In the renowned ancient treatise The Art of War, Sun Tzu famously states that every battle is won before it is fought.
To paraphrase a paraphrase, “The effective general makes sure his army has ample resources, he manages those resources wisely, and actively controls his forces. He trains his troops and his soldier’s specialties, sets up supply lines, or, better yet, forages for his enemy’s resources.
“He makes sure that his army is ready at the right time -- but not too early -- the troops should be available just-in-time. He develops a strategy.”
Isn’t this just like managing contact centers?
While this comparison is stretched and tongue-in-cheek, planning for any complex and expensive endeavor is of critical importance. And contact centers are complex and expensive operations.
Here is a funny fact: with a few recent exceptions, contact center managers do not have real systems and technologies available to help develop longer term operational plans. Really.
Contact center management has invested in important workforce management (WFM) products, but the software only serves to manage the workforce you have already hired.
The Strategic Planning Problem
Determining how many resources you need is usually left up to another process, the capacity planning or strategic planning process. For the vast majority of contact center operations, both big and small, that process is managed through a homegrown Erlang-based spreadsheet.
The contact center strategic planning problem can be stated like this: “What is the resource plan and budget, over the next 18 or so months, which will allow us to maintain our service standards at least cost?”
This problem is both complex and significant and it entails predicting the peaks and valleys of seasonal contact center demand. It requires understanding the seasonality of agent shrinkage (sick time, vacation, absence, etc., ...) and agent attrition. It maps the differences in performance across different staff groups and geographies.
Most importantly, it includes knowing the relationship between the number of people you hire and the service you deliver. In an omnichannel/multichannel operation, across states and countries, this is not a simple calculation. Yet, “simply” is how we have traditionally tried to solve this problem.
...planning for any complex and expensive endeavor is of critical importance. And contact centers are complex and expensive operations.
Strategic planning is a big problem, but if you solve it well, your contact center operation will run more smoothly, and your costs will be lower. Solve it perfectly, and you always have the exact right number of agents logging in to help your customers every single day. It may be the single most underappreciated and critical process left to tackle of contact center technologies.
I’ve always wondered why more WFM software companies haven’t developed strategic planning systems to complement their current offerings. My guess is that strategic planning systems are both different enough from standard WFM and complex enough that it is just too much of an investment. Let’s discuss.
Workforce Management and Strategic Planning
WFM has been with us since the early 1990s when TeleCenter Systems (TCS), IEX, and Cybernetics developed the first such tools for call centers. Each focused their pitch around automated and optimization-based scheduling. They promised that the work schedules they produced for agents would be much more efficient than their prospects could possibly build by hand. They were wildly successful.
Similarly, the first strategic planning system was built by Bay Bridge Decision Technologies in 2000 (where I was a founder). Bay Bridge’s CenterBridge, later Decisions, pitched a similar value proposition: we could develop much more efficient plans than a company could create using a spreadsheet. It sold well, and we were very proud of our team and our results.
But other than the efficiency pitch, there are significant differences between the business questions each system is developed to solve. Therefore the technologies needed to make the systems successful were necessarily different, too.
First, WFM systems are focused on four main aspects of planning. These are: 1) short-term interval forecasting of volumes, 2) scheduling agents and managing the scheduling process, 3) real-time management of the agent’s time and exceptions, and 4) performance management and reporting.
The sorts of analyses that WFM performs are tactical: “Do I need to cancel training today to meet service levels?” or “How many agents will I be short this afternoon?” A lot of effort goes in to managing schedule requests and daily exceptions.
Second, capacity planning systems provide similar sounding functions, but with significantly more complexity. These are: 1) long-term forecasts of volumes, shrinkage, handle times, pay rates, etc., 2) simulations of the contact center network to determine how many agents are required, 3) optimization of hiring, training, and overtime plans, and 4) budgeting and reporting.
Great strategic planning systems can handle more critical business questions, like “What is the best use of in-house versus outsourced agents?”, “How can my operation best save $2 million?”, or “What is the risk to the operation if our attrition continues to climb?”
WFM is, by its nature, tactical and capacity planning is inherently strategic. While both are necessary to run an efficient operation, capacity planning has been given the short shrift in applying resources to this important business process. Spreadsheets are just the wrong technology for it.
What Does a Great Planning Process Look Like?
Contact center strategic planning, at its best, is fast, provably accurate, and mathematically optimal. It automates the planning process and automates the development of what-if analyses. Its formulas and models are locked down, so hidden and costly spreadsheet errors are not possible.
At its core, a great planning process involves three different mathematical modeling processes, forecasting, operations simulation, and resource optimization models.
Each serve to explain and solve different parts of the capacity planning puzzle. When coupled together, it allows rapid answers to big picture, boardroom-level business questions. Let’s discuss each.
