It's budget season. Your CFO wants to know how many CSMs to hire next year, and you reach for the ratio you heard at a conference: 50 accounts per CSM, or was it 2 million in ARR? The trouble is, that figure doesn't survive long once it meets the reality of your customer base. Gainsight CEO Nick Mehta suggests a different approach: triangulate. Here are his three methods, with benchmarks and worked examples.
'What's the ideal account ratio for a CSM?' In a LinkedIn post published in autumn 2023, Nick Mehta said this is one of the questions he gets asked most often, and that it comes back every autumn when companies build next year's budget. His summary: a very simple question with a complicated answer.
The stakes are real. Under-resource your team and your CSMs are permanently short of time, quieter accounts slip under the radar and churn creeps in unannounced. Over-resource it and Customer Success becomes a cost line that's hard to justify to the finance team.
Rather than hunting for THE right number, he recommends combining three lenses: the number of accounts, the ARR managed and the cost of Customer Success as a share of total ARR. Each one sheds light on a different constraint, and the answer comes from setting them against each other.
Why a single ratio isn't enough
The problem with a single ratio is that it assumes all your accounts are equal. A customer paying €8k a year and one paying €300k expect neither the same level of support nor the same frequency of contact. And they carry very different weight in your retention.
In Nick Mehta's framework, each method answers a specific question:
- the number of accounts measures capacity: how many customers can a CSM realistically look after with the time they have?
- the ARR managed measures value: how much recurring revenue rests on each CSM?
- the cost of CS as a percentage of ARR measures sustainability: how much can the business afford to spend on Customer Success?
Taken on its own, each method is blind to the other two. An account ratio can look reasonable while blowing the budget. A percentage of ARR can hold up on paper while your CSMs have only twenty minutes a month to give each customer. Hence his recommendation: triangulate.
Method 1: start from the time available per account
The first approach Nick Mehta describes is to look at how many customers a CSM can handle depending on the engagement model. He gives a rough guide: a high-touch CSM might look after 5 customers, whereas a low-touch CSM can take on 50 or more.
'Time and motion': putting a clock on the job
To set these ratios, he encourages running a 'time and motion' study, in other words quantifying how much time a CSM should spend on each customer. His reasoning:
- a CSM has roughly 160 working hours a month;
- time actually spent in front of customers accounts for around 25% of that, so about 40 hours;
- the rest goes on training, preparation, reporting, internal meetings and so on.
With 5 accounts, a CSM can therefore give each one about 8 hours a month. With 100 accounts, those 40 hours shrink to 24 minutes per customer per month. Nick Mehta draws the logical conclusion: a low-touch CSM managing 100 customers has to space out touchpoints, to every other month or once a quarter.
The question to ask: what cadence for which segment?
The real starting point, in his view, is the cadence your customers need under each engagement model. And he warns against a classic illusion: don't assume you'll be able to run Quarterly Business Reviews with every customer if your CSMs each manage 100 of them.
Do the maths. 100 QBRs a quarter is roughly 33 a month. Allow even one hour of preparation and one hour of meeting for each, and you've already used up more than the 40 hours of customer time available. Before anything else.
Nick Mehta also points out that these targets can vary with a CSM's fully loaded cost. In other words, the right ratio also depends on what a CSM costs you.
Benchmarks: median number of accounts per CSM
To give some reference points, Nick Mehta shared data from his company's customer base. It shows the median number of accounts assigned to a CSM by average annual contract value (ACV):
- ACV of $0 to $10k: 105 accounts per CSM
- ACV of $10k to $25k: 44.1 accounts
- ACV of $25k to $50k: 39.2 accounts
- ACV of $50k to $100k: 20.5 accounts
- ACV of $100k to $250k: 17.1 accounts
- ACV of $250k to $500k: 15.3 accounts
- ACV of $500k to $1m: 8.5 accounts
- ACV above $1m: 12.9 accounts
Two things stand out. First, the drop between very small contracts and everything else is steep: the figure falls from 105 to 44 accounts as soon as ACV goes above $10k. Second, between $50k and $500k, books of business stay within a fairly narrow band of 15 to 20 accounts. Above that, it's around ten accounts per CSM.
