The Agreement Is Not the Partnership

The most widely read piece of legal advisory published this month contains, buried in its third paragraph, a sentence that should stop every founder cold: earn-outs in founder-led acquisitions are "inherently litigation-prone." Skadden, one of the most respected deal law firms in the world, spent several thousand words advising buyers on how to structure hybrid investment-plus-commercial-partnership arrangements, persona rights agreements, and equity rollovers, all of which are designed, at their core, to manage the risk that the founder and the buyer will eventually disagree about what they agreed to. The paper is framed as a guide to deal terms. What it actually describes is the architecture of a relationship that was never quite built.

This is not an acquisition-specific problem. Founders who have watched a partnership dissolve under the weight of documentation they both signed will recognize the pattern immediately. The agreement looked complete. The incentives looked aligned. The terms looked fair. And somewhere between month six and month eighteen, both parties discovered that the document they had executed with such care was not a record of what they had built together, but a record of what they had each hoped the other had agreed to, and the gap between those two things is where every arbitration clause eventually earns its keep.

What the Structure Is Actually Telling You

PwC's mid-year M&A outlook for 2026 notes that capital is increasingly flowing through "minority investments, joint ventures, partnerships, and other structures rather than control-based acquisitions." The implication is that the market has matured past the need for full ownership, that sophisticated parties now prefer shared upside structures because they are more flexible and more aligned. What the analysis does not say, but what the Skadden paper makes clear by implication, is that these structures proliferate precisely because alignment is difficult to establish and even harder to maintain. They are not evidence that the parties are aligned. They are evidence that the parties have agreed on a mechanism for managing the consequences of not being aligned, and called it a deal.

The earn-out is the most honest document in any founder-led transaction. Its existence is an admission that both parties believe the value of the business depends on behaviors that one party controls and the other cannot reliably verify. Logan Paul's Prime Energy produced £112 million in U.K. sales in 2023 and declined 70% in 2024, which is not a failure of the deal structure, it is a failure of the relational assumptions that were priced into the deal before it was done. The earn-out did not cause that decline. The misunderstanding about what created the value in the first place did, and that misunderstanding existed long before any lawyer entered the room.

Founders who are building strategic partnerships rather than selling companies are working through a version of this same architecture. The partnership agreement specifies the revenue share, the deliverable cadence, the exclusivity windows, and the exit conditions. What it does not specify, because no agreement can, is how the two parties will behave during the months when performance falls short of the projections that made the deal feel worth signing. The agreement captures the parties at their most optimistic. It is tested at their most pressured. Those are different people, and most agreements are not built to accommodate that difference.

The Moment the Agreement Becomes the Partnership

The specific pattern that produces most partnership failures is not conflict. Conflict is late. What produces most partnership failures is the moment, usually in month three or four, when one party experiences a gap between what they expected and what is happening, and decides to carry that concern privately rather than surface it directly. The decision feels reasonable at the time. The relationship is new, the goodwill is real, and raising a concern before there is documented evidence of a problem can feel like introducing friction into something that is still gaining momentum. The concern gets filed, the pattern continues, and by month eight both parties have an internal account of the partnership that they have never compared with each other.

By the time the formal conversation happens, both parties arrive with evidence. The founder cites missed timelines, shifted priorities, a partner whose engagement looked different in year two than in the meeting that produced the agreement. The partner cites unclear expectations, moving targets, a founder who could not distinguish between a disagreement and a betrayal. Both accounts are accurate. Both are also incomplete. And the agreement that was supposed to resolve this moment cannot, because it was written to describe what each party wanted at the beginning, not what the relationship had actually become in the middle.

The Skadden paper recommends "careful drafting around performance metrics and operating covenants" as a partial solution. That is sound legal advice, and it will not solve the underlying problem. The drafting cannot create the trust it describes. The operating covenant cannot substitute for the conversation neither party had in month four. What careful drafting produces, at its best, is a clearer record of what both parties owed each other, which makes the eventual dispute marginally more legible to an arbitrator and considerably more expensive for both founders involved.

The Partnership That Earns the Agreement

The partnerships that hold are not the ones with the most precise documentation. They are the ones in which both parties established, before the paperwork was completed, a pattern of surfacing discomfort directly and without consequence to the relationship's standing. That pattern is built in the early months through small tests, concerns raised with a low cost of acknowledgment, expectations examined before they become grievances. The founders who built this kind of infrastructure before signing typically find, when they do sign, that the agreement is almost anticlimactic, a formality that describes a relationship that already exists rather than one they are hoping to produce.

This is what makes the proliferation of earn-out structures in the 2026 deal market a useful diagnostic for founders who are not selling their companies. If the party across the table is proposing a structure designed to manage the risk that the relationship will fail after you sign, and you have not yet had a conversation that would tell you whether it will, you have not done the foundational work. The structure is not the solution. The structure is the signal that the solution is still outstanding. Platforms like onSpark are built around exactly this gap, matching founders with partners whose track record in the early months of a relationship has already demonstrated the behavioral pattern, before either party picks up a pen.

The agreement is not the partnership. The partnership is the set of behaviors both parties have already demonstrated before the agreement is signed, and whether that infrastructure exists is knowable before the term sheet, if both founders are willing to look for it.