The Conversation Is Not the Structure

The most widely circulated partnership advice published this week tells founders they should have seven critical conversations before signing anything, and the premise is correct: if you never discuss roles, equity expectations, exit scenarios, and decision rights, you will fight about them later.

What the advice cannot account for is that founders who do have every conversation also fight about them later.

The framework, published in Inc. on May 31 by experienced operators who have run a 25-year partnership, is genuine and hard-earned. The seven questions are sensible. The problem is structural, and it has nothing to do with the quality of the questions. It has to do with what founders believe they have completed when they can answer them.

When Agreement Becomes Its Own Risk

A conversation held at the beginning of a partnership takes place in a specific set of conditions: peak optimism, zero pressure, shared excitement, and no decisions that have cost either party anything yet. Both founders answer honestly. Both founders mean what they say. In that moment, the answers are probably true.

The partnership structure will eventually be tested in the opposite conditions: low revenue, competing priorities, a deal that benefits one party more than the other, a slow quarter that neither person will name out loud as evidence that the original thesis was wrong. In those conditions, the answers from the opening conversation are not irrelevant, but they are not load-bearing either, because they were never tested under the weight they are now being asked to carry.

Founders who complete the seven conversations feel resolved. That feeling is the hazard. They leave the room believing they have done structural work because the conversation required courage and specificity and produced agreement. What they have actually confirmed is that both parties can reach alignment in a low-stakes room when no real decision is required. That is a different capability than the one a partnership needs when month nine arrives and one partner has stopped generating the revenue the original thesis assumed.

The clarity produced by a good pre-partnership conversation is real, but it is also temporary, and the period it covers is precisely the period during which the partnership was not going to fail anyway. The first three months of a partnership almost always feel functional. The conversation did not create that; the honeymoon conditions did. What the conversation cannot do is extend those conditions into month eleven, and founders who believe it can are solving for a problem that was never their actual problem.

Conversation as Infrastructure

The M&A data illuminates this pattern at scale. Between 70 and 90 percent of mergers and acquisitions fail to create shareholder value, and the most instructive category of failure is not the deals that collapsed before the letter of intent was signed. It is the deals that cleared legal, commercial, and financial due diligence and still failed, because behavior under pressure diverged from the agreement made at the table. The parties talked. The parties agreed. The parties documented the agreement. Then the conditions changed, and the agreement turned out to have been a record of intent rather than a mechanism of governance.

Founders building partnerships at the operator level face the same gap at a smaller scale. A conversation produces a shared record of intent. Structure produces a mechanism that governs behavior when intent is no longer aligned. The two things are not the same, and conflating them is what makes the conversation feel like sufficient preparation.

The distinction matters precisely because conversations are comfortable to have and structure is uncomfortable to build. Talking about what happens if the partnership ends is easier than building a legal mechanism that governs the dissolution. Agreeing that "we'll handle conflicts by talking through them" is easier than establishing who holds the deciding vote when talking through them produces no resolution. Founders choose conversation because it is available, because it signals good faith, and because it produces the feeling of clarity without requiring either party to encode the actual work into a system.

This is not a failure of intention. Founders who skip the structure are rarely cynical about it. They genuinely believe that the agreement they reached in the opening conversation is a durable one, because in the moment they reached it, it was. The problem is that durability under good conditions and durability under bad conditions are different properties, and a conversation can only test for the first one.

What Structure Actually Requires

Structure requires that the agreement continue to function when both parties stop wanting it to. That is a different design challenge than producing mutual agreement in a single session, and it requires different inputs: documented decision rights, a defined review cadence, an escalation path that does not depend on both parties being cooperative at the same time, and specific criteria that trigger re-evaluation rather than a vague commitment to "revisit as needed."

Founders who walk away from the seven-conversation framework with a decision log, an escalation protocol, and a 90-day review milestone have used the conversation as a starting point for structure. Founders who walk away with a handshake and shared relief have done the version of the work that performs well under no conditions except the ones that already exist, which are the easiest possible conditions for a partnership to survive.

The founders who build durable partnerships do not stop at the conversation. They translate the answers into mechanisms, review those mechanisms against actual behavior within the first quarter, and accept that the first version of the agreement will require revision because no conversation held before the first real test can anticipate what that test will cost. onSpark sees this pattern consistently across its network of operators: the partnerships that dissolved fastest were rarely the ones that skipped the opening conversation. They were the ones that mistook it for structural preparation and built nothing afterward.

The seven conversations are better than nothing. They are also not close to enough. A founder who leaves that meeting feeling resolved has completed the easiest possible version of the work, in the most cooperative possible version of the relationship, under conditions that will never exist again in exactly that form. The conversation is the beginning of the structural work. The most dangerous thing a founder can do with it is treat it as the end.