on that logic for amber and red intervals.

Imagine how easy it is to look at a color coded representation of your Service Level performance that day by interval.  How many greens?  Ambers?  Reds?  Where do patterns emerge?  Because you can only fix what you can find.

And use cool graphical representations of Service Level performance to share with everybody in the Centre. It takes everyone in the boat, rowing in the same direction, to achieve Service Level.

It shouldn’t be a secret – it should be displayed everywhere.


3.  Choose a Service Level calculation formula

A lot of folks don’t realize that there are at least 4 different formulas out there to calculate your actual Service Level performance.

And that this formula has been input into your ACD.

So the question is always this.  Do you know what your formula is?  Are you happy with it?  Is it consistent across the organization to allow for some level of apples to apples comparison?

It’s an important decision.

Average Speed of Answer

Are you using Average Speed of Answer?  If so, why?

If you have an ACD, Service Level is the best and key metric to measure the Customer Wait Time experience.

Leading Centres don’t use ASA because it is an outcome of Service Level performance.  If Service Level goes down where does ASA go?  It goes up!

Learn to chase drivers – not outcomes.  Fix a driver and you automatically fix an outcome.

Where does ASA come into play?

Well it’s in the Erlang B calculation to calculate how many trunk lines you need for your Centre.  But that’s automatic.

And admittedly it’s easier to graph using ASA vs. Service Level.  Simply run Erlang C for your desired Service Level, identify the outcome ASA figure (let’s say 12.7 seconds) and use that for graphing.  It’s ‘equivalent’.

But if you have an ACD and you’re using ASA as a target or important metric stop, pause and ask yourself why.

It’s very ‘Jurassic Park’ and was in use in the days before ACDs were commonly installed.


Longest Wait Time

One Customer is going to experience a ‘Longest Wait Time’ even when you’re achieving your interval based Service Level.

Firstly – everyone should know what that is.  Because even if you’re achieving your 80/20 or 90/10 or 50/40 someone is going to wait a long time.

There are often widely different wait time experiences for Customers within a single half hour interval.  It’s good to be aware of that.  Particularly if you survey Customers on their wait time experience.

Lastly – and perhaps most importantly – whoever is reading your Readerboard and making decisions based on that data MUST know the Longest Wait Time for that interval.

The Longest Wait Time is an important piece of data for interpreting your Readerboard and making smart Service Level recovery decisions.

Fortunately Erlang C helps us calculate that piece of data for each and every interval.

And just like ASA, Longest Wait Time is an outcome of Service Level.  If your Service Level goes down your Longest Wait Time goes up.

But as I mentioned before – don’t chase outcomes, find and fix the driver(s) – in this case Service Level.


Abandonment Rate

There’s a lot of confusion around this metric.

Clearly, Centres that generate revenue – such as food orders or hotel reservations – care a lot their Abandonment Rate.  But they address that ‘care’ by setting very high Service Level objectives.  That means that Customers get answered so quickly they hardly have time to abandon – though of course an Abandonment Rate still exists.

For non-revenue generating Centres – and that’s most of them – Abandonment Rate is an outcome.  An outcome of Service Level.

If Service Level goes down, it is likely (though not assured) that Abandonment Rate will go up.  Or when Service Level goes up, it is likely (though not assured) that Abandonment Rate will go down.

Smart Centres see Abandonment Rate as an outcome.

So rather than targeting it they examine it.  When do most people abandon?  What intervals experience higher or lower Abandonment Rate?  When do we play our delay announcements?  Should we move our announcements around?

Because Abandonment Rate is a human behaviour – not a mathematical behaviour.  It lies in the hands of the Customer.

We’re in control of (drum roll please) Service Level.  When we’re achieving Service Level by interval we’re accomplishing our mission. We’re delivering a consistent Customer Wait Time, Agent Occupancy and Organizational promise experience.

Trying to chase abandoned calls is like trying to catch a greased pig at the county fair.

Getting a handle on Abandonment Rate in the Contact Centre


Service Level (again)

I like to say that Service Level has a Driving License and the other metrics are all passengers in the car.

So if Service Level turns left – they all turn left.  If Service Level turns right – they all turn right.

You get the idea.

And if it helps remember that there are only 3 drivers of Service Level for any interval.  They are:

  • Contact Volume
  • Average Handling Time
  • Agent Capacity

If you’re struggling to achieve Service Level for any interval or set of intervals your root cause lies within one of more of these 3 variables.

More articles on Operations soon!

Thank you for reading!

Thank you for the time you took to read this today!

Stay up to date on our articles and insights! Send me a message if you have any questions.

Daniel Ord

[email protected]

Contact Center
Send me a message

Decode the Customer Ecosystem

Want to stay in touch with our articles, insights & offers?