For years, our hiring strategy for the HubSpot customer support team was simple. When support reps were overwhelmed with the volume of work, we'd hire another one. Simple, right?
This simple growth model worked to get our team to about 25 people. We had funding, an accelerating sales team and customer count, and awesome technology. Our startup hopes and dreams were coming true.
And at least for me, I never even saw the growth storm coming. I was "managing hyper-growth" day-to-day in support, making sure the team was happy and ensuring customers were served quickly -- all part of a customer success manager's regular duties.
The storm was a combination of rapidly accelerating customer acquisition plus the launch of a totally new platform, and it nearly destroyed HubSpot in 2013. As we signed up customers at an ever-faster rate and rewrote the platform at the same time, every one of our simplistic hiring models broke with violent speed.
Within one week of the growth storm hitting, support was completely overwhelmed by volume. Our ticket wait time went from "basically none" to "just call us; don't bother submitting a ticket this week." Within one month, customer churn tripled.
In general, users were furious, the customer team was shattered and out of control, and the entire company was scrambling. And me? I had just taken over the support team at HubSpot.
It took us four months to recover. Professionally, these were the longest days of my career.
Plan for Scale Before it Happens, Not After
But how did we go so quickly from growth dreams coming true to a four-month customer support nightmare?
Spooky how fast things fall apart when they're broken, right? Even spookier is that, in retrospect, it was 100% avoidable. We needed to define how we scale before our system broke, not after.
We could've managed the pace of customer acquisition and platform rewrite smoothly if only we had understood the input and model behind our support team and planned for it all. But, we didn't, and we got crushed.
If you're in the early or mid-stage of your growth, this is the right time for you to understand how to scale support so you don't end up like me in 2013. You want to have your model and understanding locked in, so when your sales and marketing teams start hitting their stride in the growth stage of your business, or you rewrite your whole platform (or both), you're ready -- and you'll avoid churn, angry customers, and a sad tale of scale.
How Should You Hire for Your Customer Support Team?
How to Build a Hiring Model for Your Customer Support Team
Your hiring model doesn't need to be complicated. In fact, simpler models tend to be more nimble. The model below alone should be enough to get you through your growth stage smoothly (and as you hit the scale-up stage, you'll probably want a little more nuance):
Let's talk through each of these data inputs -- how to build them up, and some additional considerations to keep in mind when starting your calculations.
Incident rate is a simple ratio that helps you understand, "If my sales team is planning to add 50 new customers this month, my team is going to get X new tickets/month in perpetuity as a result."
There are two scenarios in which you may want to add additional nuance to your incident rate model:
If you have customer segments with wide variations in behavior:
For example, if you have a freemium product that has a very low support load, you'll need to have two incident rates and customer-add line items (e.g. 0.05 tickets/month/customer on freemium vs. 1.00 on another product).
If you have front-loaded support demand to your customer experience:
For example, new customers have a radically higher incident rate than existing customers (for example, if new customers have five tickets/month/customer in the first month, then that dips to one ticket per month).
If these two points don't strike you as familiar when you think about how your business works, omit them and keep it simple.
You may also want to consider:
- What is your net customer count growth rate (i.e. adds minus churns)?
- How predictable/steady is that growth (i.e how good is your sales and retention forecasting?)
Again, be careful about unnecessarily complicating your first real model. These are just things you'll want to start thinking about as it makes sense for you and your business to have more predictability around this model.
Hopefully, you have a central system in place that is managing your customer support queue! But even for companies that do, they often have some "extra" support work lying around.
For example, legacy customers might still email your founder directly when they have a problem, or maybe there's an old support email that gets occasional use and one of your support reps manages it on the side.
Part of modeling your growth is to not have any more "on the side" work. This is a good time to do a quick check and make sure you know all of your ticket streams (even better if you can consolidate them), and get this count right.
This is a universally reachable bar that every person you hire can hit. So how do you determine this number? Take a look at historical reporting, and be conservative.
If people tend to do 300 tickets in a month, take that down 20% or so to 250. Keep in mind that rep working time comes with "shrinkage" -- hours lost to meetings, breaks, training, knowledge base article-writing, etc. Assuming 20% shrinkage time is a good approach in your first model; you want to keep things loose enough to function smoothly.
As time goes on and your team becomes more sophisticated, you'll want to sharpen your model around performance and shrinkage. The "average rep" may not ever exist, and it's probably not realistic to expect every new hire will be trained for three months and then turn into your new top performer. Plan accordingly.
Total Support Reps Needed =
Total Monthly Tickets / (Tickets/Reps/Month)
Simple math right? But, one issue: what to do about that fractional rep? When do I tick over and add the next hire vs. stretch the team? My advice is to always advocate strongly for hiring ahead of your model.
Have a model that demands 6.23 people? You should be pounding the table with your CFO to hire a seventh rep. Customer support teams with high stress and high turnover almost always have a leader who isn't advocating for hiring ahead of the model -- and the small incremental expense will be worth it to your team experience and customer experience.
The above is a strong warning to fast growth companies: Get a hiring model, and keep that model simple. If you're lucky, you'll get a perfect storm of growth as your start-up scales up, and if you're prepared in support, your customers and your team will weather it well.
I spent the last seven years of my career building HubSpot's customer support team. Today, I'm leading the development of HubSpot's suite of tools designed for past-me -- a new product line called Service Hub (coming in 2018). If you'd like to be notified about this release, you can sign up for info here.
A version of this blog post originall appeared on Sales Hacker.