We are currently spending ~7% of revenue on dedicated GTM. That is extremely lean for a company trying to invent a category, shape a market, and move high-trust, high-friction decisions.
To be fair, last year at ~$1.5M revenue we were closer to ~17% — still lean, but not quite as much of an outlier. At ~$3.6M revenue today, ~7% starts to look unusually low for the stage/problem we are tackling.
For context, construction-tech companies like Procore and ServiceTitan spend roughly ~30–45% of revenue on Sales & Marketing alone because changing workflows in messy physical industries requires humans, trust, onboarding, enablement, and repeated customer interaction — not just software. A Series A startup at our scale (burning other people’s money) could easily spend 50–100%+ of revenue on GTM while trying to discover repeatable motion.
To be clear: I do not view this as a backward-looking critique of decisions we made. GTM has largely been in my court, and frankly the company has had many other things to figure out at the same time. But I do think TC’s broader point is fair: for a company that has been at this for a few years, we probably have fewer deployment/GTM learnings than we should — in part because we have had so little sales surface area in market.
Part of this may also be role design. In my head, my highest leverage role was never really “do every transaction sale myself.” It was:
build distribution channels (Aurora, FSE, NYCA, etc.),
shape positioning,
help define what we build,
and have complementary people/resources move projects forward day to day,
with me coming in selectively for:
trust-building,
technical credibility,
packaging,
and harder judgment calls near the end.
In practice, we have not really had enough of that middle layer to learn what parts of Marc are compressible versus non-compressible.
TC’s push was not “hire more salespeople tomorrow.” It was:
stop hand-waving and instrument the motion.
If deployment/customer success is core, then we should define explicit hypotheses, funnel stages, KPIs, and learn faster.
Where does the motion actually break?
Top of funnel?
Trust?
Packaging?
Procurement?
Sequencing?
Follow-through?
The uncomfortable realization for me: by being this lean on GTM, we may have also limited our learning surface area. We simply have not had many humans in market testing messaging, workflows, objections, sequencing, partner motions, or different ways of getting owners to “yes.”
We may have been under-learning as much as under-selling.
The more I reflect on it, the question is less:
can software replace me?
and more:
how do we systematically compress all my voice over shit?
If we can reduce the amount of scarce expert orchestration required per building while maintaining trust and outcomes, that becomes very interesting leverage.
My main takeaway / to-do before the next convo with TC (~3 weeks): move from philosophy to instrumentation. I want to define a few explicit GTM hypotheses (e.g., Daisy Chain as a wedge, direct owner outreach vs. channel motion through FSE/Aurora, and what parts of Marc are actually compressible), implement a very simple funnel with stage tracking, and intentionally increase learning surface area. Concretely, that means running more reps in market, tracking where buildings stall, being more systematic about post-mortems after owner conversations, and getting clearer on where I uniquely add value versus where repeatable workflows or other people could increasingly step in. The goal is less “close a bunch of deals immediately” and more “learn where the motion breaks and what actually scales.”
One other benchmark I found grounding: I asked Sri what Steven Winter’s revenue per employee is and he said ~$225k — almost exactly where we are today — and frankly not that far from Procore and ServiceTitan.
I found that kinda interesting.
Consulting businesses can absolutely be profitable, but they are still hard to scale because revenue tends to rise primarily with expert headcount
Technology businesses with true “lego bricks” are theoretically easier to scale because workflows, knowledge, and infrastructure become reusable — but that does not necessarily mean dramatically higher revenue per employee, especially in messy real-world industries like construction.
If Procore, ServiceTitan, and others are any barometer, success still requires a lot of people — and many of those people are going to need to be in... sales....
I am honestly really excited to put these pieces together with all of you in July. But after 4.5 years of doing this, I also feel real urgency. It kinda feels like now or never.
So… LFG.
Reference: ConTech & PropTech SaaS Platform Comparison (Ranked by Revenue per Employee)
The table below ranks companies by Revenue per Employee (RPE) from highest to lowest. A row representing a typical Series A-ready startup has been added to show how early-stage venture metrics compare to scaled market leaders.
AppFolio takes the top spot because its model captures automated transaction fees on electronic rent payments and tenant screenings, scaling revenue instantly without requiring a proportional increase in headcount.
Scaled, pure-play vertical SaaS players that require significant onboarding support — like Procore, ServiceTitan, Bentley, and Buildertrend — naturally cluster together in the tight $240k–$300k revenue per employee range.
