6 How to track customer success: the 9 metrics you should know
You’ve got more customers than last month. And the last user you talked to was really happy with your support. Things are going great, right?
Well, maybe. Because if you aren’t measuring, you’ll never really know.
So where do you start? What data should you track? Which metrics really matter? What key performance indicators (KPIs) should guide your ship?
Wait, metrics vs. KPIs? Good question.
Metrics are just numbers that measure some type of behavior—visitors to your site, number of logins to your platform, and so on. KPIs are metrics that reflect some objective or key result you’re trying to hit. So all KPIs are metrics, but not all metrics are KPIs.
The bad news: there’s a lot of numbers you could be tracking. And there’s no god-given KPIs that work for everyone.
The good news: there are some staple metrics that just about every team should have their eye on. And they fall into three big buckets:
- Business metrics
- Support metrics
- Product usage metrics
But hold on—why measure in the first place?
Measure to learn, not to track numbers
The most important question is not what you should measure, but why.
So don’t set KPIs without knowing why they’re important—how they connect and contribute to larger company goals.
Just ask Avinash:
If you don’t actually know what you’re trying to achieve, it’s also hard to measure success. Try it, you’ll see what I mean.
— Avinash Kaushik (@avinash) December 9, 2014
And remember: metrics are only meaningful against some benchmark or objective.
Think $20M in sales during Q2 is impressive? Maybe for your company, but if you’re ExxonMobil, you’re probably considering bankruptcy.
So before you start tallying numbers across your whiteboard, ask yourself:
- What are my most important objectives?
- What metrics reflect those objectives?
- What benchmarks should I be measuring against?
Now get your slide rules out, it’s time to talk metrics.
Start with the big picture
If your customers aren’t sticking around, you’re not achieving customer success. So if you had to pick just one metric to improve, the general answer would be clear: churn.
Problem is, there are lots of ways to measure churn—customer churn, revenue churn, and more.
How should you do it? That’s for you to decide. But let’s start from the top.
1) The big one:
Net MRR Churn
Customer success is about retention and growth. So if you have to pick one overarching metric—your key performance indicator—this is it: net MRR churn.
Why? Because net MRR churn not only tracks lost accounts, but also changes to different plans.
Here’s how you calculate net MRR churn:
First, let’s make sure each part is clear.
MRR Churn: Revenue lost from customers who cancel.
Contraction (downgrade): Revenue lost from paying customers who downgrade to a lower-paying plan or receive a discount. Counter-intuitively, this also includes people who switch from a monthly to a yearly plan, since yearly plans are often discounted.
Expansion (upsell): Revenue gained from paying customers who upgrade to a higher-paying plan or when a discount expires.
Reactivation: Revenue gained from customers who jumped shipped that you’re able to pull back on board.
MRR: Monthly recurring revenue at the beginning of the period–month, quarter, year, whatever makes sense for your business.
So, imagine at the beginning of January, your MRR is= $100,000.
During the month, you lost $4,000 to churn, and another $1,000 from downgraded plans. But you also managed to upsell $8,000 to existing accounts, while reactivating another $2,000.
Plug it all in and…
Wait, a negative churn? It’s possible, and it’s what you—and investors—want to see if you’re aiming at high growth. Boom! A good month.
You want that too? Negative churn takes good retention strategies and carefully targeted upsales. But it all boils down to providing customers more value.
If you’re consistently hitting <1% net MRR churn, you’re on the right track.
2) Net Promoter Score
NPS measures customer loyalty and sentiment about your brand. It’s the pulse of your company taken with just two questions:
- How likely are you to recommend (my company/product/service) to a friend of colleague?
- What is the most import reason for your score?
Question 1 gives you a quantitative measure of how customers feel about you. Separate responses into three categories: promoters (9 or 10), passives (7 or 8), and detractors (0-6). Subtract the detractors from the promoters and you have your company’s NPS.
What do you do next?
- Find out why promoters are so happy, and keep them that way.
- Aim to push passives to promoters.
- And for goodness sake, have a talk with detractors to see what needs fixed.
This is where Question 2 steps in with the why—the reason for their feelings. Group responses into similar why buckets. Prioritize based on the number of related concerns, business values and goals, and available resources. Then make changes.
How often should you measure NPS? Good practice says check in at three points along the customer journey: at 1 month, 6 months, and 1 year.
But remember to send it out at least 10 days before each of these milestones. Why? So you still have time to proactively reach out to detractors and passives to stave off potential churn.
Some NPS tips:
Nurture through support and education
The reason you provide support is to keep your customers moving toward their goals. So make sure the customer leaves every interaction feeling better than when they arrived.
