Part one built the machine: metric, period, slope. This half is the failure catalog, drawn from the disputes that actually happen, because earnout litigation is common enough that the patterns are well documented, and every pattern has a contract clause that would have prevented it. Read this before you sign, not after, since afterward the only people it helps are the lawyers.
Failure mode one: the integration that eats the metric
The most common story isn’t villainy. It’s a buyer running their standard playbook on a business whose earnout never occurred to them. They migrate your clients to the platform PSA in month four (a disruptive quarter, some churn, your retention metric bleeds). They fold your help desk into the regional one and service quality wobbles through the transition. They reassign your best account manager to a platform project. They reprice the acquired base per the consolidation model, and a few flagship clients walk. Each decision is defensible company strategy. Each one also moves your number, and the purchase agreement’s silence on the subject means the loss is yours.
The preventive language: operating covenants for the earnout period. Realistic ones, because a buyer won’t (and shouldn’t) freeze their company for your metric: commitments to run the acquired business consistent with past practice in defined respects, advance notice and consultation on changes that foreseeably affect the earnout metric, protection of the specific inputs your number depends on (pricing authority over the acquired base, your role’s scope, dedicated sales resources), and, at minimum, the covenant courts read into ambiguous deals anyway, stated explicitly: the buyer won’t take actions primarily intended to avoid earnout payments. Pair it with the acceleration triggers from part one (resale of the company, your termination without cause, material covenant breach), so the worst cases pay out rather than play out.
Failure mode two: definitions that drift
The second pattern is quieter. Nobody changes the business; they change the arithmetic. “EBITDA of the Business” turns out, at measurement time, to bear new corporate overhead allocations, a management fee that didn’t exist at signing, integration costs booked against your unit, revenue reclassified between entities as the buyer restructures, or the buyer’s accounting policies (different rev-rec, different capitalization thresholds) applied retroactively to your trailing baseline. Each adjustment is arguable. That’s the problem: you’re now in an argument, annually, with the side that keeps the ledger.
The preventive language is accounting precision that feels excessive at signing and priceless later: the named accounting standard as historically applied by the seller, an explicit exclusion list (allocations, affiliate fees, transaction and integration costs, non-cash purchase-accounting effects), a locked baseline exhibit showing the trailing numbers both sides verified in diligence, and a worked example with real dollars attached as a schedule. If your earnout section contains no arithmetic, it isn’t done. Then bolt on process: your audit and information rights (monthly reporting of the metric, access to supporting detail), a defined objection window, and a pre-named independent accountant as referee whose determination is final and whose fees the loser pays. That last clause converts disputes from lawsuits into a few weeks of accounting review, and its mere existence prevents most of them.
Failure mode three: the earnout that was the price gap wearing a costume
Some earnouts fail at conception. The buyer valued the business at $2.8M, the seller needed to hear $3.5M, and the extra $700k became an earnout with targets set off the seller’s own optimistic projections, which the buyer never believed and the seller now has to hit inside a company they don’t control. Missing was always the likely outcome; the earnout’s real function was letting both sides sign while disagreeing about the truth. When it misses, the seller feels robbed and the buyer feels vindicated, and both are responding to the same original sin: targets anchored to hope instead of to the trailing performance diligence verified.
The prevention here is posture, not drafting: treat any earnout whose payout requires growth above your verified run-rate as what it is, a lottery ticket stapled to a lower price, and compare offers accordingly. Targets at or near trailing performance, escalating modestly, are bridges. Targets 30% above trailing are marketing.
When to trade earnout dollars for certain ones
All of which sets up the only question that matters at the term sheet: how much earnout should you accept at all? The honest framework: an earnout is appropriate when it covers a specific, named, bounded risk the buyer genuinely can’t price (the anchor client’s renewal, a pending contract award), sized to that risk and no larger. It’s inappropriate as a general price gap-filler, as a substitute for the buyer’s conviction, or as the mechanism carrying more of your proceeds than you could afford to lose outright.
So run the trade explicitly. If the buyer offers $3.2M with an $800k earnout you now know to risk-adjust to maybe $400k, ask for $2.9M with $200k contingent, and watch what happens: a buyer who believes their own model will often move most of the way, because certainty is cheap for them to grant. A buyer who won’t move is telling you the earnout was never expected to pay, which is the most valuable disclosure in your whole process, delivered free, before you signed. Either way you’ve applied the rule that survives both parts of this series: take the smaller number you can count on over the larger one you’d have to sue for. Sellers who regret that trade are rare. We’ve never met one.