GTM Spotlight: Professional Services

What KBCM Measures

Four instruments, all live, spread across the Retention and Contracts/Pricing sections. PS is not under-measured. It is unrecognized, which is a different failure.

1. Gross dollar churn by PS attach, the dataset’s steepest causal gradient, measured in four survey years:

The intermediate bands step down through the middle in every year (14.3% to 11.1% across the 2018 bands; 13%, 11%, 10% in 2020), and the 2023 survey’s rebanding (PS as a share of upfront ARR) carries one honest wobble: its top band prints 11%, above the 8% of the band beneath it. The endpoints never wobble. Zero-attach companies churn at 1.7x to 2.8x the rate of the lowest-churn attach band in every survey year that asks. Inference

2. Attach by GTM motion, a row in the field-versus-inside comparison:

Two years, one shape. Attach is not a delivery decision made after the sale; it is a property of the GTM motion, fixed at the deal. That matters later on this page, because it means the churn gradient starts forming before any post-sale team meets the account.

3. PS as a share of first-year ARR, by target customer (2018 data, 2019 Survey, p. 35):

The larger the customer, the more of first-year ARR goes to services. That gradient is the front end of the attach-and-churn story above: bigger buyers take on more implementation, and implementation is what the churn data rewards.

4. PS as a revenue-composition line (2025 Survey, p. 14):

Through the actuals the line rose. The forecast reverses it, down to 7% by 2026E, projecting a cut to the one line the churn data says is working. The Forecasting a Cut to the Cure section below takes that up.

That is the measurement inventory. Here is the recognition inventory beside it: no PS margin question. No PS headcount. No question about who owns services. No expense-allocation row: the survey asks where Customer Success cost is booked and where Support cost is booked (2024 Survey, p. 37, among other years), and has never once asked where PS cost lives. PS has no expense line of its own.

What KBCM Concludes

Nothing. The certified series contains no KBCM or Sapphire commentary on Professional Services in any year: the churn-by-attach chart is printed, four times, without a sentence of narration attached to it.

Absence of commentary is not neutrality. The churn cut sits in the Retention section as a cross-tab, the attach rate sits in Contracts/Pricing as a deal attribute, and the revenue line sits in the composition stack as mix. Filed that way, PS reads as a pricing feature of contracts rather than an organization that does work, and the reader inherits the framing without ever being told a decision was made.

What We Conclude

The survey’s strongest causal lever has no functional home

Run the simple math. The 2018 spread is 10.5 points of gross churn between zero attach and high attach (16.3% against 5.8%), Calculated and the churn floor the companion analysis documents has not improved in any measured year, and runs 14–15% on the current edition (see The Churn Tax). One variable spans nearly the full height of the floor. The >50% attach cohorts print 5% to 9% churn in the three survey years that use that band: below the floor the series treats as structural. Causality deserves its caveats here (attach correlates with enterprise motion and longer contracts, and the high-attach cohorts are small), but no other operating variable in the dataset separates outcomes this widely, and the survey has never asked one question about the function that produces the separation.

PS is instrumented exactly as far as it behaves like an input

Every one of the four measurements records PS from the outside. Attach is wired to churn and to nothing else; there is no attach-to-expansion cut anywhere in the series. The survey knows what services suppress. It has never asked what services produce. With expansion supplying 52% of pooled new ARR in 2024 (2025 Survey, p. 18; see Growth and New-ARR Composition) and expansion’s coverage of churn thinned to 1.07x per the companion analysis (see The NDR Crisis), the untested half of the wire is now the half the industry’s growth model depends on.

Services attach is how the outcome sold becomes the outcome delivered

The companion analysis puts one condition on CS’s ownership of the churn side of coverage: CS controls churn only for deals sold on outcomes the product can deliver. Delivery is not self-executing. Implementation, integration, configuration, migration: that is the labor through which the sold outcome materializes or fails to, and attach is the dataset’s only trace of it. Read that way, the churn gradient stops being a pricing curiosity. It is the deliverability condition surfacing in the data: accounts that received delivery work churn like retained customers, and accounts that received none churn at rates near unretained cohorts. PS is a structural input to the 1.07x coverage margin, not a revenue sideline.

The industry is forecasting a cut to its proven churn suppressor

The revenue-composition line rose through the actuals, 8% to 10%, and then the forecast reverses it: 7% in 2026E, roughly a 30% reduction from the 2024 actual. Calculated Set that against the gradient. The same respondents whose pooled data shows high attach coinciding with sub-floor churn are planning to shrink the services line, and the survey’s architecture guarantees the consequence will be unmeasurable: no PS margin row to price the cut, no attach-to-expansion wire to catch the damage, only gross margin, which will improve. The companion analysis documented this exact illusion once already, on the R&D line: margins rose, retention did not.

The Blindspot

The survey has causal evidence without a causal agent. It can show what services attach does; it can say nothing about the function that does it. The absences, in a series that measures four PS quantities:

• No PS gross margin: revenue share exists; profitability was never asked, so services can only be read through the software-margin lens, as dilution
• No PS headcount: the survey counts CSMs and quota-carrying AEs; delivery capacity is uncounted
• No ownership question: whether services report into CS, Sales, or a standalone org is invisible, so the industry’s actual delivery structures cannot be benchmarked
• No expense address: CS and Support each get an allocation question; PS cost is booked somewhere in every respondent’s P&L, and the survey has never asked where

The distortion is directional, not random. A function with a revenue line and no margin line looks like mix. A function with a churn gradient and no headcount looks like a contract term. Every lens the survey offers renders PS as something to minimize, and none can register what minimizing it costs, which is how a dataset containing a churn spread as wide as 2.8x comes to host a forecast that cuts the spread’s cause. The churn-floor conclusion inherits the same flaw: the series finds that no strategy has broken the 14–15% floor while carrying, unconnected, the one variable whose high band lives below it. Inference

What a Complete Picture Would Require

The through-line: the survey already trusts this function’s evidence enough to print it four times. The proposals extend to PS the courtesy extended to every function the survey recognizes: a cost, a capacity, an owner, and an output.

The survey measures PS four ways and asks no question that recognizes it as a function. That is an omission, not an oversight, and the evidence for reversing it has been sitting in the survey’s own pages since 2019.

Related Spotlights

• Customer Success: the scorecard that inherits whatever churn services delivery prevents or fails to prevent; the deliverability condition binds these two pages into one argument
• Support: the other unrecognized function; PS shapes outcomes before go-live, Support detects them failing after it, and the survey measures neither as a function
• Sales: attach is set at the deal, and the field-versus-inside gap makes it a GTM-motion decision before it is a delivery one
• Product: outcomes sold beyond deliverable value surface as churn on someone else’s scorecard; services are the labor that closes that gap, or doesn’t

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.