The Metrics Brothers

Dave "CAC" Kellogg and Ray "Growth" break down one of the oldest productivity metrics in business and explain why, in the age of AI-native software, it has never mattered more. This episode covers the full arc from Frederick Taylor's factory floors to Cursor's $3.3M per employee, with the rigorous definitional discipline the Metrics Brothers are known for.
What We Cover:
The metric's 100-year history. Revenue per employee traces its roots to scientific management in the late 1800s, gained traction as a Wall Street efficiency screen in the 80s and 90s, and became a standard signal of business model quality in M&A diligence. The core math is simple: annual revenue divided by headcount. What is not simple is how you define the denominator.

FTE vs. employee: why the definition matters more than the formula. The E in FTE stands for full-time equivalent, not full-time employee, and that distinction drives real measurement decisions. How do you count a part-time contractor? What about 200 offshore developers on a third-party vendor's payroll? Ray and Dave walk through the practical choices, including why offshore headcount is almost never counted on a 1:1 basis and why that decision can dramatically change your benchmark comparison.

Public SaaS companies in 2025: the benchmark is $395K. Using the Benchmarkit SaaS 100 index (134 public SaaS companies), the median revenue per employee in 2025 is $395K, up from $327K in 2022, a 21% improvement in three years. ARR per FTE runs about 5-7% higher at $413K. The shift reflects the industry's move from growth-at-all-costs to efficient revenue growth.

Private SaaS companies: size matters. ARR per employee scales materially with company size. At the $5-20M ARR stage, the median is $144K. By $100M+ ARR, the median reaches $300K. The recurring-revenue tailwind from a large renewal base is a significant driver as companies scale.

AI-native companies have reset the benchmark entirely. Where the historical range for enterprise software was $200-400K per employee, AI-native companies operate at a fundamentally different level. Cursor reached $1.67M per employee at 60 people, and now runs at $3.3M per employee at 300 people. Midjourney is at $4.7M. Anthropic is in the $3-5M range on a run-rate basis. This is not a modest improvement over traditional SaaS. It is a 10x shift.

One important caution on the AI numbers. Many of the figures being cited by AI-native companies are monthly run-rate revenue annualized (last month times 12), not trailing 12-month GAAP revenue. When growth is compounding fast, that distinction can dramatically inflate the productivity figure. The Metrics Brothers flag this as a meaningful source of confusion in how the benchmark is being discussed today.

The AI tailwind may be temporary, at least in part. Current customer acquisition friction for AI software is unusually low, given experimentation budgets and departmental purchasing. As enterprise procurement tightens (74% of enterprise AI purchases now involve IT), GTM investment will likely increase, and revenue per employee for AI-native companies may stabilize or compress. Ray and Dave estimate that steady-state productivity is more likely to be in the 3-5x range over traditional SaaS, not 10x.

Revenue will replace ARR as the standard numerator. The rise of usage-based and hybrid pricing is rendering ARR less meaningful for a growing share of companies. Snowflake, Datadog, and MongoDB do not report ARR. As AI-native pricing models proliferate, Ray and Dave expect the industry to converge on revenue as the standard numerator across productivity benchmarks.

What about revenue per agent? Ray raises the forward-looking question: as AI agents take on SDR, sales, and other GTM functions, how do we measure agent productivity? Dave's take is that "revenue per agent" is likely a dead end, partly because agent instances are nearly impossible to count and partly because the right way to price and measure agents is to decompose their capabilities, not to anthropomorphize them as headcount equivalents.

The Bottom Line:
Revenue per employee is a deceptively simple metric with genuinely complex definitional choices underneath it. For B2B SaaS executives, the 2025 benchmarks are $395K (public) and $144-300K (private, depending on scale). For AI-native companies, the numbers are in a different category entirely, though some of that gap reflects accounting choices as much as true productivity gains. The metric is worth tracking closely, both as a board-level efficiency signal and as a leading indicator of business model quality.
See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

