Most people are familiar with a forecast in the context of the weather. For example, your local news will tell you that it might rain tomorrow.
But, when it comes to workforce management, a forecast isn’t about whether or not you need to pack an umbrella. Instead, your forecast is the backbone of…well, everything.
Your forecast is what you’ll use to make informed and educated decisions about everything from staffing to activities to ensure that your agents are supported, your stakeholders are happy, and your customers are getting the service they deserve.
Does this matter to you? Keep reading if…
- You’re consistently finding your support team understaffed or overstaffed.
- You’re being reactive rather than proactive and frequently need to scramble to adjust your plans.
- You’re getting questions from stakeholders about your strategy and staffing levels.
So, what exactly is forecasting?
Let’s dig a little deeper into what forecasting actually is. A forecast itself involves making an informed estimate about what will happen based on the information you have available (like a meteorologist warning you of a thunderstorm based on what they see on the radar).
That “information” piece is crucial. Forecasting doesn’t involve operating on a hunch or throwing a dart at a board. To create an accurate forecast, you combine historical data with future assumptions.
So, for example, if you need to estimate how many calls your support team can expect to receive next week, you’d look back at past contact volume numbers to get a more realistic grasp on what future weeks will look like.
Don’t worry—we’ll get into the nitty gritty of how to do that a little later. But, in the meantime, it’s important to note that forecasts aren’t a “set it and forget it” sort of thing. You’ll get better and better at forecasting if you continually go back and compare what actually happened to your forecast and make improvements and adjustments from there.
In short, forecasting is a constant learning process (but it doesn’t need to be a painful one—we promise).
What are the different types of forecasting?
Forecasts can be created for anything—from weather patterns to budgets. But here, we’re talking specifically about contact volume forecasting, or how many contacts your support team will receive during a given time period.
Under the umbrella of forecasting (sorry, we can’t resist the weather puns), there are a number of different forecast types you can implement. These include:
- Long term forecasting: Makes predictions looking out at a longer period of time (usually more than three months).
- Short term forecasting: Makes predictions looking at a more immediate period of time (usually under three months).
- Special event forecasting: Makes predictions for events that don’t follow your normal pattern for contact volume (like Black Friday, as one example).
All of those have a place in your workforce management strategy, but let’s focus on the more operational and intricate types of forecasting: short-term and special event forecasting.
Let’s dig into short term forecasting
Maybe you want to figure out how many contacts your team will receive next week. Or over the next few weeks. Or even over the next two or three months.
That’s where short term forecasting comes into play. You’ll look back at your historical data to make contact predictions for the next couple of weeks or months.
Think you can look back at last week and say, “Huh, we received 7,500 contacts last week! I’ll plan the same for next week”? Not so fast. Accurate forecasts are a bit more intricate than just using last week’s numbers.
In fact, there are several different methods that you can use for short term forecasting. Some of the most common ones include:
- Running averages methodology: This one is the simplest option. You look back at the running averages of contacts in recent weeks to forecast for upcoming ones.
- Momentum model: This is the next iterative step from running averages. Instead of just using your pure averages, you’ll figure out your growth rate week-over-week. You’ll use those recent week-over-week volume changes to apply them forward and then update each week with the latest data.
- Regression: This one is the most complex as it involves looking at other correlations in your historical data. There are tons of different variables you could use to build a more accurate forecast. But in all honesty, it’s time intensive and isn’t always worth the effort.
Running averages are straightforward (provided you paid attention in math class) and regressions are super complex. So, let’s talk more about the option that’s right in the middle: the momentum model.
Calculating your forecast using the momentum model
You’re going to use the momentum model to forecast your contact volume for future weeks. Here are the steps to follow:
1. Pick a number of historical weeks to look back at
We think 8-12 weeks is a good rule of thumb as it gives you enough trends to work with. If you know your queue volume fluctuates a lot, you might want to look back at fewer weeks—like only two to start.
2. Calculate the growth percentage between each week
That might sound complex, but you don’t need a statistics degree to do this. Here’s the formula:
end week volume – start week volume
start week volume
So, if you had 1,335 contacts last week and 1,223 the week before that, here’s what that would look like:
1,335 – 1,223
That gives you an answer of 0.0915. Multiply that by 100 and you roughly have a growth rate of 9%
3. Estimate your future weeks
Now you have an average growth percentage that you can use to estimate future weeks. The momentum model considers a little more than just your running averages to give you a more accurate short term forecast.
