The Exit Interview Nobody Conducts

The founding team gets assessed at the start and taken for granted ever after. Then the project runs long — as they always do — and one by one people's lives move in directions the original agreement never accounted for. What nobody planned for, and nobody measured, is the question that should have been asked long before the exit: when exactly did this become inevitable?

The Assumption

The team that validates the idea is the team that will commercialize it.

This assumption is almost never stated. It does not need to be — it is embedded in the structure of how founding teams are formed, how equity is split, and how investor narratives are constructed. The team slide in the pitch deck presents a unit. The founding agreement distributes shares across that unit. The early milestones are achieved by that unit.

What the assumption does not account for is time.

The Reality Check

Deep tech ventures, hardware startups, and innovation-heavy B2B plays do not take three years. They take seven. Sometimes ten. The academic literature on deep tech commercialization is consistent on this point: TRL progression from lab to market routinely runs 8–12 years for complex hardware systems (Auerswald & Branscomb, Research Policy, 2003; Rothaermel & Thursby, Research Policy, 2007). Even software-heavy ventures with hardware components routinely exceed initial timeline projections by a factor of two or three.

A founding team assembled in 2018 is being asked — implicitly, without anyone revisiting the agreement — to still be the right team in 2026. That is a long time to hold.

Three things happen over that span, consistently, and consistently underpriced:

Life intervenes. Co-founders in their early forties are typically in the most compressed decade of their adult lives — aging parents, school-age children, partners who have been patient for years, a mortgage that made sense on an exit timeline that has since moved twice. These are not weaknesses; they are the normal arithmetic of a life at 43 or 47. But venture structure was not designed around them. Equity vesting schedules were not built to accommodate a co-founder who needs to step back for eighteen months to manage a family health crisis. The founding agreement assumed continuity. Life does not.

The required skillset rotates. The person who is exceptional at zero-to-one — building in ambiguity, tolerating chaos, staying motivated without external validation — is often not the person who is exceptional at one-to-ten. Scaling requires different instincts: process discipline, hiring judgment, managing downward, navigating institutional relationships. A technical co-founder who built the prototype cannot always build the organization, and a commercial co-founder who opened the first ten accounts cannot always build the sales function. This is not failure — it is a normal mismatch between human wiring and stage-dependent demand. But because nobody mapped the required skillset evolution at founding, because that conversation felt premature or unkind, the mismatch accumulates quietly until it becomes undeniable.

The financial load becomes unevenly distributed. Founders in the 40–55 bracket are rarely in the same financial position as they were at founding. One co-founder has family obligations that have grown. Another has a secondary income stream that changed the calculus. A third was counting on a bridge round that did not close on schedule. The shared sacrifice that felt equitable at year one looks different at year five — and the founding agreement, which codified the original equity distribution, does not automatically rebalance as circumstances diverge. Resentment accumulates. Commitment fragments. The person who appears least engaged is often the one whose financial exposure has grown most unsustainably.

None of these three dynamics announces itself. Each one accumulates slowly, across dozens of small decisions and undiscussed accommodations, until the gap between the team on the pitch deck and the team in the room is too large to ignore.

The Forensic Pattern

I have worked inside founding teams across multiple continents and two decades. In several cases, I joined them — carrying the dual role of advisor and operational co-founder that the venture needed at a specific stage. What follows is not a single case but a pattern, assembled from direct observation.

The founding agreement was written for the idea, not for the journey.

Equity splits negotiated at formation reflect founding contributions, risk tolerance, and relationship trust at a specific moment. They do not reflect what will be required of each person at year four or year seven. They do not contain a mechanism for reassessing whether the current team composition matches the current challenge. And because revisiting equity feels like an accusation — I am suggesting you are worth less than we agreed — the conversation that should happen repeatedly never happens at all.

The exit, when it comes, is almost never the first signal.

By the time a co-founder formally steps back, the signals have been present for twelve to eighteen months. Slower response times. Reduced initiative. Conversations that circle without landing. Commitments that slip without explanation. None of these are necessarily bad faith. Most of them are a person managing an impossible tension between what the venture needs and what their life now requires. But because the founding team has no structured mechanism for surfacing these tensions — no regular review of team fit, no agreed framework for role evolution — the signals accumulate unaddressed until exit is the only remaining option.

The venture loses more than a person.

When a founding co-founder exits, particularly past year three, the disruption is rarely limited to the operational gap they leave. Investor confidence recalibrates. The team’s narrative — the cohesion story that underpins the pitch — must be rewritten. Institutional relationships that were managed through personal trust must be re-established under new ownership. The knowledge that lived in one person’s head, never fully documented because there was always something more urgent, is now partially inaccessible.

The exit interview — the structured post-mortem on what the team assumed, what the reality turned out to be, and what should have been built differently into the founding structure — almost never happens. The departure is managed however it can be, the gap absorbed or not, and the venture continues carrying the same unexamined assumption into its next hiring decision, its next leadership structure, its next stage.

That is the immediate problem: not the loss of a person, but the absence of any mechanism to understand what the loss revealed — and what it now requires of the team that remains.

Open Questions

These are the questions the exit interview would ask — if anyone conducted one.

On the founding structure: — At what point did the equity distribution reflect something other than the current contribution? When did that gap first become visible — and to whom? — Was there a mechanism in the founding agreement for revisiting role definitions as the venture evolved? If not, when did the absence of that mechanism first cause a problem? — Which assumptions about each co-founder’s financial capacity to sustain the venture were never made explicit — and when did those assumptions first stop being true?

On the life-stage dynamics: — What external obligations were present at founding that were treated as manageable? Which ones grew beyond what was originally anticipated? — If each co-founder had been asked, privately, at the eighteen-month mark: “Are you still the right person for the role you have in year five?” — what would the honest answer have been?

On the skillset evolution: — What did the venture require at founding that it no longer requires? What does it require now that was not required then? — Which co-founders’ skills compounded with the venture’s needs — and which diverged? When was that divergence first measurable? — Who in the founding team is currently operating in a role that someone else would perform better — and what is preventing that conversation?

On the counterfactual: — If the founding agreement had included a formal annual team-fit review — not a performance review, but a structured assessment of whether each person’s role still matched the venture’s stage — would the exit have been earlier, smoother, or avoided entirely? — Who knew, and when, that the departure was becoming likely? What stopped them from raising it?

What This Means for How You Build

No founding agreement survives contact with a seven-year timeline intact. The question is not whether team composition will need to evolve — it will — but whether the founding structure contains a mechanism to manage that evolution, or whether it leaves the problem to accumulate until exit is the only option.

Three structural additions that almost no founding team builds in, and that consistently matter:

A role definition review, annually. Not a performance review — a forward-looking question: does each person’s current role match what the venture needs for the next eighteen months? This is a different question from “are you doing your job.” It is the question that makes visible the mismatch before it becomes irreparable.

An explicit financial capacity check, built into the governance rhythm. Not an interrogation — a structured acknowledgment that financial positions change, and that founding agreements can be adjusted when they stop reflecting reality. The alternative is watching resentment accumulate in silence until departure is the only release valve.

A transition protocol, agreed at founding, before anyone needs it. What happens when a co-founder needs to step back? What does partial exit look like? What is the vesting acceleration clause? What is the knowledge transfer process? These conversations are easy to have before they are necessary and nearly impossible to have when they are urgent.

The founding team is not a static asset. It is the most assumption-laden element of any venture — and the one that receives the least structured scrutiny over time.

The exit interview nobody conducts is the autopsy that could have been a design choice.

Destruction Desk
We perform autopsies on innovation’s failed assumptions.

This newsletter was edited by Manfred Lueth.