Forecasting
When discussing strategic planning forecasting, what needs forecasting is more important than how to perform it. Forecasting for WFM usually focuses on predicting contact volumes. That is important for capacity planning, too, but there is a lot more forecasting still to be done.
To determine how many agents are needed week over week into the future, forecasting volumes is necessary, but so to is predicting when agents will call in sick (it is fairly seasonal), or when agents will quit. Any metric that affects staffing or productivity should be forecasted.
Another key point about long-term forecasting: often, certain metrics are differently seasonal by location. It is both a stereotype and a fact borne out in handle time data that in the U.S. certain locations speak slower than other areas. That should be considered.
...a solid plan brings appropriate and lower costs. Just-in-time staffing means just-in-time salaries.
But it is also true that certain locations might have more absenteeism than others. Certainly, agent attrition has more to do with the local economy than national trends. It is a best practice to forecast at the level of detail that is predictable.
There are many great time-series forecasting models available. And state-of-the-art contact center forecasting is still either a “best of the best” model or an ensemble (average) of several forecasting models.
Operations Simulation
If I were to rank order all features of a capacity planning system (data load, forecasting, operation simulation, staffing optimization, reporting and what-if analyses), number one, the most important would be the operation simulation.
What it does is to model the operation and to answer key questions. Like:
- “Given our forecasts (volumes, shrink, attrition, handle times, etc.) how many people do we need each week to hit our service goals?”
- “If we had the following week-over-week staff plan, what would our service levels, abandons, occupancies, speeds of answer be each week into the future?”
In effect, the simulation becomes a crystal ball that can be used to answer any question about service, abandons, staffing, costs, sales, and agent occupancy.
There are many approximations available to help answer these sorts of questions, like the old Erlang equations. But in a complex operation like ours, these approximations do not come close to accurately capturing the nuances of the contact center operation. While an Erlang model might be OK when used for tactical WFM interval staffing, it is simply not accurate enough for capacity planning.
There is a step to building models that our industry has seemed to have forgotten and that is we should prove our models are accurate.
Given that contact center operations are complex, a generic model is suspect. And whatever means you or your vendor take to perform long-term staffing and analyses, you should make sure to include a model validation exercise of this critical model. Just because a vendor says “we can predict staffing accurately” it doesn’t mean you should believe them. Ask for a model validation on your data.
Resource Optimization
The final major step in your capacity planning modeling process is resource optimization. Algorithms that are similar to those used in WFM scheduling can be employed to come up with just-in-time resource plans.
These types of models, if built well, can design resourcing that does not overstaff or understaff. With them, the right number of agents will be onboarded, be through their learning curves, and ready for work when needed.
There are several levers when it comes to longer-term planning. Hiring, overtime, undertime, training, and controllable shrink are available to be used to make sure the technology’s resource optimizer can staff just so.
However, it is important to understand each operation and location’s staffing constraints. Like certain locations can only handle so much overtime or undertime. Or that there are fewer trainers or training rooms available for onboarding agents in some centers.
What you can do by automating all three modeling processes
When all three of these models work together a beautiful thing happens, contact center management is given a crystal ball to answer all sorts of big-picture questions.
- Do you want to quickly change your planning assumptions and understand the costs of new scenarios?
- Do you want to know the marginal benefit of hiring one more agent?
- Do you want to understand the risk to your operation of weather events happening at some of your offices?
- Will an impending recession require changes to your operation?
Your crystal ball can help you answer all of those questions.
The Benefits of Planning Well
Contact center planning, when done well, has enormous benefits. Here they are.
First, a well-developed capacity plan will result in consistent service. Simply, right staffing leads to service consistency.
Further, the variability of service delivery is narrowed. Meaning, natural variability in staffing or contact volumes or handle times, will be part of the simulation and be considered. It’s not that these events don’t happen, but your models consider them and staff appropriately.
Second, more consistent service means a better and more stable agent experience. With service consistency comes steady agent occupancy, the pace of work is more even.
Third, a solid planning process allows more agent flexibility. Because management knows when it is necessary to offer both overtime and undertime, these can be offered well in advance to the agent and company’s benefit. A focus on the future allows more flexibility for your staff.
Lastly, a solid plan brings appropriate and lower costs. Just-in-time staffing means just-in-time salaries. Your agents will be hired to come onto the floor the day they are needed, not a day early or a day late.
There is also an adage that no battle plan survives first contact with the enemy. Dwight Eisenhower, the Allied Supreme Commander in World War II, was famously said to have spent D-Day playing bridge with friends. He had planned, resourced, and trained his army; there was nothing left for him to do.
The wonderful truth was that even though he had no active role in the important day’s events, he had designed a remarkable plan that served the great purpose. Every battle is won before it is fought.