GTM Spotlight: Customer Success

What KBCM Measures

Three survey years give a clean longitudinal view of CS capacity. Read the table and notice what kind of number every cell is: an input.

Where CS costs live, the survey’s only other CS question, is an accounting classification, not a performance measure:

And the outcomes CS is held accountable for, measured at the company level with no line back to the function:

Interpretive framing on three cells.

The SMB load rose from 100 to 110 accounts per CSM between the 2023 and 2024 data years: a 10% coverage degradation at exactly the segment where logo pressure is highest, part of the full series in Customer Success Coverage Data.

The revenue-at-stake spread inside “the CSM role” is 20x. The SMB book works out to roughly $11K of ARR per account ($1.2M / 110) against roughly $238K per account in Enterprise ($2.5M / 10.5). Calculated The survey flattens both into one job title and one capacity benchmark.

A quarter of the industry is already losing its base. The 25th-percentile NDR of 95% (2023 and 2024) means one company in four is in net dollar contraction (full NDR and GDR trend in Retention Data: GDR, NDR, and Loss), and the VC-backed median sits precisely at the 100% boundary. These are the companies for whom CS output measurement would matter most, and for whom the survey offers coverage ratios instead.

What KBCM Concludes

KBCM offers almost no CS commentary; the function’s most revealing treatment is definitional. The 2024 survey’s expense-allocation page defines Customer Success as “renewals + upsells,” distinct from Customer Support, “service + product assistance” (definitions, 2024 Survey, p. 37).

That is the survey’s clearest statement of what it believes CS does: commercial work on the installed base. Nothing else in seven survey years evaluates whether CS does that work well. One in three sales-org employees is a CSM (32%, 2025 Survey, p. 29); the survey benchmarks the productivity of none of them.

What We Conclude

The survey defines CS as a revenue function, then declines to measure its revenue

Hold the two structures side by side. Sales gets headcount and productivity, quota and attainment: capacity paired with output at every step. CS gets headcount, accounts, book size. Capacity, capacity, capacity. If “CSM Book of Business” were a real output metric it would come with a result attached: the NDR of that book, the save rate on it, the expansion sourced from it. It comes with nothing. The survey’s own definition says CS owns “renewals + upsells”; its own metric set makes both invisible.

Expansion ARR appears in exactly one place: inside the AE Productivity formula

(New Logo ARR + Expansion ARR) / quota-carrying AEs (2025 Survey, p. 31). This is the architectural fact the companion analysis’s Conclusion 2 ownership note establishes, and it deserves restating from CS’s side: expansion is the growth motion whose share of pooled new ARR crossed 50% for the first time in 2024 (42% → 45% → 52% across the 2022–2024 data years; 2023 Survey, p. 10; 2024 Survey, p. 9; 2025 Survey, p. 18), the same growth-mix shift examined in The Expansion Myth, and the survey’s data model assigns every dollar of it to Sales by construction. By segment in 2024, the share runs 48% below $10M, 47% at $10–25M, 53% at $25–50M, and 56% above $50M (2025 Survey, p. 18): expansion-majority growth is earned above roughly $25M ARR, exactly where CS books are deepest. Two ownership models are common in industry practice, split/handoff and CS-owned end-to-end, and a CS-owned expansion dollar has no cell to land in anywhere in this dataset. A reader absorbing the KBCM figures inherits a Sales-owned framing without ever being told a decision was made. Inference

CS cannot even be costed, let alone valued

Between 57% and 68% of CS expense is booked into S&M depending on the year, commingled with sales commissions and demand generation, with the remainder split across COGS and G&A (full cost-structure series in Cost Structure and Profitability Data). CS investment as a percentage of revenue, the single number a CFO would need to evaluate CS ROI, is unrecoverable from the survey by design. R&D gets its own OpEx line and a five-year trend; CS expense is buried inside three other functions’ budgets.

Meanwhile the highest-leverage retention lever is under the most pressure

The 2024-data economics (14–15% gross churn, derived from the 85–86% GDR medians, expansion coverage of churn at 1.07x per the companion analysis, and a 37-month new-only CAC payback) make first-year retention the highest-leverage dollar in the business. Calculated The survey’s answer across that same period: SMB coverage degraded from 100 to 110 accounts per CSM, and the field’s largest cost share moved deeper into the S&M line, where the next efficiency cut will find it.

The Blindspot

The survey has no concept of CS output. Specifically absent, in a series that asks Sales equivalents of every one of these:

• No NRR-per-CSM or GDR-per-book: the direct analogue of AE Productivity does not exist
• No CS-sourced expansion ARR: structurally impossible while expansion lives only in the AE formula
• No save rate or churn-prevented dollars: the defensive work that justifies the function is unrecorded
• No CSM compensation or quota-structure question: the survey asks AE quota every year; it has never asked whether CSMs carry one, which is precisely the question the two expansion ownership models turn on
• No CS cost line: allocation percentages across three other categories, never a percentage of revenue

Why the omission distorts conclusions: the survey’s cost-cut era narrative (Conclusion 3 of the companion analysis) shows profitability improving while retention stayed flat, and CS is the function whose investment level cannot be read out of the data. If CS spend per dollar of ARR fell during the cut years, flat retention is a resilience story: the churn floor held while its defense was defunded. If CS spend held steady, flat retention is an efficacy question. These imply opposite strategies, reinvest versus rethink, and the survey cannot distinguish them. Inference

Worse, the capacity-only lens shows CS to executives as one thing: a cost ratio to drive down. Accounts-per-CSM can always go up, and nothing in the data model registers what that costs. The 100 → 110 SMB move is the visible edge of a change the dataset can only score as an improvement.

The asymmetry also contaminates the expansion conclusion. Because CS-owned expansion is invisible, companies running that model successfully are recorded as having productive AEs. The dataset structurally cannot surface the finding that would matter most to CS leaders: whether CS-owned expansion outperforms. Inference

What a Complete Picture Would Require

The pattern: every proposed metric pairs an existing capacity number with the output it exists to produce. The survey already believes CS owns renewals and upsells; it wrote that definition itself. A complete picture would simply take the survey’s own definition seriously.

Related Spotlights

• Sales: the function that got the capacity-plus-output pairing CS was denied, and whose productivity formula absorbs the expansion CS may be sourcing
• Marketing: the other function with zero measured retention accountability, from the opposite end of the customer lifecycle
• Product: the origin of the churn CS inherits and cannot close alone; the capability-gap decomposition both pages call for is the same metric
• Support: the adjacent post-sale function, classified as COGS where CS is classified as S&M, and the early-warning data CS-run plays depend on

All certified values from the KBCM/Sapphire survey series. Analytical interpretations are the work of SuccessCOACHING and are not attributable to KBCM, KeyBanc Capital Markets, or Sapphire Ventures.