Newsletters

Customer Support:   (972) 395-3225

Home

Articles, News, Announcements - click Main News Page
Previous Story       Next Story
    
Managing The Time Crunch: Getting Proactive About Performance Management

by By Ric Kosiba, Ph.D., Bay Bridge Decision Technologies - January 1, 2013

Managing The Time Crunch:
Getting Proactive About Performance Management
 By Ric Kosiba, Ph.D., Bay Bridge Decision Technologies

Busy, Busy, Busy

We have been doing a fair amount of traveling this year, and have met with many people in workforce management. We all have a lot in common: we’re mathematically inclined with an unusual obsession with efficiency, we work to very unworkable deadlines, we are able to assimilate data and ideas from many individuals and organizations, and we are universally leaned upon to save our companies a lot of money while keeping our customers happy.

However, because of these responsibilities and the complexity of our jobs, we workforce managers are almost universally playing catch-up. As Yogi Berra says: “It’s tough to make predictions, especially about the future.” Is there value in getting proactive? Can this change? Can we go from fixing problems to proposing solutions ahead of the problem?

In this article, we will discuss the problems and their causes that drive us into “catch-up” mode, strategies for getting ahead of the curve, and the value you’ll see from doing it.

Our Friends in Marketing

Let’s start by looking at a common source of this “catch up” problem: the give and take (mostly take) between workforce management groups and marketing. Marketing develops wonderful sales campaigns and other incentives that cause our customers to call us (e.g., to place orders or ask us questions). Best case, they tell the contact centers about the campaign with enough lead-time to allow us to prepare. But sometimes our call center performance suffers from the impact of marketing campaigns because:

· Contact centers are not told of campaigns, communications, or events that may increase call volumes or handle times with enough lead time to adjust staffing.

· Expected campaign call volumes, based on marketing response expectations, are inaccurate.

· Handle times increase unexpectedly based on marketing communications (e.g., complicated programs, unclear collateral information).

· The operational impacts of new programs on center operations are not sufficiently understood and have unintended consequences.

So what is the result?

· Low and inconsistent customer service quality (e.g., missed service goals and/or large day to day swings in customer wait times).

· Burn out of our agents and our workforce management staff.

· A last minute scramble to analyze program impact on staffing needs.

· Unmet management expectations.

· The workforce management team’s reputation for accurate planning and forecasting takes a hit.

Certainly these issues are not always caused by communication problems with marketing: weather and other unpredictable events also have a big impact on our ability to do our job perfectly.

The Truth About Forecasts

The truth is that forecasts are expected values, highly dependent on assumptions we get from other departments, and the actual performance will vary. Because of this, our forecasts always have an inherent “plus or minus” value associated with them—even if we don’t state it that way.

Our challenge is to educate our constituents in marketing, center management, and the executive suite about the ranges of outcomes we expect in our forecasts as well as their sensitivity to key assumptions. How do we set the appropriate expectations?

First, Manage Expectations

First we need to manage the expectations of others by letting them know the likely performance ranges that result from their decisions. By knowing ahead of time the likely range of performance that result from a new policy, marketing initiative, or customer communication, we can be better prepared for what happens, and, more importantly, modify or phase the program to reduce its risk.

While it might take senior management some time getting used to the idea, we recommend providing the expected call volume and handle time forecast and then adding an upper and lower boundary to provide a range of likely results. For example, you may be able to state that the forecast is 225,000 plus or minus 5,000 calls with 90 percent confidence. This presentation reinforces the idea that workforce management is probabilistic in nature with the benefit that it is easier (and more meaningful) to come within a forecast range than to match a forecast number exactly.

What will happen if calls come in at the higher or lower ends of the forecast? To develop forecast intervals, you will need sensitivity analysis. In the winter 2002 edition of On Target, we discussed how to produce sensitivity analysis of our centers using a discrete-event simulation model. As we all know, the most effective way to communicate with marketing and senior management is through pictures and data. In determining the effect of variable call volumes of handle times, with simulation you can produce the following two graphs:

Having these graphs available will help manage expectations. When the marketing department hands you a call volume forecast based on their activities, you can ask them about their confidence in their forecast numbers (or keep tabs on their past forecast performance), and quickly highlight the range of likely service performance expected.

Second, Plan Ahead

The past few years have been among the most difficult for forecasting that we can remember. World events have changed customer behavior. Forecasts that used to run error rates of around 2% are now running 6% or higher. This has in turn spawned more forecasting activities and more plans than ever. We’ve a lot more work to do, but the number of hours in a day remains, sadly, constant.

In order to “proactively respond” to business events, both foreseen and unforeseen, it is important to do not just forecasting, but detailed operational response planning based on our range of assumptions. We can develop a series of plans to respond to a range of scenarios and keep these in our back pockets. Say, for example, that we develop complete hiring plans (and even budget forecasts) around the ranges of our confidence levels for the call volume forecast. When call volume trends start favoring one scenario over the others, we are ready to respond immediately.

This has tremendous value. First, the business is already prepared and can respond to these changes quicker. This fast response will save money and time to market. Second, our organizations will gain credibility. We’ve prepared senior management to know what the range of performance expectations are, and when one of the scenarios is looking more definite, our plan is already in place, and the business is already responding. Can you say “promotion?”

How to Get There

In order to achieve all of this we need to automate! This process needs to be quick, accurate, and seamless. It is perhaps time to bite the bullet, and build or purchase a planning tool that really meets your needs. Going through the effort of streamlining the planning process will pay dividends in more than time. You can reevaluate overly simplistic assumptions and algorithms (e.g., finally get rid of Erlang!) and improve both the speed and the accuracy of this process. We recommend, when engineering your new planning process, that you build new staffing algorithms as well, in order to improve the efficiency of your long term hiring plans.

Summary

Tough times are good and bad times for forecasters. We are needed more than ever, but our work product, by the nature of real-world instabilities, is more difficult to get right. Because of this, our reputation can suffer. We need to educate our organizations on the range of likely outcomes in our forecasts. By providing them with these ranges, we are inviting the discussion of the array of likely performance results and what can be done to narrow those ranges. These are good discussions to have.

At the same time, we must be prepared for unforeseen events with multiple planned responses. By being prepared for multiple contingencies, we will greatly speed our “time to market” with our responses. This is very valuable.

With spreadsheet infrastructure, getting to this level of preparedness will be difficult to achieve. We need to automate this process and develop new algorithms and technologies to make planning as effective as it can be. Proactive management of the planning process will reap huge rewards, both for the company and for us.

You can reach Ric at EDK@baybridgetech.com or (410) 990-1079.

 
Return to main news page