One word can reveal a lot about someone's analytical depth, and on this episode of The Metrics Brothers, Dave "CAC" Kellogg and Ray "Growth" Rike break down one of the most commonly misused pair of terms in metrics analysis: segments and cohorts.
Dave shares what sparked this topic, a Norwest benchmark report that used the word "cohort" when it clearly meant "segment," and explains why the mix-up matters far more than a simple vocabulary error. In this episode, Ray and Dave cover:
Segments vs. Cohorts Defined: A segment is a slice of data defined by a shared attribute such as company size, vertical, or deal size. A cohort is a group anchored to a shared event and tracked over time, such as all opportunities created in Q1 or all customers acquired in a given year. The two are orthogonal concepts, not synonyms, and confusing them can signal a lack of the numerical fluency that sharp operators and investors expect

Snapshot vs. Cohort Analysis: Standard dashboard win rates are fast and stable, but they only capture what crossed the finish line in a given period with no visibility into where those deals came from or how long they were in the pipeline. Cohort analysis rides along with a group of opportunities from creation to resolution, revealing how process and personnel changes actually affect outcomes over time

Win Rates and Pipeline Coverage: Ray walks through a real example where cohort-based win rate analysis exposed a breakdown in discovery quality after a Q3 process change, something a standard dashboard completely masked. Dave explains why pipeline coverage goals should not simply be calculated as the inverse of a snapshot-based win rate, and how close rate (a cohort-based metric) gives a more accurate picture of both yield and timing

NRR, GRR, and Customer Expansion: Dave makes the case that tracking ARR by customer acquisition cohort over time is far more predictive of long-term retention and expansion behavior than NRR alone, which only looks back one year. Ray adds how cohort analysis helped him identify a high-value expansion window between months 18 and 30 of the customer lifecycle, enabling smarter allocation of sales resources towards existing customers

Combining Both for Maximum Insight: The most powerful approach is a segmented cohort analysis, tracking time-based behavior across meaningful attribute-based cuts of your customer or pipeline data. Segments tell you what kind of customer. Cohorts tell you what happened over time. Together, they tell the full story.

If you use metrics to help inform decisions in your company, and have a goal to help build a culture of numeracy in your company, this is a great listen!
See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

The Metrics Brothers, Dave Kellogg and Ray Rike open with their introductory bit, which this week included Statler and Waldorf, the grumpy old men in the balcony from the Muppets, before diving into a thorough review of the Norwest B2B Sales and Marketing Benchmark Report, a102-page study published in November 2025, that included 177 participants (77 Norwest portfolio, 100 third-party VC/PE-backed).

Key Topics Covered:
Overall Report Assessment: Praised for its breadth, year-over-year trend data, and even split of marketing (40%), sales (42%), and combined (18%) respondents

Marketing Budgets: Smaller companies ($5M–$15M ARR) saw dramatic budget cuts — down from $3.3M to $825K, nearly a 75% decrease

Top GTM Challenges in 2025: #1: Positioning product as a "must have" — 44%, up 6% YoY

Revenue Re-Forecasting: 66% of respondents changed their revenue plan mid-year; 43% increased it (down from 48% prior year), while 23% decreased (up from 18%)

Renewals Ownership Shift: Customer Success owned renewals 56% of the time in 2023 — now just 29% in 2025

MQL Scoring Model Collapse: Use of formal scoring models (demographic fit + engagement)

Top Marketing KPIs: #1: Dollar value of opportunities (pipeline) was number one metric at 56%

CAC & Cost-Per-Lead Awareness - The "Bonus" Topic: 45% of respondents didn't know their CAC; 41% didn't know their cost per lead

Closing Recommendations: Both hosts recommend reading the report, including pages 80–98 covering AI adoption in sales and marketing, that they were not able to cover in this episode

If you are a GTM executive leading a software company or the CFO responsible for driving revenue growth and profitability - this episode and the associated report is a great source of insights!
See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

