Subscription metrics to track, and how they connect

Every subscription business ends up with a dashboard full of numbers, and most of them are the same few facts seen from different angles. Once you see how they connect, you stop tracking twenty metrics and start watching the four or five that tell you something at your stage. This post walks through the core set, shows how each one feeds the next with one worked example, and ends with what to watch when.

Start with MRR

Monthly recurring revenue is the amount your active subscriptions bring in each month, normalized so an annual plan counts as one twelfth of its price per month. It leaves out one-off charges, setup fees and taxes. Almost every other metric here is MRR sliced, divided or compared over time, so it is worth getting right first.

Annual recurring revenue is MRR times twelve. It says nothing new; it is the same number in the unit that annual contracts and investors think in. If you mostly sell monthly plans to small customers, MRR is the more honest headline.

If you want to check your own figure, the free MRR calculator handles mixed billing periods.

Net new MRR: the four movements

MRR on its own tells you where you are. Net new MRR tells you why it moved. Every change in MRR over a month falls into one of four buckets:

• New MRR: customers who started paying this month.
• Expansion MRR: existing customers paying more, through upgrades, more seats or add-ons. See expansion revenue.
• Contraction MRR: existing customers paying less, through downgrades or fewer seats.
• Churned MRR: customers who stopped paying entirely.

Net new MRR is new plus expansion, minus contraction and churn. Two months with the same net new MRR can be very different businesses: one adding lots of customers while losing lots, the other adding few and losing almost none. Looking at the four parts separately is what tells them apart.

A worked month

Say you run a scheduling app. On October 1 you have 400 customers paying a total of $20,000 MRR. During October:

• 40 new customers sign up at $50 each: $2,000 new MRR.
• Existing customers upgrade and add seats: $620 expansion.
• A few downgrade: $200 contraction.
• 20 customers at $50 cancel: $1,000 churned.

Net new MRR is $2,000 + $620 minus $200 minus $1,000 = $1,420. You end October at $21,420 MRR, which is $257,040 ARR. You have 400 + 40 minus 20 = 420 customers. Every metric below comes from these same figures.

Churn rate

Churn rate comes in two flavors, and people mix them up constantly.

• Customer churn: customers lost divided by customers at the start. Here, 20 / 400 = 5%.
• Revenue churn: MRR lost to churn and contraction divided by starting MRR. Here, ($1,000 + $200) / $20,000 = 6%.

They diverge when the customers leaving are not average. If your smallest customers churn most, customer churn looks worse than revenue churn. If a large account leaves, the reverse. Track both.

Also split out involuntary churn, the customers lost to failed payments rather than a decision to leave. It is the easiest churn to reduce, because the customer still wants the product. Our guide to failed payment recovery covers how.

Gross and net revenue retention

Retention looks at the customers you already had at the start of the period and asks how much of their revenue is still there at the end. New customers are left out on purpose.

Gross revenue retention counts only losses: starting MRR minus contraction and churn, over starting MRR. Here, ($20,000 minus $200 minus $1,000) / $20,000 = $18,800 / $20,000 = 94%. It can never go above 100%, which makes it a clean measure of how well you keep what you have.

Net revenue retention adds expansion back in: ($20,000 + $620 minus $200 minus $1,000) / $20,000 = $19,420 / $20,000 = 97.1%. When NRR is above 100%, your existing customers grow revenue faster than you lose it, so the business grows even in a month with no new signups.

Note these are monthly figures. Retention is often quoted over a year, computed on the customers you had twelve months ago, so check which one a number means before comparing it to anything.

ARPU

Average revenue per user is MRR divided by paying customers. At the start of October that is $20,000 / 400 = $50. At the end it is $21,420 / 420 = $51. A rising ARPU means expansion and pricing are working; a falling one can mean you are winning smaller customers, or discounting too much to close them.

Customer lifetime value

Customer lifetime value estimates the gross profit an average customer brings in before they leave. Start with revenue: ARPU divided by monthly customer churn, $50 / 5% = $1,000. At 5% monthly churn the average customer stays about 20 months, and 20 months at $50 is $1,000 of revenue. The version to compare with acquisition cost multiplies by gross margin: at 80% margin, LTV is $800.

The value of LTV is the comparison: it tells you what you can afford to spend acquiring a customer. Treat it as a rough guide. A small change in churn moves it a lot, and a young business does not yet have enough history to know its real churn.

Trial conversion

If you offer a free trial, your trial conversion rate is the share of trials that become paying customers. Measure it by cohort: of the 200 trials that started in September, 36 paid, so September converted at 18%. Counting "conversions this month over trials this month" mixes cohorts and swings with your signup volume. Our post on free trial vs freemium goes deeper on measuring it.

How the metrics connect

Put together, the chain looks like this:

1. Trials convert into new MRR.
2. Existing customers produce expansion, contraction and churn.
3. Those four sum to net new MRR, which moves MRR and ARR.
4. Contraction and churn give you GRR; add expansion for NRR.
5. MRR over customers is ARPU; ARPU times margin over churn is LTV.

If one number looks wrong, walk back up the chain. A falling NRR is either more churn, more contraction or less expansion, and the four movements will tell you which.

What to watch at each stage

Just launched (first paying customers). Watch trial conversion and customer churn, customer by customer. At this size every cancellation is a conversation worth having, and percentages on small numbers jump around too much to mean much.

Finding growth (tens to hundreds of customers). Watch MRR and its four movements every month. This is where you learn whether growth comes from new signups or from customers paying more, and whether churn is eating it. Start tracking involuntary churn separately.

Scaling (hundreds of customers and up). Watch GRR and NRR on annual cohorts, ARPU by plan, and LTV against what you spend to acquire a customer. These are the numbers that tell you where to put the next dollar of spend.

At every stage, resist adding metrics nobody acts on. A number earns its place on the dashboard when a change in it would change what you do.

How yRecurring handles it

yRecurring computes these from your billing data rather than a separate analytics tool. The analytics dashboard charts MRR, ARR, net new MRR, MRR growth rate, ARPU, and active subscriptions and customers, alongside revenue, payments, refunds and credits. The reports library lets you click any number through to the rows behind it and export them to a spreadsheet for anything the dashboard does not chart. Plans are a flat monthly price, see pricing.