CSM to Customer Ratio: Why 1:144 Misleads

The short answer

A good CSM to customer ratio is one where every account gets a planned step each quarter and CSM hours go to the accounts that need a person. There is no universal number. The famous 1:144 counts accounts, not the work behind them.

So stop debating the ratio. Count touches, then report which accounts actually got one.

Where 1:144 comes from

The number comes from Gainsight's July 2022 benchmark of its US enterprise customers above $100M ARR. Gainsight competes with Base AI, which publishes this post, so weigh that as you read. It is still a fair starting point.

The averages, by touch tier:

• High touch: 22 accounts each
• Mid touch: 49 accounts each
• Low touch: 144 accounts each

Here is the part to read closely. Gainsight said low-touch accounts "likely have automated CS processes" to "help the CSMs scale". So 144 may not measure what one person can handle. It may measure one person plus a set of programs.

Benchmark your team against it and you may be comparing a CSM to someone else's CSM and their machinery.

More accounts, less revenue

The second problem shows up when you look at money. Gainsight set low touch at under $10K in annual contract value and high touch at over $100K. Multiply that by the account averages, and the arithmetic, which is ours, looks like this:

The CSM with six and a half times the accounts carries less revenue. Account count tells you little about either the load or the value in a book.

What is a good CSM to customer ratio?

The CSM to customer ratio is the number of accounts assigned to one customer success manager, usually set by segment. A good ratio is one where every account gets a planned, account-specific step each quarter. Two teams at 1:144 can be running completely different businesses.

Picture two CSMs with 144 accounts each.

The first works the 30 loudest accounts. The rest wait for a renewal date or a ticket.

The second runs programs that reach all 144, and spends personal time on the 25 accounts where usage says a person is needed.

Same ratio. Different business. If you can't say what your 144th account received in the last 90 days, your ratio describes your org chart, not your coverage.

How do you plan CSM capacity?

Plan customer success capacity with a touch budget instead of an account count. For each segment, multiply accounts by the touches each one needs per quarter and by the CSM minutes per touch. Compare the total with the CSM hours you have. The gap is the list of accounts that won't get what you planned.

A touch is any step that reaches one specific account: a call, a business review, a success plan update, an adoption tip tied to its usage. A newsletter to everyone doesn't count.

Here is an illustrative book of 150 accounts, with one CSM and 120 hours a quarter:

• Strategic: 10 accounts, 6 touches, 60 minutes each. 60 hours.
• Mid-market: 40 accounts, 3 touches, 30 minutes each. 60 hours.
• Long tail: 100 accounts, 2 touches, 15 minutes each. 50 hours.

That is 170 hours of work against 120 available. The missing 50 don't come out evenly. They come out of the long tail, because strategic accounts call back and long-tail accounts don't.

The plan says 1:150. The budget says 100 accounts get nothing this quarter.

Move touches, not accounts

The usual fix is to move accounts into a lower tier. The ratio goes up, and the touches those accounts needed stop being anyone's job.

The better move is to sort the touches. Every touch belongs in one of three lanes:

• On schedule. Fixed steps with no judgment, like a welcome sequence or a renewal reminder. Base AI's glossary covers these under zero-touch automation.
• AI agent decides. Steps that need a decision per account but not a relationship, like which tip to send. On the Base AI Customer Growth platform, an AI teammate decides the next step, takes it and brings in a person when needed.
• Person owns it. Steps where a person changes the outcome, like a renewal at risk or an executive conversation.

Our guide on what to automate in customer success gives four tests for sorting them.

Now rerun the same book. Strategic stays the same. Mid-market moves one status update to an agent. The long tail moves both touches off the CSM's plate.

CSM demand drops from 170 hours to 100. The spare 20 hours go to the accounts an agent hands back, like the customer whose usage fell off.

One warning. A lane is only as good as the data behind it. Our guide to customer success automation lists what a playbook needs before an agent can run it.

How Okta covered accounts with no CSM

Okta's story is the clearest live example we can point to. Okta is a Base AI customer, so weigh it with that in mind.

Its Digital Success team needed success plans for accounts without a CSM. Plans built by hand didn't scale, and the team set a hard limit: no added manual work for CSMs.

Customers now pick their goals from a curated list. Live usage data places each one on Okta's maturity model and generates a plan. Completed actions check off automatically.

The results, as Okta's story reports them:

• Success plans 100% automated for digital-only accounts
• Weekly adoption of an underused feature tripled in one week
• No added manual work for Okta's CSM or Scale teams

This is automation, not an AI agent: each plan follows Okta's own model. A ratio would never register the change, because no CSM was assigned to those accounts at all. That is the idea behind success plans in Base AI's retention solution.

What should you report instead?

Report covered accounts instead of the CSM to customer ratio. Covered accounts is the share of accounts in a segment that received at least one account-specific program step in the last 90 days, from any lane. It is our definition, not an industry standard, and it shows what the ratio hides.

Three rules keep it honest:

1. The step depends on that account's goals, usage or stage. A launch announcement to everyone doesn't count.
2. The step reached the customer. A task in a CSM's queue isn't coverage.
3. You split the number by lane, so you can see where your people spend their time.

In the example book, the long tail goes from none of its 100 accounts covered to all of them. The ratio never moves.

Pair coverage with an outcome you already track, like adoption or renewal rate. Otherwise a team can raise coverage by sending steps that change nothing.

If your activity data doesn't share an account ID, that is the gap the Base AI platform closes. Then run the same count on new accounts: Base's onboarding solution tracks milestones per account and flags the ones that need attention.

Start with one query

You don't need a new tool to start, if your CRM and program logs share an account ID. Pull the last 90 days of activity for your lowest-touch segment. Count the accounts that got at least one account-specific step, and divide by the accounts in the segment.

That is your first covered-accounts number.

Keep the 1:144 for the board deck if you have to. Just put this number next to it.