Selling Your Startup: LOIs, Earnouts, Escrow and Reps & Warranties

A founder announces a $50 million exit. Two years later, over a drink, the real number emerges: $31 million at closing after escrow and adjustments, $9 million of earnout that never paid because the buyer reorganised the sales team, and a retention pool that came out of the purchase price rather than the buyer's pocket.

Nothing improper happened. Every one of those outcomes was written into documents the founder signed, mostly after exclusivity had been granted and leverage had evaporated.

This guide covers the sequence of an acquisition, why the letter of intent is the most important document in it, how the deal-value components actually work, what to do in the year before a process, and where sellers routinely lose money they did not know was at risk.

1. The Sequence, and Where Leverage Sits

1. Approach or process. Either an unsolicited approach, or a structured process run by a banker. The difference is enormous: a process creates competition, and competition is the only real source of seller leverage.
2. Indications of interest. Non-binding valuation ranges.
3. Letter of intent. Headline terms, and almost always exclusivity.
4. Confirmatory diligence. Six to twelve weeks of legal, financial, technical, tax and commercial review.
5. Definitive agreement. The purchase agreement, disclosure schedules and ancillary documents.
6. Signing, then closing — sometimes simultaneous, sometimes separated by regulatory approvals.
7. Post-closing: purchase price adjustment, escrow release, earnout measurement.

Your leverage peaks the instant before you sign the LOI and collapses the instant you grant exclusivity. Before exclusivity you have alternatives. After it, you have one buyer, a burning clock, mounting legal bills, and a team who knows something is happening. Every re-trade in the history of M&A has happened after exclusivity.

The practical conclusion: negotiate everything you care about into the LOI, even though it is mostly non-binding. It is far easier to hold a buyer to a written LOI position than to introduce a new demand in week nine of diligence.

2. What Belongs in the LOI

• Purchase price and form of consideration — cash, buyer stock, or a mix, and if stock, whether it is public, liquid, or subject to lock-up
• Cash-free debt-free basis and how the working capital target will be set
• Escrow size and duration, and whether it is the sole remedy for breaches
• Earnout structure, if any — metric, period, and the operating covenants protecting it
• Retention and equity pools for employees, and critically who funds them — buyer or seller
• Exclusivity period — 30 to 60 days, with a hard end date and no automatic extension
• Treatment of unvested equity and any acceleration
• Who pays for D&O run-off cover — a real cost, addressed in our guide to startup insurance
• Expense responsibility and any break fee
• Conditions to closing, stated as specifically as possible

The exclusivity clause is typically binding even when the rest is not. Read it as a contract, because it is one.

3. Purchase Price Versus Proceeds

The headline is a gross number. What arrives is what remains after:

• Working capital adjustment. The deal assumes a normal level of working capital; the price adjusts for any difference at closing. Negotiate how the target is calculated, not just the number — the methodology decides the outcome and is frequently drafted by the buyer's accountants.
• Debt and debt-like items. Buyers define these broadly — deferred revenue, accrued bonuses, unpaid taxes, capital leases, and sometimes uncollected sales tax.
• Transaction expenses. Banker fees, legal fees, accounting, and the D&O tail, generally deducted from seller proceeds.
• Escrow. Held back, released later, subject to claims.
• Retention pools. If funded from the purchase price, this is money moving from selling shareholders to employees who are staying. It may be entirely fair, and it must be understood before it is agreed.
• The preference stack. Then the remainder runs through the waterfall — which is why liquidation preferences decide what common stock actually receives.

Model your personal net proceeds before signing the LOI, using conservative assumptions on every one of these. Founders who do this sometimes discover that a lower headline with a cleaner structure pays them more.

4. Earnouts

An earnout defers part of the price against future performance. Buyers like them because they bridge valuation gaps and retain founders. Sellers should treat them with considerable suspicion.

The structural problem: once the buyer owns the business, the buyer controls the inputs to the metric. They set pricing, allocate sales resource, reorganise teams, change the product roadmap and decide how revenue is recognised. Even acting in complete good faith, they may make decisions that are right for the group and fatal to your earnout.

