TL;DR: The slowdown you’re feeling between Series A and Series C is structural, not motivational. Around 60 people, the configuration that won you Series A — flat decision-making, a founder in every loop, scope defined by who shows up — actively starts breaking the company. The fix isn’t hiring a COO or telling people to “step up.” It’s redesigning decision rights, reporting depth, and founder scope before the symptoms harden into culture.
The diagnostic question you’re already asking
You’ve noticed it. Decisions that took a day now take three weeks. People you trust are escalating things they used to own. Your calendar is the bottleneck for the company, and the same problem keeps surfacing in different costumes. You suspect this is a people problem. It almost never is, at this stage.
The transition from Series A to Series C is the most predictable structural failure window in a startup’s life. It’s not where companies fail — it’s where they slow. And the slowdown is rarely diagnosed correctly, because the symptoms feel like personnel issues while the cause is the org chart you stopped looking at six months ago.
The transition window — what changes at 60 to 150 people
The A-to-C window is defined by a specific structural shift: the company crosses the threshold where every person can no longer hold every context. Research on organizational scale, including Robin Dunbar’s well-known work on group cognition, places that threshold somewhere between 50 and 150 people, depending on density of interaction (Dunbar, 1992). AI-native companies tend to hit it closer to 60 because product surface area expands faster than headcount.
Before this threshold, informal coordination is faster than any process you could design. After it, informal coordination is the bottleneck. The team that got you to Series A was optimized for speed under shared context. The team you need at Series C is optimized for speed under distributed context. Those are not the same machine, and you cannot get from one to the other by hiring harder.
The structural failure modes below all share the same root: the operating system you ran at 30 people is still running at 90, producing the symptoms you’re trying to fix at the layer of individuals. Here is the map — five failure modes, what each looks like from the inside, and the structural intervention for each. The rest of the article walks through them in order.
# | Failure mode | What you see | Structural intervention |
|---|---|---|---|
1 | Decision rights vacuum | Senior hires asking permission; cross-functional loops | Written decision rights per function (owner / consulted / informed) |
2 | Flat-org tax | Founder in three stand-ups; 12 direct reports | Cap founder span at 5–7; add layers with explicit scope |
3 | Founder scope creep, in reverse | Calendar full of everything; "needed everywhere, excellent nowhere" | Contract scope to 3 things only the founder can do |
4 | Trust-context mismatch | Early hire has more authority than the VP who owns the function | Publish decision rights for every senior role in week one |
5 | Process built for the last failure | Three-week planning, eight-stage hiring loop | Audit each process against current top-five risks |
Failure mode 1 — The decision rights vacuum
The structural failure mode: Decisions that should be owned by a function are still being made by the founder, or, worse, by no one in particular.
You see it as: senior hires asking permission for things they were hired to own. Cross-functional decisions that loop indefinitely because no one role has the authority to close them. A growing list of “we need to talk about this” items on your 1:1s.
The root cause is not that the people are weak. It’s that decision rights were never written down because they didn’t need to be when everyone sat in one room. As the company scales, the absence of explicit rights defaults to founder rights — every ambiguous call routes upward. By Series B, the founder is the rate-limiter on the entire company, and the strongest hires either start working around the founder (creating shadow authority) or quietly check out.
The structural intervention: Define decision rights at the function level, not the title level. For each major decision class — pricing, hiring above a threshold, infrastructure spend, customer commitments — name the single role that owns the call, the roles that must be consulted, and the roles that get informed. The form matters less than the fact that it exists in writing and is referenced when decisions are made. The McKinsey work on decision-making at scale puts this bluntly: organizations that explicitly assign decision rights move roughly twice as fast on cross-functional calls as those that don’t (McKinsey, “Decision making in the age of urgency”).
Failure mode 2 — The flat-org tax
The structural failure mode: The company is still operating with one or two layers between the founder and the IC, well past the point where that span is feasible.
You see it as: the founder in stand-ups for three different teams. Senior leaders managing twelve direct reports because adding a manager felt “too corporate.” 1:1s compressed to fifteen minutes because the math doesn’t work. Decisions delayed not by debate but by calendars.
The root cause is a misread of what flatness costs. Flatness is fast at small scale because the founder can hold every context. At 80 people, flatness is no longer the absence of bureaucracy — it’s the presence of a single overloaded human. Adding a layer is not the same as adding bureaucracy. Bureaucracy is process applied to decisions that don’t need it. A layer is a person whose full-time job is to hold context, set direction, and remove blockers for a defined surface area. Companies confuse the two and refuse the layer, then accumulate the bureaucracy anyway as a coping mechanism.
The structural intervention: Reduce founder span to no more than five to seven direct reports, each of whom owns a clearly bounded function. The added layer is not a “VP of Stuff.” It’s a person with explicit scope and explicit decision rights. If you can’t name what they own without using the word “and” three times, the scope isn’t ready and neither is the hire.
