X% answered in Y seconds. That’s the formula we use for service level calculations in contact centers. It can (and often should) be used to set objectives for any channel, but is most commonly utilized for inbound calls.
Here’s what I love about service level objectives: They are the tangible proof of how an organization calculates the trade-off between cost and customer satisfaction. If you truly value customer satisfaction, you will set an aggressive service level standard and routinely meet it. If cost drives the day, your objective will be much more modest, and you likely fall short during busy times.
Here’s what I hate about service level objectives: Most contact centers have had the same goal in place for so long that no one can trace its roots. It may no longer be relevant, but without a firm grasp of the many implications of a service level objective, no one feels comfortable making a change.
Despite any discomfort, refreshing a service level objective is an activity that cannot be avoided forever. It is a great subject of discussion for all levels of the leadership team, and you can be sure that many opinions will be presented. Before going there, though, everyone in the room can be grounded on some mathematical facts related to different service levels, such as the impact on:
- Answer times outside of the threshold
- Pacing of calls
- Staffing expense
Each of these factors can be calculated for many potential service levels, and doing so helps shed light on the trade-offs that must be faced.
Outside of the Threshold: The Unfortunate Few
When we set a standard of, for example, 80% of calls answered in 30 seconds, our focus is on the 80%. If we meet our mark, we know that 4 out of 5 callers will get answered in what most any customer would consider a short timeframe. It’s a comforting thought, and so we set about doing whatever it takes to make sure we meet the objective.
But… there is still another 20% out there. What happens to them? They represent the “tail” of the call answer time graph, and that tail is always longer than people think. The actual outcomes will depend on the size of your organization (smaller centers have longer “tails” than larger ones). As an example, though, a typical 200-seat contact center, with an 80/30 service level objective, will subject the unfortunate 20% to the following:
- Five of that 20% will be answered between 21 and 47 seconds
- Another five of that 20% will be answered between 48 and 70 seconds
- A third five of that 20% will be answered between 71 and 105 seconds
- The final five of that 20% will wait more than 105 seconds… with a longest wait at well over three minutes
So, yes, in this example, answering 80% of calls in 30 seconds is the same as answering 95% in 105 seconds. That helps define the tail, and everyone needs to understand it when setting your objective.
Impact on Abandonment
The previously described “tail” is where abandonment exists. Because it is such a clear indicator of caller dissatisfaction, the abandoned percent is one of the most critical metrics we have.
Once we have fine-tuned our on-hold messaging, the only way we can positively influence the abandoned rate is by answering the call more quickly. A relationship exists between service level and abandonment, and it differs by organization. That’s why every contact center should know their abandonment curve (an example of one is in the sidebar below).
This curve is a critical piece of information to reference when setting the service level objective. It defines the one answer time metric—abandonment—which is a true reflection of customer behavior, so ignoring it is not an option.
Setting the Pace
Service level directly affects your occupancy rate (the percent of manned time that is spent in either talk or after-call work). The occupancy rate is what affects the pacing of calls received by agents. If the occupancy rate is too high, agents get burned out. They will do whatever it takes to catch a break, from inflating after-call work to extending breaks to calling out sick.
Occupancy can also be too low. If there is too much time between calls, agents get bored. Time moves slowly, and quality actually decreases. So is there an ideal occupancy rate? Opinions vary, and there are a number of factors that affect the ideal occupancy rate. For a typical center, though, many agree that burnout starts showing up at 92% and beyond, and boredom sets in under 80%.
Service level and occupancy have an inverse relationship. As service level increases, occupancy decreases, and vice versa. This is a mathematical fact, and no amount of hoping, coaching or micromanaging can change the occupancy rate. It is basically set when you determine your service level objective, so it needs to be well understood by those setting the standards.
The Cost Factor
We all realize that better service levels require more staff, and more staff requires more money. What needs to be known, though, is how these two affect each other. It is a common misconception that great cost savings can be achieved by small reductions in service level objectives. Moving from 90% in 20 seconds to 75% in 20 seconds may reduce those getting a quick answer by 15 percentage points, but the cost savings in most instances will be under 4%. The relationship between service level and cost is not linear, and this needs to be clear so that better decisions are made.
Finding the Right Balance
The four factors discussed here are not the only issues to consider when setting a service level objective. They are the objective ones, though, and since they are based on mathematical fact they carry considerable weight.
They also “ride together”—a given service level objective will have a given cost, abandoned rate, occupancy rate and tail. You cannot change one without changing the others. And the desire to find the right balance is what makes this such a great discussion.