If you cannot avoid one:

• Use the simplest, most objective metric available. Revenue beats gross profit; gross profit beats EBITDA; EBITDA is highly manipulable through allocated overhead.
• Keep the period short — twelve months is far better than three years.
• Negotiate operating covenants: the business will be run as a separate unit, with a defined budget, retained sales capacity, and no reallocation of accounts, for the earnout period.
• Define the accounting. Which policies apply, how intercompany revenue is treated, what overhead may be allocated.
• Include acceleration on a further sale, on a material breach of the covenants, or if you are terminated without cause.
• Provide for information rights and a dispute mechanism — an independent accountant, with defined timelines.

And then discount it in your own model. Assume it pays little or nothing, and decide whether the deal is still one you want. Earnout disputes are among the most common forms of post-closing litigation for a reason.

5. Reps, Warranties, Escrow and Indemnity

You will make extensive representations about the company — ownership of shares, title to IP, accuracy of financials, tax compliance, material contracts, employment matters, absence of litigation, data protection, and much else. If a representation turns out to be wrong, the buyer can claim.

The mechanics that matter:

• Escrow / holdback. A portion of the price held to satisfy claims. Historically often around 10% for 12 to 18 months; R&W insurance has pushed both figures down substantially in many deals.
• Survival periods. How long each representation lives. General reps commonly 12 to 24 months; fundamental reps — authority, capitalisation, title — far longer, sometimes indefinitely. Tax reps typically run to the statute of limitations.
• Caps. The maximum recoverable, usually the escrow for general reps, and often the full purchase price for fundamental reps and fraud.
• Baskets and deductibles. A threshold before claims can be made. A tipping basket pays from the first dollar once crossed; a true deductible pays only the excess. Deductibles are meaningfully better for sellers.
• Materiality scrapes. A provision reading materiality qualifiers out of the reps when calculating damages. Buyer-friendly, and worth resisting or narrowing.
• Sandbagging. Whether a buyer can claim on something it already knew about. Anti-sandbagging language protects sellers; buyers resist it.

Representations and warranties insurance

R&W insurance covers breaches of the seller's representations, with the buyer claiming against a policy rather than against the sellers. It has become standard in middle-market deals and has genuinely changed structure: escrows shrink, sometimes to a small retention or nothing, and sellers receive far more at closing.

The cost is a premium plus underwriting fees, frequently split or borne by the buyer, and it requires clean, well-documented diligence — which is another argument for having your house in order before a process starts. If your deal is large enough to qualify, ask about it early.

6. The Twelve Months Before a Process

Almost everything that costs a seller money is cheaper to fix before anyone is looking. A year of unglamorous preparation reliably returns more than any amount of negotiation once exclusivity has been granted.

Twelve months out — fix the record. Reconcile the cap table to signed documents, chase every missing board consent, obtain invention assignments from every contributor including former contractors, and confirm the 409A series is complete and defensible. These are the findings that reduce price, and every one of them is fixable now and awkward later.

Nine months out — quantify the tax exposures. Multi-state sales tax nexus, contractor classification, payroll registrations in states where you have remote staff. If there is exposure, get it measured by your accountant and, where sensible, remediated through a voluntary disclosure programme. A quantified and addressed exposure is a footnote; an unquantified one becomes an indemnity or a price reduction.

Six months out — read your own contracts. Specifically the change-of-control and anti-assignment clauses in your largest customer agreements. If your top accounts can terminate or must consent on a sale, your buyer will discover it and your customers acquire leverage over your deal. Knowing which consents you need before you start is the difference between a scheduling exercise and a crisis.

Six months out — build the quality-of-earnings story. Buyers run a quality-of-earnings analysis on your revenue. Anything unusual — one-off deals presented as recurring, revenue recognised early, related-party arrangements — will surface. A sell-side quality-of-earnings report commissioned by you costs real money and frequently pays for itself by removing the buyer's ability to re-trade on a finding you could have disclosed.

Three months out — assemble the data room properly. Complete, current, indexed and reconciled to the deck, as our guide to data rooms sets out. A sale process needs a transaction-grade room with bidder separation, not the one you used for your Series B.

Throughout — keep running the business. Roughly a third of signed LOIs never close, and a company that took its eye off execution during a failed process is in a materially worse position than when it started. Missing your numbers mid-diligence is also the single most common trigger for a re-trade.