What do you do with that information?
Understanding your volume and growth percentages is one thing. But, the goal here isn’t just to have a bunch of numbers—you need a plan you can work against. You might have 7,500 contacts next week, but how do you staff against that?
You need to take that weekly volume and get into something more tactical. That’s where your arrival pattern comes in. An arrival pattern is a volume agnostic representation of customer contact patterns within your business. To put it simply, you’ll represent volume with a percentage instead of a raw digit.
Choose an interval (whether it’s an hour or 15 minutes—whatever works for you) and get your volume for each of those intervals throughout the day. Then, you’ll use another simple calculation:
total weekly volume
So, if you’re looking at each hour and you had 60 contacts at 8 am on Monday and a weekly volume of 7,500, that means 8 am on Monday made up 0.8% of your weekly contact volume. That feels a little more digestible and tactical to staff against, doesn’t it?
You can repeat that and get arrival patterns for each hour (or whatever interval you chose). But, you can also dedicate arrival patterns for each day. For example, Monday makes up 10% of our call volume. Tuesday is 12%...you get the idea. It’s not necessary, but it’s another way to break up your volume in a way that feels more actionable.
Now you can combine your weekly totals and the arrival pattern to create your forecast for the next few weeks or months. Just remember to check your work and continually update your forecast—things can change fast!
Let’s dig into special event forecasting
The tips and calculations above can help you get a forecast that’s grounded in reality. But, you know the saying about best-laid plans, right? They tend to fall apart—and that’s especially true if you have a special event in the mix.
Special event forecasts can help you account for those times that are out of the ordinary. That could be a special sale or marketing campaign, a maintenance period, a new product launch, a holiday…basically anything that could throw your “normal” volume way off.
Keep in mind that sometimes the impact on your contact volume doesn’t happen entirely during the event itself. You could also feel some ripples before or after. For example, if you’re closed over the holidays, you can likely expect higher-than-average contact volume when you get back.
This is obviously something you want to account for in your forecasts, so let’s briefly cover the gist of special even forecasting. If you really want to familiarize yourself with the ins and outs, we highly recommend checking out our comprehensive guide.
On a surface level, here’s how special event forecasting works:
1. Look at last year’s event volume
Find the relevant event last year and ask yourself questions like:
- How many contacts did we receive?
- Was there a clear arrival pattern?
- Did volume spike in the morning and then level off or did it stay steady throughout the day?
Make sure that you also forecast for the correct dates. Some holidays occur on the same day of the week each year but on a different date. Others occur on the same date but a different day of the week.
2. Compare to the preceding days and weeks
Now, look at several weeks (we recommend four to eight weeks) prior to the holiday last year. What was the average volume you received over that time?
You'll use that number as a proxy for the volume you would receive if the day were not a holiday. Dividing your holiday volume by this proxy gives you your holiday multiplier, the number that you'll use to multiply your usual forecasted projections.
That holiday multiplier is your secret sauce for more accurate special event forecasts.
3. Adjust your forecast
With your holiday multiplier ready to go, you’ll multiply the number of contacts you’d normally expect for that time by your holiday multiplier.
Special events don’t have to be surprises that leave you no choice but to hold tight and hang on. Finding your multiplier will help you create special event forecasts so that you can plan (and staff!) accordingly.
4. Review your work
So, the big day came and went and your special event forecast wasn’t exactly spot on. This isn’t your time to shrug and say, “Huh, better luck next time…”
It’s important to look back at what actually happened so you can create even more accurate and helpful special event forecasts moving forward. Here are two things you can do:
- Get a real multiplier: Now that the special event has come and gone, you can look back and get a real multiplier instead of your best-guess one. You’ll use the same formula but with real data from the actual holiday.
- Find causes: Dig in and figure out what caused reality to be so different from your forecast. You could find that the overall day was higher than planned, but not because of the event itself. Instead, your baseline volume actually increased. That’s helpful information to use moving forward.
Stop guessing and start forecasting
When the meteorologist calls for rain and you see nothin’ but blue skies, it’s tempting to write off forecasts as conjecture. Hunches. Gut feelings.
In reality, forecasts use data to form educated predictions about what will happen. And when you need to staff a support team? They can help you create plans that keep everybody—your team, your stakeholders, and your customers—happy.
But, you know, it never hurts to grab an umbrella. Just in case.