79: Forecasting in Support with Matt Dale

Charlotte Ward: 0:13

Hello and welcome to episode 79 of the Customer Support Leaders Podcast. I'm Charlotte Ward. The theme for this week is forecasting, so stay tuned for five leaders talking about that very topic. I'd like to welcome to the podcast again today, Matt Dale. Matt, it's lovely to have you back. This week we're talking about forecasting in support. Now I know that you work in a business that has a fair amount of seasonality, and I think that brings its own particular set of challenges, doesn't it?

Matt Dale: 0:48

It does, and it's it's great to be back, Charlotte. I'll as happy to join you on the podcast here. Yes, so we work in education, a public school, the K-12 school market. And for us, that means that in the fall, during back to school season, things get really crazy. We basically go from kind of our normal in May and June to nothing during July because everybody shuts down and goes on holiday. And in August, people come back having forgotten everything they knew about our product. Um and they're supposed to be using it at a day-to-day basis. And so we see volume on most of our products double. Um, and for a very short period there were on a couple products we see it almost triple. Um and that's a challenge, as you can imagine, from a forecasting perspective, knowing what is that, what is double acting, is it actually gonna be double? And how many people, how many agents do we need to have staff so that um we have reasonable call time call wait times? Obviously, we can't staff for 100% a peak. Um, but you know, at what point are we going to be causing real frustration with our customers? The other challenge that we have, I think, that that may be unique or at least different than some of your listeners is that our products are really complex. So it takes quite a bit of learning curve to get people staffed up on that. One of the things that many of my peers look at is they say, hey, we've got a peak season. Let's say I've got I've got a buddy in e-commerce. He knows during you know Black Friday to Christmas that his his business is going to be very busy. And they have folks that they'll bring in for a few weeks to handle you know the basic questions and stuff. And his onboarding usually takes about three days. Um, our normal onboarding process is about three weeks for the basics. And we really joke with our new agents that, hey, you know, you're basically useless for the first six months. And we're okay with that. You need to be okay with that. But from a staffing perspective, that's a real challenge because I can't just bring in temporary folks or part-time folks to help us with that high peak volume. I have to figure out a way to still provide good support for our customers, but but do it in a cost-effective way that doesn't break the budget the rest of the year.

Charlotte Ward: 2:39

And you have a sense of of that seasonality from previous seasons, but how how accurate is your your perception of what the load might be next peak season, would you say, based on last peak season?

Matt Dale: 2:51

So right now all bets are off. Um, this whole COVID thing has messed up all of our models. And I think that's an important thing to think about too. If you're if you're listening to this and you're thinking about, hey, I've got this model or I'm trying to figure out how it works. I want to compare last year, last quarter. That's really difficult because we don't know what our contracts are going to look like. Last year, or in a normal year at this time, we'd be able to look at the sales pipeline and say, hey, here's what we're expecting to sell. ARR is one of the metrics we look at, and the other is the number of students that we're bringing on. And the third is the number of school districts. And each one of those shows us a little bit different picture. Um, a small school district tends to actually need more help because they don't have the staff to do the work that they need to do. So under a certain size, they're a little bit more of a challenge. But but typically I would take a look at the sales forecast and say, hey, this is where we're at. I can look at what we did last year. I can look at, you know, this many new dollars in ARR is going to roughly represent, you know, 20% increase, and this is gonna be a 12% increase in tickets. And I can kind of back into the numbers and and get an idea of, okay, I'm, you know, in in the middle of September, I'm expecting this kind of volume. And based on what we did last year at our team functioning at this level, you know, I think we need new a couple new people or or you know, whatever it is. A lot of stuff is changing all the time. And so you can build a model. In fact, I I've been working on our model for the last six or seven years, and every year it gets a little bit better, but every year there's a curveball, there's a change. And I think it's important that when we think about forecasting, that we're making sure that you have um that you're you're understanding your how how good your model is, but also that it's there's going to be a level of uncertainty.

Charlotte Ward: 4:22

And when you you're talking specifically about how you extract likely growth and and likely upcoming numbers, how in broad strokes do you do that? You you talked a little bit about knowing specific types of customer and specific student numbers.

Matt Dale: 4:39

Really, we're we're basically trying to find the correlation between whatever the number is, whether it's dollars in new sales or schools or or number of schools in certain size, and and what what sort of volume you're gonna see with that. Um also you have to kind of take into account what sort of volume you, you know, things tail off. So, for example, for us, a second or third year uh district that's been with us three years, they're gonna have different support needs than a brand new first district. They they don't drop off in our case as quickly as you might expect because as they finish their initial implementation, then they say, cool, we'd like to add this part, or we'd like to add that part. But understanding your customers and understanding that relationship between, hey, we got a new customer, or we have we have this many customers that are two years, this many customers that are three years old, based on historic data, we can see that assuming the the engineering team doesn't you know release a new feature that doesn't work that everybody's frustrated at, we can expect this sort of volume and and kind of getting a rough range there. So, you know, working with the sales team, making sure the Salesforce numbers are accurate and or or understanding the level of accuracy. If you make decisions based on knowledge that you think is is true but but isn't, then you then you can be in a real tough situation. So understanding what sort of model you're working under, understanding what your your leadership and the company is trying to accomplish, and then trying to kind of fit that in with the reality that you've been given. We had a really bad year. CEO came to me and said, What's the minimum number of people we can hire for support this year? And I'm like, I don't know. And I didn't ask good questions. I didn't say, what are our sales look like this year? What are we doing from a product perspective? Are we making any big changes? Those that those were questions I should have asked and information I should have had before I said any number. I said, Hey, it feels like we need probably eight or ten. He's like, I'll give you four. August 15th, we realized that we were grossly understaffed, but because of our onboarding process and the inability to train people while you're trying to triage stuff, it meant that from August until December, we just got hammered. It was terrible. That year I said, okay, this I can't do this again. We can't have bad information. I need to do a better job. And that's when I started developing the model.

Charlotte Ward: 6:41

That's it for today. Go to customersupportleaders.com forward slash seventy-nine for the show notes, and I'll see you next time.