7. The Diligence Findings That Actually Reduce Price

From experience across the market, the recurring ones:

• Missing IP assignments. An early contractor who never signed an invention assignment can hold up an entire transaction.
• Uncollected sales tax. Multi-state nexus exposure becomes a direct price reduction or an indemnity — see our guide to startup accounting.
• Contractor misclassification. Back taxes and penalties, quantified and deducted.
• Cap table discrepancies. Options granted without board approval, missing consents, a share count that does not reconcile.
• Defective 409A history. Gaps in the series, or valuations that cannot be defended.
• Customer concentration and non-assignable contracts. Change-of-control clauses requiring consent give your largest customers leverage over your deal.
• Open-source licence non-compliance. Particularly copyleft licences in a proprietary codebase.
• Unresolved employee equity promises made by email and never documented.

Every one of these is cheaper to fix in advance than to negotiate under exclusivity.

Frequently Asked Questions

Asset sale or stock sale?

Buyers generally prefer asset purchases — a step-up in tax basis and the ability to leave liabilities behind. Sellers generally prefer stock sales, which are cleaner and, for holders of qualified small business stock, can be dramatically better after tax. This is a genuinely large economic question and belongs in the LOI, not in week eight.

What happens to unvested employee options?

It depends entirely on the plan and the agreement. Options may be assumed, cashed out, accelerated, or cancelled. Employees with double-trigger acceleration are protected only if they are terminated after closing. This is one of the most emotionally charged parts of any deal, and getting ahead of it with clear communication matters.

Do we have to tell employees?

Not until late, and confidentiality is usually contractual. Practically, a small number of people must be involved in diligence. Plan communications deliberately — an information leak mid-process is disruptive and can cost you the deal.

Can our investors block the sale?

Frequently yes. Protective provisions typically give preferred holders a veto over a sale, sometimes with a price threshold. Drag-along provisions cut the other way, forcing minority holders to go along with an approved deal. Read both before you start — our guide to board governance and protective provisions covers the mechanics.

How long does it take?

Four to nine months from first serious conversation to closing is typical, with confirmatory diligence and documentation consuming most of it. Regulatory clearances extend it further. Run the business as though the deal will not happen, because roughly a third of signed LOIs do not close.

Should we run a process or negotiate with one buyer?

A process, almost always, if you have any choice. Competition is the only reliable source of seller leverage, and a single buyer who knows they are unopposed prices accordingly and re-trades more readily. The exception is a genuinely pre-emptive offer at a price a process would not beat — which does happen, and which you can only assess if you have a realistic view of what the market would pay. Even then, running a short, quiet process among three or four credible parties usually improves the outcome more than it costs in time.

What if the buyer re-trades after diligence?

Establish first whether the finding is real. Buyers sometimes discover genuine problems, and a price adjustment for a quantified tax exposure is legitimate. What is not legitimate is a vague reduction late in exclusivity with no specific basis — that is a negotiating tactic that works because your alternatives have disappeared. The defences are all upstream: a hard exclusivity end date, a clean pre-process clean-up so there is nothing to find, and a genuine willingness to walk. Founders who have modelled the standalone plan negotiate very differently from those who have not.

How much of the price should be cash at closing?

As much as you can get. Every other component — escrow, earnout, buyer stock, deferred consideration — carries risk you do not control, and each should be discounted accordingly in your own model. Buyer stock deserves particular attention: if it is illiquid, locked up, or in a private company, you have exchanged a known amount for an unknown one and taken on a second investment decision you may not have wanted to make.

What do we owe our investors during a process?

More than founders sometimes provide. The board should be informed of any serious approach immediately, and directors have fiduciary duties that become sharper the moment a sale is contemplated — particularly where preferred and common receive materially different outcomes. Where the board is conflicted, a special committee of disinterested directors is the standard mechanism. Handling this properly protects the deal as much as it protects anyone; a transaction approved through a defensible process is far harder to challenge afterwards.

The Bottom Line

Get everything into the LOI, because that is when you still have leverage. Model net proceeds rather than headline price. Discount the earnout heavily. And clean up the recurring diligence findings in the year before anyone looks, because each one becomes a price negotiation you conduct from the weaker side of the table.

Global Capital Network connects founders with acquirers, investors and the advisors who run these processes. Get in touch or see our upcoming events.

This article is general information, not legal, tax or financial advice. M&A structures are highly fact-specific. Engage experienced transaction counsel before signing anything, including an LOI.