Failure mode 3 — Founder scope creep, in reverse
The structural failure mode: The founder’s scope has expanded with the company instead of contracting.
You see it as: a calendar that contains product reviews, sales calls, hiring loops, infrastructure debates, and investor updates — all in the same week. A vague but persistent feeling that you are needed everywhere and excellent nowhere. The strongest signal: your top hires are starting to do the same thing, because they’re modeling you.
The root cause is the most common founder error in this window, and it’s the inverse of what most advice tells you. The advice says “delegate more.” The structural reality is that founder scope must contract, not just delegate. At 30 people, the founder’s job is everything that isn’t done yet. At 100 people, the founder’s job is the two or three things only they can do — usually some combination of vision, capital, the most senior hires, and the highest-stakes customer relationships. Everything else must be owned, not delegated. Delegated work routes back. Owned work doesn’t.
The structural intervention: Write down the three things only you can do this year. Audit your last two weeks of calendar against that list. Anything that doesn’t map is either a transfer of ownership (not a delegation) or a hard stop. This is not “letting go.” Letting go implies the work was yours to release. Contracting scope means recognizing the work was never supposed to be yours at this stage, and the company has been compensating for an undefined founder role.
Failure mode 4 — The trust-context mismatch
The structural failure mode: Trust in the company is still being granted based on tenure and proximity to the founder, not based on role and decision rights.
You see it as: the early hire who joined at twelve people having more informal authority than the VP you hired at sixty, even though the VP owns the function. New hires sensing — correctly — that the real org chart is different from the published one. Quiet attrition among senior hires who can’t tell whether they have authority or are auditioning for it.
The root cause is that trust at small scale is personal — you trust people because you’ve watched them operate. Trust at scale must be structural — you trust the role, and the role is held to its decision rights. When trust stays personal past 60 people, every senior hire enters a company where the rules are unwritten and the existing team is the unwritten rulebook. The strong ones leave. The accommodating ones stay and underperform. Either way, the founder concludes the senior hires “weren’t the right fit,” and the cycle continues.
The structural intervention: Make trust portable by making roles explicit. Three concrete moves: (1) publish a one-page decision-rights map for every senior hire within their first week — what they own, what they consult on, what routes through them. (2) For each early-team member with informal authority, formalize the scope into a role title with explicit rights, or retire the informal authority into a peer relationship — never both. (3) When a senior hire and an early team member disagree, escalation routes through the rights map, not through the founder’s relationship history. There is no third option that scales.
Failure mode 5 — Process built for the last failure, not the next one
The structural failure mode: The company has accumulated process in response to specific past incidents, none of which is the failure mode it’s currently facing.
You see it as: a planning ritual that takes three weeks because of one bad quarter eighteen months ago. An approval chain for spend that exists because of one rogue purchase. A hiring loop that has eight stages because of one bad hire. Every individual process is defensible. In aggregate, they are the bureaucracy that the flat-org refusal was trying to avoid.
The root cause is that process is almost always written reactively and almost never retired. Each layer of process was rational when it was added. None of it is reviewed against the current failure mode of the company. By Series B, the company is running on a stack of fixes for problems it no longer has, while under-resourced for the problems it does.
The structural intervention: Audit process by failure mode, not by department. For each significant process, ask: what specific failure is this preventing, and is that failure still the highest-cost failure at our current stage? If the answer is no, the process is overhead. The hardest part of this is political, not analytical — every process has someone who remembers why it exists. The intervention is to retire process explicitly and publicly, not to let it decay.
Diagnostic mistakes founders make in this window
Even founders who correctly identify this as structural still get the intervention wrong in predictable ways.
Confusing structure with personnel. “We need to upgrade the team” is the most common response to a structural failure, and the most expensive one. Replacing a person inside a broken structure produces the same outcome on a six-month delay. Diagnose structure first; personnel decisions become clearer once the structure is right.
Hiring a COO as the answer. A COO is a structural intervention only if the founder has already done the work of contracting scope and defining decision rights. A COO hired into an undefined operating environment becomes the most expensive symptom of the original problem. The structure has to be designed before the role is filled, not after.
Treating layers as failure. Adding a layer of management feels like a loss of velocity. At this stage, refusing to add a layer is the loss of velocity — it just shows up two quarters later as turnover and missed plans.
Running the diagnosis once. The A-to-C window is not a single transition. It’s a moving threshold the company crosses at roughly 60, 100, and 150 people, and each crossing changes the math on decision rights, span, and founder scope. The diagnosis has to be re-run, not solved.
Mistaking culture for the cause. When the symptoms surface — slowness, escalation, attrition — they look cultural. They are downstream of structure. Investing in culture before fixing structure is laying carpet over a cracked foundation.