📍 Want to see what this looks like with your own numbers? Our calculator combines all three methods in two minutes and gives you an ARR-per-CSM range.
Method 2: think in terms of ARR managed per CSM
The second lens, which Nick Mehta sums up in a sentence: not all customers are equal in terms of spend. Your high-touch customers might average $500k in ARR, while your smaller customers sit around $50k. Counting accounts without looking at what they're worth misses the point.
The finding: high-touch CSMs manage more ARR
This is where his data gets interesting. On average, a high-touch CSM manages more ARR than a low-touch CSM, despite looking after far fewer accounts.
Here is ARR per CSM by ACV band:
- ACV of $0 to $10k: $525k ARR per CSM
- ACV of $10k to $25k: $772k
- ACV of $25k to $50k: $1,470k
- ACV of $50k to $100k: $1,538k
- ACV of $100k to $250k: $2,993k
- ACV of $250k to $500k: $5,738k
- ACV of $500k to $1m: $6,375k
A CSM looking after 105 small accounts carries about $525k of ARR. A CSM looking after 8 or 9 very large accounts carries more than $6m. Twelve times fewer accounts, twelve times more ARR.
An order of magnitude, not an exact science
One detail is worth highlighting. If you redo the calculation, these figures match (give or take rounding) the median account count from Method 1 multiplied by the midpoint of each ACV band: 105 accounts at $5k gives $525k, 8.5 accounts at $750k gives $6,375k, and so on. On the vendor's community, the chart is in fact presented as an estimate of ARR per CSM by ACV.
So these figures are an order of magnitude, not ARR actually observed in real books of business. They're still very useful for benchmarking your own ratios, provided you don't read them to the last pound.
What it means for your targets
The direct consequence of this method is that a single ARR-per-CSM target for the whole team doesn't make much sense. Asking a low-touch CSM to carry the same ARR as an enterprise CSM is the same as asking them to manage several hundred customers. ARR targets are set segment by segment, not across the team as a whole.
Method 3: the cost of Customer Success as a percentage of ARR
The third approach, and the simplest according to Nick Mehta: take the fully loaded cost of the CSM team (including CS Ops) and divide it by the company's total ARR. That ratio represents the overall cost structure of Customer Success.
Its value goes further than that. This budget constraint is what sets the account and ARR targets to aim for in Methods 1 and 2. In other words, you no longer start only from what your customers need, but also from what your business model can afford.
The benchmark from large US SaaS companies: 5.3% of ARR
At Gainsight's CXO Conference, Nick Mehta shared a figure: his customers spent an average of 5.3% of their ARR on the fully loaded cost of Customer Success.
For a SaaS business with €10m ARR, that means a CS budget of around €530k a year, including employer costs and CS Ops.
Jason Lemkin's view: a level to aim for, not a rule
SaaStr founder Jason Lemkin picked up this benchmark and set it against the rules of thumb he has championed for years for sizing CS:
- $2m ARR per CSM once at scale;
- $1m ARR per CSM during the scale-up phase, if you can afford it;
- an even more generous investment in the very early days, if cash allows.
He also notes that the companies in this sample are fairly large and enterprise-focused. His advice: invest in CS earlier if you can, and treat 5.3% as a level to converge towards once you reach scale. For him, the figure is mainly a reminder that at scale, CS teams stay fairly lean: 5.3% of ARR in fully loaded cost is actually a little below his $2m-per-CSM benchmark.
Moving from one ratio to the other
That point is worth unpacking, because it links Methods 2 and 3 through a very simple calculation. The ARR a CSM needs to carry equals their fully loaded cost divided by the share of ARR spent on Customer Success.
Let's apply it. At 5.3% of ARR, a CSM with a fully loaded cost of €90k needs to carry around €1.7m of ARR. To land exactly on Jason Lemkin's $2m, their fully loaded cost would need to be around $106k. Put simply, the more your CSMs cost, the larger the book of business the same percentage of ARR forces them to carry.
What the other ratios say
Nick Mehta's 5.3% isn't the only benchmark doing the rounds. But you need to compare like with like: some ratios measure cost, others book of business, others still headcount share.
- Fully loaded CS cost = 5.3% of ARR (Nick Mehta): average across companies at his CXO Conference, which are fairly large and at scale.
