Accurate Forecasting: Your Call Center Depends on It

Original Publication: Customer Management Insight - May 2008
View Online

There’s a well-known proverb, handed down through the centuries, that illustrates the logical progression of actions and outcomes:

For want of a nail the shoe was lost.
For want of a shoe the horse was lost.
For want of a horse the rider was lost.
For want of a rider the battle was lost.
For want of a battle the kingdom was lost.
And all for the want of a horseshoe nail.

While you may not lose your kingdom, inaccurate call center forecasts lead to cascading consequences that could cost you dearly. Getting this area of planning right is a high-leverage activity essential to the health of your business.

The workload forecast is the basis for determining staffing needs — as well as the requirements for other resources, such as how many workstations and phone lines are needed. The forecast provides the foundation for:

• Calculating base staff required to meet your service level objectives
• Engineering network and system requirements
• Minimizing abandoned and blocked calls
• Organizing accurate, workable schedules
• Predicting staffing and network costs
• Establishing an environment in which quality service can be
• provided
• Ultimately, meeting customer expectations.
  

Small Inaccuracies = Big Impact

What happens when forecasts are off the mark? Let’s look first at staffing trade-offs.

The example in Table 1 (based on the industry standard queuing formula, Erlang C) illustrates the relationships between staffing level, service level (percentage of calls handled within a specified timeframe), average speed of answer, agent occupancy (percentage of time agents handle calls versus wait for calls) and trunk load. Assume the highlighted row is your intended service level. See how being just a few agents off the mark affects everything else?

Table 2 gives further insight into what happens to callers. With the intended service level, 65 of the 250 total callers will wait five seconds or longer. In the next five seconds, seven of those callers reach agents, so only 58 callers are waiting 10 seconds or longer. When you’re a few agents (say 30 versus 34) off the mark — Ouch! — look at those wait times! That’s the kind of thing that drives agents to throw in the towel and consumers to blog to the multitudes about your (lack of) service. Overstaffing, on the other hand, creates its own unsavory set of problems — high costs, underused staff and loss of credibility in budgeting, to name a few.

 Table 1. When Service Level Goes Up, Occupancy Goes Down

 

A Unique Challenge

The big challenge with forecasting in customer contact environments is that the inbound portion of the workload arrives randomly. Customers decide when and how they will contact the organization, and the resulting work does not arrive in a nice, even flow. This puts a premium on knowing how to forecast for this type of environment.

Many call centers use time-series methodologies common in workforce management software — the governing assumption being that past data will reflect trends that will continue into the future. But the best forecasts also incorporate more advanced approaches — e.g., explanatory methods that reveal linkages between variables, as well as judgmental forecasting that accounts for factors not reflected in historical patterns.

So what’s accurate? As a rule, the negative consequences of under- or over-staffing begin to take root at just five percent off the mark. And — very important — accuracy can only be measured by interval (e.g., half-hour), not by day!

Today’s always-on, real-time world doesn’t give margin to the unprepared. Trying to provide customer services without a forecast that’s consistently on target is like going into to battle on an unshod horse. Don’t do it!

When Forecasts Are Off the Mark

Consequences of Being Understaffed

• Frustrated customers
• Abandoned calls
• Longer calls
• More errors and rework
• Higher telephone network usage and costs
• Staff stress, burnout, turnover
• Consequences of Being Overstaffed
• Unnecessarily high staffing costs
• Underuse of staff
• Boredom of staff

Factors Affecting Call Center Workloads

• Calling patterns
• Revenue growth (or decline)
• Marketing activities
• Changes in customer segmentations
• Competitor activities
• Technology changes (internal and external)
• Web site revisions (content or structure)
• Laws and regulations
• Customer experience levels
• Agent experience levels
• New product rollouts
• Customer relationship initiatives
• Restructuring
• Quality improvement initiatives
• Publicity
• New suppliers and business partners
• Human resources policies
• Cost-cutting or growth initiatives
• Economic developments