The self-diagnostic — run this in the next two weeks
Print this. Fill it in honestly. The diagnostic works only if your answers are specific enough to test. Example entries are filled in to show what useful answers look like.
-
Decision rights audit List the five most consequential decisions made in the last 90 days. For each: who owned it (one name), who was consulted, who was informed.
Example: Pricing change for enterprise tier — Owner: VP Product. Consulted: CRO, Founder. Informed: Sales, CS. Took 9 days from proposal to call. Flag: any decision where the owner is “the founder and ___” or “we discussed it as a team.” -
Founder span check List your direct reports.
Example: 9 directs — VP Eng, VP Product, Head of Sales, Head of CS, Head of Finance, Head of Recruiting, Head of Ops, Lead Designer, Chief of Staff. Flag: more than seven names. Flag: any name without a one-line scope you could write without using “and.” -
Founder scope contract Write the three things only you can do this year.
Example: 1) Lead Series C raise. 2) Own top-ten enterprise customer relationships. 3) Hire CTO and CRO. Audit last 14 days of calendar against the list. Anything below 60% mapped is a scope problem, not a time-management problem. -
Trust map List the five people in the company with the most informal authority — the people whose opinion changes outcomes.
Example: Founding engineer (joined at 8 people, no direct reports). VP Eng (hired at 70). Founder. Head of Sales. Founding designer. Flag: any name without a corresponding formal role and decision rights. That’s an unwritten org chart. -
Process audit List your three heaviest processes — the ones that consume the most calendar time across the company.
Example: Quarterly planning (3 weeks). Hiring loop (6 weeks for senior roles). Pricing approval (avg 11 days). For each, name the specific failure it prevents. If you can’t name it, or the failure is no longer in your top five risks, the process is overhead. -
The structural-vs-personnel test For each underperforming senior hire, ask: would a high performer in their role succeed inside the current decision rights, span, and process? If the honest answer is no, the problem is structural. Fix structure before changing personnel.
If three or more of these flags fire, you are in the A-to-C structural failure window. The good news is that the interventions are concrete and the timeline is short — most of them can be in motion within a quarter. The cost of waiting is not the next quarter; it’s the senior hires you’re about to lose because the structure can’t hold them.
Closing synthesis
The slowdown between Series A and Series C is the most predictable thing that will happen to your company, and the most miscoded. It is the signal that the operating system you built for thirty people is now running ninety, producing exactly the failure modes that operating system was always going to produce at scale. Decision rights, reporting depth, and founder scope are the three levers. All structural. None solved by hiring harder, working longer, or letting go more. The founders who navigate this window cleanly diagnose it as structure — and intervene before the symptoms become culture.
What changes in 90 days vs. 12 months: The decision-rights map, founder scope contract, and span audit can land inside a quarter. Re-running the diagnosis at the next headcount inflection, and retiring process that no longer matches the current top-five risks, is a 12-month discipline. The first set is a project; the second is a habit.
If you’re in this window now and want a second pair of eyes on the diagnosis, the door is open.
FAQ
Q: How do I tell if my scaling problem is people or structure? A: Ask whether a high performer in the same role would succeed inside the current decision rights, span, and process. If the honest answer is no, the problem is structural. Replacing the person without fixing the structure reproduces the same outcome on a delay.
Q: When should a founder change decision rights — at Series A, B, or C? A: The forcing function is headcount, not stage. Around 60 people, informal decision-making stops being faster than explicit decision rights. Most companies hit that point in the late-A or early-B window. The cost of acting late compounds quickly; the cost of acting early is negligible.
Q: What is the right founder scope at Series B? A: Three things only you can do — typically some combination of vision, capital, the most senior hires, and the highest-stakes customer relationships. Everything else should be owned by a named role with explicit decision rights, not delegated. Delegated work routes back. Owned work doesn’t.
Q: Isn’t adding management layers just adding bureaucracy? A: No. Bureaucracy is process applied to decisions that don’t need it. A layer is a person whose full-time job is to hold context and set direction for a defined surface area. Companies that refuse the layer accumulate the bureaucracy anyway, as a coping mechanism for an overloaded founder.
Q: Should I just hire a COO to fix this? A: Only after you’ve defined decision rights, contracted founder scope, and named the operating environment the COO will work inside. A COO hired before that work is done becomes the most expensive symptom of the original structural problem.
Q: How often should I re-run the structural diagnosis? A: At each headcount inflection — roughly 60, 100, and 150 people. Each crossing changes the math on span, decision rights, and founder scope. The diagnosis is not a one-time fix; it’s a recurring discipline through the entire A-to-C window.
Part of The Four Scaling Decisions That Compound — and How to Tell You’re About to Get One Wrong — the operator’s-seat overview of the four scaling decisions that compound between Series A and Series C.
I unpack observations like this every week in Operator’s Log. Grounded AI perspective from the operator’s seat. Free. Subscribe →