For more than a decade, the Rule of 40 has been the gold standard for measuring SaaS performance:
Growth Rate + Profit Margin ≥ 40
But in today’s environment of higher interest rates, multiple compression, private equity leverage, and AI-driven cost pressures, that benchmark may no longer be enough.
In this episode of the Metrics Brothers, Ray “Growth” Rike and Dave “CAC” Kellogg explore why many investors and operators are now targeting something far more ambitious:
The Rule of 60.
Dave walks through the history of SaaS economics, from the growth-at-all-costs era, to the rise of balanced metrics after 2015, to the capital reset that began in 2022. The discussion then shifts to the math driving today’s expectations: if private equity firms buy companies at high multiples but must sell them later at lower multiples, Rule-of-40 performance simply doesn’t always generate acceptable returns.
In many leveraged SaaS deals today, hitting Rule of 60 can be the difference between a 1.1x return and a 3x outcome.
Ray and Dave also dig into how the Rule of 40 is evolving in practice, including:
Why growth still matters far more than margin in valuation models
How companies organically converged on the “20/20” Rule-of-40 profile
Why PE investors increasingly expect Rule-of-60 performance
The impact of debt service, CAC inflation, and AI cost structures
Why achieving Rule of 60 often requires radical operational changes

The takeaway: this isn’t a temporary metric trend.
For many SaaS companies, the math of modern software investing now demands it.
One interesting comment that reflects the reality of the Rule of 60, especially for companies that have been funded with leverage debt is: “This isn’t a fad. The math of the deal breaks without it.”
See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

In this episode, Dave "CAC" Kellogg and Ray "Growth" Rike go point-counterpoint on two high-profile articles making waves across Wall Street and Silicon Valley: Citrini's provocative February 2025 report, The 2028 Global Intelligence Crisis, and Citadel's rebuttal, The 2026 Global Intelligence Crisis.
Dave and Ray unpack whether AI is truly triggering an unprecedented economic collapse or whether Citrini's dark simulation is, as one economist put it, just "a scary bedtime story." They dig into the SaaS private credit contagion theory, the historic parallels of labor displacement, the role of government regulation, and why this particular AI scare hits closer to home than any previous tech disruption. As always, the brothers bring the receipts, including nearly 20 sources and 20 hours of research - so you don't have to.

Full Episode Summary:
Dave Kellogg and Ray Rike open by framing the episode as a tale of two AI futures: Citrini's alarming speculative simulation versus Citadel's data-driven rebuttal.
The Citrini Case (Bear Case): Published February 22nd, Citrini's report simulates a scenario in which rapid AI agent adoption triggers a global intelligence crisis by mid-2028 featuring 10.2% unemployment and a 38% drop in the S&P 500. The report argues AI is categorically different from prior technology waves because it displaces cognitive workers, who represent roughly 75% of U.S. labor income.
Citrini further warns that SaaS, already accounting for 23% - 25% of the $3 trillion U.S. private credit market could become the chip in the windshield that cracks the broader financial system, with ripple effects into insurance and the broader economy. Dave and Ray note that Citrini's word choices ran 3.4-to-1 negative, and flag that the firm may hold short positions — characterizing the piece as well-crafted "bear porn."
The Citadel Rebuttal (Bull Case): Two days later, Citadel, a $65B AUM asset manager with 35 years of credibility responded with a data-driven defense. Software engineering jobs are up since January 2024, AI CapEx is 2% of GDP and AI-adjacent commodity pricing is up 65%. Citadel argues AI follows historical S-curve adoption patterns, that "recursive capability doesn't equal recursive adoption," and that technology has always complemented rather than replaced labor - pointing to Microsoft Office as a historical analogue.
Dave and Ray's Take: Both hosts find Citadel more credible, but acknowledge real displacement risks ahead. Their key insight: the reason this particular AI scare is generating 10x more fear than past labor disruptions (auto workers, telephone operators, elevator operators) is that this time it's us — white-collar knowledge workers facing displacement. Ray adds that blue-collar jobs (truck drivers, Uber drivers, warehouse workers) face equal or greater long-term risk from AI plus robotics, but those disruptions don't generate the same visceral fear in the media and investor class. Both agree the timing of adoption is the biggest unknown. Long-term, history favors the Citadel view. Short-term, the transition could be painful.
On Government Response: Dave and Ray agree that political and regulatory intervention is inevitable if unemployment spikes materially, whether through labor protections, AI regulation, or fiscal stimulus.
On Economists' Reactions: Real economists, including Noah Smith (Noahpinion) and Wharton's Jeremy Siegel, largely dismissed the Citrini piece, wi Siegel arguing that productivity gains generate new income and demand, Smith calling it a "scary bedtime story." Dave's takeaway for operators: let the Metrics Brothers do the 20 hours of reading so you don't have to.
See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.