Long Term Planning, Twelve Years In

By Ric Kosiba, Vice President, Interactive Intelligence

Twelve years ago, we started a conversation with the workforce planning community about the importance of long term planning and what-if analyses. Usually when we spoke to members of our community, we received a response somewhat akin to “is there software for this?”

This sort of planning is called many things — strategic planning, capacity planning, budget planning, operational planning, and long range planning are the most common names. Whatever it is called, we  have never spoken to a contact center operational manager who did not have this function represented at their operation in some way. Every operation does it, and back in the day, every operation provided it ad-hoc using large spreadsheets.
However, in the past four years or so, there has been a common understanding of the practice of strategic planning by our industry and the analysts that cover it. This understanding goes something like this:

  • Strategic planning is a critical first step in planning for any operation. In the contact center industry, it is the first step in the workforce optimization suite. According to Gartner, “Strategic Planning software can greatly assist with  decision making processes, such as staff recruitment needs, skills requirements and budgetary planning.”*
  • It is understood that strategic planning is the proper place to provide analyses for decision-making that affects the medium- and long-term (e.g., staffing and resource planning).
  • A spreadsheet is not the appropriate technology for planning an operation.
  • There is a common set of basic features that your strategic planning process should have.
  • Business variability has made strategic planning more important, as operations need to be more nimble and analyses for executive level decisions need to be provided more often and with more accuracy.

Let’s discuss these further.

Why is the spreadsheet not the right technology for planning?

The most common planning system in the world is the spreadsheet. However, it has one major drawback (and several only slightly more minor drawbacks): it is prone to terrible error. There has been a growing list of decision-making debacles that have been the fault of complex spreadsheets containing errors deep in their complex and linked equations. Frankly, we have stretched spreadsheets way beyond their abilities. From a casual internet search of “Excel mistakes” there are some doozies, including spreadsheet errors that helped bring down MF Global and contributed to a $6B trading loss at JP Morgan Chase. Do your own search and enjoy.

But what about contact center planning spreadsheets?

Clearly, these are not as complex or prone to error, right?

We have a bit of a unique perspective on this question — we have seen and analyzed dozens of contact center planning spreadsheets over the years, and our experience tells us:

  • Spreadsheets are not designed to be easy to alter and because plans change often, every contact center planning spreadsheet we’ve encountered had calculation errors, many significant, and often because the spreadsheet was inherited from earlier planners.
  • As operations have become more complex (multi-skill, multi-site, multi-channel), the spreadsheets have become slower, more error prone, and more difficult to use.

    • Many spreadsheets — in order to be useful at all — oversimplify the planning process and the simplification introduces significant sources of inefficiency to the operation.
    • An efficient plan — well managed — can significantly improve the performance of an operation.
      Your planning process should have some basic features. A contact center planning spreadsheet or system is really a series of mathematical models tied together:
    • Forecasting models to forecast volumes, shrink, attrition, and handle times
    • Simulation models to predict center performance (abandons, service level, occupancy, ASA, customer satisfaction)
    • Staffing optimization models to determine the best resource plan (hiring, overtime, controllable shrink, undertime)

Variable labor financial models, variance analysis, and automated budgeting.

As such, the most basic features our planning process should have is the ability to forecast all performance drivers, the ability to predict the performance expected accurately under various staffing scenarios, the ability to determine the  appropriate week by week staffing levels and the best resource plans, and the ability to cost all plans out accurately. Implicit in this list is the guarantee that any performance prediction model (i.e., Erlang C or discrete-event simulation) used is accurate (Erlang is not) and works in two directions, predicting performance given staffing or staff required given a goal.

Similarly, any system would have a data model that includes historical weekly calls or contacts, staffing, shrinkage,  handle times and components, dialer probabilities, customer experience scores (however defined), at the appropriate level of detail (usually by staff group or contact types).

Business variability makes forecasting and planning more important.

The final point we’ve all learned from the last decade is one that surprised us a bit. Walking into the 21st century, we never dreamed that business variability would be as drastic as it has been. History is much less of a guide than it was 15 years ago, as the world economy and normal business cycles are much less stable.

But this has made the planning process more, not less, important. We still are required to make longer term decisions, such as hiring, developing new customer servicing strategies, or exploring and developing new channels. But given the dynamics of business forecasts, how can we plan if everything is changing?

This question has been asked and answered in many business articles and analyst’s papers, and it is this: by regularly monitoring variance, reforecasting as the business environment changes, and revisiting all long term decisions as new information becomes available, we can ensure that our company and contact center operation is well informed, well managed, and nimble.

Without a well-disciplined planning process, we are flying blind. The last place we want to be on autopilot is in the middle of a storm.

Ric Kosiba, Ph.D. is a charter member of SWPP and Vice President and founder of the Bay Bridge Decisions Group at Interactive Intelligence. He can be reached at Ric.Kosiba@InIn.com or (410) 224-9883.

* Gartner, Technology Overview for Contact Center Workforce Optimization, May 2012