But feelings aren’t enough. As Joe Daniel, director of Customer Success at Chargebee, puts it:
So what data should you use?
Support can be measured in more ways than you can hum a tune. Ticket volumes, reply times, resolution times, all of which can be averaged across support reps, topics, or user plans. Yes, it’s exhausting.
Here are the key support metrics you’ll want to keep your eyes on.
3) Time to first response
First response time is how long it takes to respond the first time a customer gets in touch.
Why should this be a support team KPI? Because first impressions matter, big time.
If you can answer with a direct solution, do it now! And if not, tell them you got the message, you’re looking into it, and you’ll be back soon.
Warning: focusing on this metric can encourage thoughtless automation.
If a customer sends a query that can be answered with “It’s in the hamburger menu,” tell them that in the first mail. Don’t auto-reply with “That’s a great question, you lovely customer! Back to you soon!” This might keep your first reply time low, but it will come around and smack you in face when users answer your customer satisfaction question.
Of course response time matters to all customer queries. So don’t get sloppy on the second response.
4) Customer satisfaction (CSAT)
How happy was that customer with the response you gave? Was it fast enough? Did your help center have the answer they were looking for? Was it easy to find?
There are lots of ways to measure satisfaction. It can be as simple as one question or as in-depth as a college entrance exam. Your best bet is this question:
“Are you happy with the support you received?”
Then follow it up with a simple ‘YES/NO’ response option. Because the more options you provide, the less likely they are to respond.
5) Help center article rating
Two-thirds of customers prefer self-service help centers over speaking with a customer service rep. And 91% say they’d use an online knowledge base, as long as it’s easy to find the info they need.
The bottom line: a better knowledge base is a huge cost-saver.
But only if customers find what they want. The easiest way to keep tabs on how well you’re doing this? One question:
“Was this article helpful?”
As with customer satisfaction, stick with a simple ‘YES/NO’ response option. If it’s a “yes,” say thanks! If they answer “no,” give them an open-ended follow-up question to find out why.
6) Ticket deflection (ticket volume per user)
Another metric to keep an eye on: the number of tickets received per user.
If this number is 1, then each user submits an average of 1 ticket. If it’s 2, then each user submits 2 tickets. If it’s 0.5, it’s one ticket per two people.
You should aim to keep this number
The fastest way to drop this number? A high-quality help center. Because better educated users mean less need to submit tickets.
7) Most common issue
Remember all those labels you’ve been using to tag customer interactions–“feature request”, “UX pain”, “bug report”, “billing issue”… You have been tagging, right? Good.
Now find out what’s eating up your support team’s time:
This will help orient your product team on how to prioritize their projects.
Peek in on the product
Now it’s time to peer under your own hood. How are users interacting with your product? What features do they use, and neglect? And which behaviors matter to your business?
8) User engagement metrics
Look for behavioral patterns tied to business outcomes. It might be frequency of logins, or number of features used. If you’re a product like Facebook, your metric might be daily active users.
The important thing? Identify the trends that matter to your business. Find behaviors correlated with churn or expansion or customer satisfaction. Then use that info to target customers appropriately.
9) Time to value
Another place to look is customer onboarding—how efficiently users get value from your product.
For example, Groove HQ finds that free users are 80% more likely to convert to paid customers if they complete all the onboarding prompts within 24 hours of signing up.
To figure out the numbers that matter to you, you’ll have to dig into your own data. Find those magic ratios. And then set targets for your new customer onboarding process.
Some data tips for beginners:
“The more you work with data, the more you’ll realize what you can ask of it.”
Data can be a huge source of insight. It can also be a huge headache. Here are a few things to keep in mind as you get started:
- Identify your KPIs. Without these, you shouldn’t do anything.
- Be data-informed, but understand that data alone doesn’t hold all the keys.
- Understand why you’re measuring before you fill in that spreadsheet.
- Focus on fewer metrics that matter more. Don’t try to track everything.
- Start general, and segment into more refined cohorts as you move along.
- Create tags or labels for support tickets. “Bug report”, “feature request”,
“UX pain,” and “billing issue” are some good ones to start with.
- Group frequently occurring tags into buckets to help prioritize requests.
- Think in volumes, not individuals. You can’t prioritize every request.
With today’s tools, you can track nearly everything with virtually no effort. And while it’s nice to have the options, don’t try to measure everything at first. Pick a few key metrics to focus on.
Remember: The right metrics depend on your business needs. Start at the top, and break it down into sub-metrics that contribute to the bigger picture.