The Series A Target (When Burning Other People’s Money...) #
A healthy Series A startup often aims for $100k–$150k revenue per employee. This marks a critical transition point, signaling to institutional investors that the company is stepping out of the early-stage building phase and beginning to capture true software operating leverage.
Strategy Notebook
A useful benchmark from today’s convo with TC:
We are currently spending ~7% of revenue on dedicated GTM. That is extremely lean for a company trying to invent a category, shape a market, and move high-trust, high-friction decisions.
To be fair, last year at ~$1.5M revenue we were closer to ~17% — still lean, but not quite as much of an outlier. At ~$3.6M revenue today, ~7% starts to look unusually low for the stage/problem we are tackling.
For context, construction-tech companies like Procore and ServiceTitan spend roughly ~30–45% of revenue on Sales & Marketing alone because changing workflows in messy physical industries requires humans, trust, onboarding, enablement, and repeated customer interaction — not just software. A Series A startup at our scale (burning other people’s money) could easily spend 50–100%+ of revenue on GTM while trying to discover repeatable motion.
To be clear: I do not view this as a backward-looking critique of decisions we made. GTM has largely been in my court, and frankly the company has had many other things to figure out at the same time. But I do think TC’s broader point is fair: for a company that has been at this for a few years, we probably have fewer deployment/GTM learnings than we should — in part because we have had so little sales surface area in market.
Part of this may also be role design. In my head, my highest leverage role was never really “do every transaction sale myself.” It was:
with me coming in selectively for:
In practice, we have not really had enough of that middle layer to learn what parts of Marc are compressible versus non-compressible.
TC’s push was not “hire more salespeople tomorrow.” It was:
If deployment/customer success is core, then we should define explicit hypotheses, funnel stages, KPIs, and learn faster.
Where does the motion actually break?
The uncomfortable realization for me: by being this lean on GTM, we may have also limited our learning surface area. We simply have not had many humans in market testing messaging, workflows, objections, sequencing, partner motions, or different ways of getting owners to “yes.”
We may have been under-learning as much as under-selling.
The more I reflect on it, the question is less:
and more:
If we can reduce the amount of scarce expert orchestration required per building while maintaining trust and outcomes, that becomes very interesting leverage.
My main takeaway / to-do before the next convo with TC (~3 weeks): move from philosophy to instrumentation. I want to define a few explicit GTM hypotheses (e.g., Daisy Chain as a wedge, direct owner outreach vs. channel motion through FSE/Aurora, and what parts of Marc are actually compressible), implement a very simple funnel with stage tracking, and intentionally increase learning surface area. Concretely, that means running more reps in market, tracking where buildings stall, being more systematic about post-mortems after owner conversations, and getting clearer on where I uniquely add value versus where repeatable workflows or other people could increasingly step in. The goal is less “close a bunch of deals immediately” and more “learn where the motion breaks and what actually scales.”
One other benchmark I found grounding: I asked Sri what Steven Winter’s revenue per employee is and he said ~$225k — almost exactly where we are today — and frankly not that far from Procore and ServiceTitan.
I found that kinda interesting.
Consulting businesses can absolutely be profitable, but they are still hard to scale because revenue tends to rise primarily with expert headcount
Technology businesses with true “lego bricks” are theoretically easier to scale because workflows, knowledge, and infrastructure become reusable — but that does not necessarily mean dramatically higher revenue per employee, especially in messy real-world industries like construction.
If Procore, ServiceTitan, and others are any barometer, success still requires a lot of people — and many of those people are going to need to be in... sales....
I am honestly really excited to put these pieces together with all of you in July. But after 4.5 years of doing this, I also feel real urgency. It kinda feels like now or never.
So… LFG.
Reference: ConTech & PropTech SaaS Platform Comparison (Ranked by Revenue per Employee)
The table below ranks companies by Revenue per Employee (RPE) from highest to lowest. A row representing a typical Series A-ready startup has been added to show how early-stage venture metrics compare to scaled market leaders.
Key Observations from the Rankings #
The Transaction Premium #
AppFolio takes the top spot because its model captures automated transaction fees on electronic rent payments and tenant screenings, scaling revenue instantly without requiring a proportional increase in headcount.
The Industry “Sweet Spot” #
Scaled, pure-play vertical SaaS players that require significant onboarding support — like Procore, ServiceTitan, Bentley, and Buildertrend — naturally cluster together in the tight $240k–$300k revenue per employee range.
The Series A Target (When Burning Other People’s Money...) #
A healthy Series A startup often aims for $100k–$150k revenue per employee. This marks a critical transition point, signaling to institutional investors that the company is stepping out of the early-stage building phase and beginning to capture true software operating leverage.