- $2m ARR per CSM at scale, $1m during the scale-up phase (Jason Lemkin, SaaStr): benchmarks drawn from his experience as an investor and founder.
- Support and Customer Success = 9% of ARR at the median (SaaS Capital, 2026 survey of more than 1,000 private B2B SaaS companies): note that this figure combines support and CS.
- Customer Success team = 11 to 13% of headcount (Serena, European SaaS Benchmark 2020, 108 European SaaS companies): this is headcount, with no account taken of salaries.
The European perspective from Serena
The benchmark from Serena, a French venture capital firm, is one of the few to offer a European view. In its 2020 edition, the CS team made up 11% of headcount in SaaS companies below €1m ARR, then 13% in every band above that, up to companies with more than €10m.
Let's break that figure down. In the €5m to €10m ARR band, median headcount is 93 people, which means about a dozen people in Customer Success. That works out at roughly €400k to €800k of ARR per person in CS, a long way from the $2m per CSM cited by Jason Lemkin. And at a fully loaded cost of €90k per person (the same assumption as in the example below), CS would represent between 11% and 22% of ARR.
A few caveats apply. The data dates from 2020, the exact scope of the 'CS team' isn't specified (support or onboarding may be included) and the calculation relies on medians. Serena also noted that European SaaS companies employ more people than their US counterparts for the same level of ARR.
The underlying message is still telling: mid-sized European SaaS companies appear to invest proportionally far more in CS than the large companies Nick Mehta refers to. That's consistent with Jason Lemkin's advice to invest more before reaching scale, and a reminder that the same ratio doesn't mean the same thing at €5m and at €100m ARR.
How to triangulate in practice
Nick Mehta's message boils down to one word: triangulate, then work out what makes sense for your business. In practice, it can look like this:
- Calculate your current CS cost as a percentage of ARR, including CS Ops.
- Segment your customer base by ACV band or engagement model.
- For each segment, define the contact cadence and time required.
- Derive a number of accounts and an ARR figure per CSM, then compare them with the benchmarks.
- Test the whole picture against your budget constraint, and make the call.
A worked example
Let's take a fictitious case, built for the exercise. A SaaS vendor generates €10m ARR from 700 customers, an average ACV of around €14k. We assume a CSM costs it €90k a year fully loaded (a working assumption).
- On the budget side (Method 3), if it aligns with Nick Mehta's 5.3%, it has around €530k, which pays for 6 CSMs. Each would then carry nearly 117 accounts and about €1.7m of ARR.
- On the time side (Method 1), 117 accounts for 40 hours of customer time a month comes to about twenty minutes per customer per month. And the benchmarks Nick Mehta shared put books of business closer to 44 accounts for contracts of $10k to $25k. Matching that would take 16 CSMs, a budget close to 14% of ARR.
- On the value side (Method 2), the same story: as an order of magnitude, ARR per CSM is more than twice Nick Mehta's estimate for this band.
📍 Want to see what this looks like with your own numbers? Our calculator combines all three methods in two minutes and gives you an ARR-per-CSM range.
The three methods won't necessarily agree. That's precisely the point of triangulating: the gap forces you to make a deliberate choice rather than live with one by default. Several levers are then on the table:
- accept a higher percentage of ARR, which Jason Lemkin considers sensible during growth phases;
- segment more finely, keeping high touch for the accounts that carry most of the ARR;
- rethink the cadence for the smallest segments, for instance by dropping systematic QBRs;
- increase the share of customer-facing time for CSMs beyond the 25% used in Nick Mehta's calculation, by cutting down on preparation and reporting.
None of these levers is right or wrong in itself. It all depends on your strategy, your stage of growth and what your customers actually expect.
Another important point: models are shifting. The vendor itself notes that these benchmarks move as companies adopt Scaled CS approaches. As soon as part of the customer journey is handled by something other than a person, the number of accounts a CSM can look after changes in nature.
Conclusion
There's no magic ratio, and that's probably the best news in Nick Mehta's approach. Rather than searching for the right number, he invites you to combine available time, value managed and sustainable budget, then look at where those three lenses diverge.
That's where the real decisions are made: which customers deserve dedicated support, what cadence to hold for the rest, and what level of investment to defend in front of the finance team.
Further reading
