Two hundred tickets a day. Average time until first reply: two minutes. Average number of replies: three (okay, three and a half.) Response time unaffected by nights and weekends, because your support team is sprawled around the globe.

Sound ideal? Terrifying? Unattainable?

Honestly, it should sound like what it is: a bunch of numbers. They can tell you that you’re getting back to customers quickly, but they can’t tell you whether or not those exchanges are successful. That doesn’t mean they’re not important—while data doesn’t deliver qualitative judgments, it can help you make reasonable deductions based on common sense and context.

When it comes to the performance of your customer service team, there are a few key metrics that will keep you pointed in the right direction.

Ask the right questions.

Asking the right questions means more than just, “Is my team performing well?” By itself, data can’t tell you that. Switch things around and ask yourself, “What type of work environment will allow my team to perform best?”

Sketching out an ideal work environment requires defining quantitative measurements: how many conversations should each person manage? How often should they be utilizing established feedback channels? What is the busiest time of day and how many people should be on duty?

Define and set those metrics with your team, but be flexible. As the workload evolves along with your product, you’ll likely need to set new goals. Making it part of a group conversation ensures total transparency and gives your team both a barometer to measure its own performance and a set of tangible goals for which to aim.

Pay attention to average volume.

There’s no template threshold for “too much” in customer support. Depending on the type and complexity of emails you tend to receive, exhaustion could set in after as few as 25 replies or as many as 75. But finding and setting that threshold, as a team, goes a long way in ensuring that you have the energy and emotional resources to engage with each conversation productively. It sounds obvious, but it’s surprising how often this number slips beneath the more obvious-seeming margins like average time to first reply and average time to resolution.

Help Scout Team and User Reports

The joint use of Help Scout’s Team Report and User Report will help you dial down into how much work—and what kind of work—each of your team is doing. When looking at the numbers keep context in mind. When reviewing the Team Report, you might want to reward the team member with the most customers helped, but make sure you’re balancing that number with the rest of the data that the report makes available.

For example, if you see a team member serving a lot of customers, but with very few replies per person and a very low average response time, that may actually be a sign that he or she is carelessly blowing through tickets.

User moving too fast

On the other hand, a high number of replies sent and a low number of customers helped might indicate that a team member is getting unfairly overloaded with higher complexity issues, thus reaching an exhaustion point more quickly than others.

User moving overloaded

In both cases, the best way to solve potential problems is to use these data points as an entry into 1-1 conversation with the team member. Other factors, like anticipating an influx of emails (new product launch, a big announcement) and tracking what days and times are most busy, can help you fluidly schedule your team so that no single person is disproportionately bearing the load.

The takeaway? Pay attention to how much work each person is doing. An overworked team member is an underperforming one, and while you can’t guarantee high quality support with an evenly distributed workload, it’s a necessary first step.

Empower the team to self-diagnose.

Pummeling users with satisfaction surveys is a good way to end up with unreliable data and an unhappy team. Nobody wants more emails in their inbox, and a customer could rate an interaction as being poor when a problem is unsolvable due to issues beyond the control of your team member. The outcome is that your team may end up feeling powerless and self-critical, which then leads to (you guessed it) subpar support. Empowered teams, who feel like they have space to both celebrate and fairly criticize the outcomes of their interactions, are also teams who are ideally positioned to—cheerfully!—handle unhappy customers.

One solution for baking a feedback loop into your day-to-day? Create a series of team-only tags that rate a customer service experience according to a vocabulary you establish. For example, “Hooray!” could indicate an above-and-beyond experience, where someone went the extra mile to solve a problem, and “Lost Cause” could indicate a customer who simply refuses to be helped. (We’ve all had ‘em.) Whatever tags you go with, it’s important to determine and define them with your entire team. Everybody should have input and everybody should feel confident about what means what. You’re essentially learning a new, private language together. When you ask a teammate why an experience was a “Total Spelunk,” both parties need to understand the important nuance of that statement.

Team-only tags

Keep things simple: 5-7 tags is a digestible number that can cover a full spectrum of customer experiences without giving you too much to remember.

With a few weeks of data under your belt, trending tags can be a helpful barometer as to who is struggling and who is succeeding. They’re also a great jumping-off point for open conversation. You could ask, for instance, “It seems like you’re rating a lot of issues as High Complexity. What are some examples?” From there, you can determine if that person is being unfairly loaded with super complicated issues or if there’s a gap in his or her fundamental understanding of the product. Then, you can solve from there. That might mean giving a refresher course on the product and/or consciously divvying up the workload differently.

At the end of the day, the best way to guarantee high-performance customer support is to build a transparent company culture that emphasizes consistent and productive communication. Processes—and data points—have to evolve as a company grows in order to stay relevant, and the feedback you receive from your team will be a critical part of that process.

Data is a great weatherman, but it’s never going to completely replace good ol’ fashioned one-on-one experience. Reality bats last. Incorporate both into your growth process, and you’ll be well-ahead in the customer support game.

Cassie Marketos

About the author: Cassie Marketos runs the show at Dollar A Day, hosts an annual co-working event called LadyParts, and works with WNYC and Longform.