Non-compete clauses: the three steps that protect the value of an M&A

In the previous article, on the earn-out, I closed with a warning: poorly designed, that clause protects no one, it merely pushes the fight past the closing. But there is a risk the earn-out, on its own, does not cover. What good is it to tie part of the price to the future performance of the business if the people who produce that performance — the seller, the key executives, the client base — can simply vanish the day after signing? The earn-out ties the price to the result. What is missing is tying down the people who deliver the result. It is this second knot, the natural sequel to that conversation, that I take up here. If you have not read the first part yet, it is worth starting there: Earn-out in M&A: the mechanism that ties the price to the future.

That ghost, paying dearly and then watching the value walk out the door, haunted Facebook when it bought WhatsApp in 2014 for roughly US$ 22 billion. On top of the price, the company set aside US$ 3 billion in restricted stock for founders and staff, with vesting over four years. It was the classic key-man retention: locking down the brains for as long as it takes the integration to catch. Did it work? Only up to a point.

Brian Acton, one of the founders, walked out early. According to the press, he left behind something close to US$ 850 million that had not yet “vested,” and departed tweeting “Delete Facebook.” The retention existed in the contract. It did not hold the man. I keep this scene because it sums up the thesis of this piece: the value of an acquisition rarely lives in the corporate registration alone. It lives largely in the people, the relationships, and the promise that no one will sabotage what was bought.

Jim Collins, in Good to Great, argued that a good leader's first question is not “what” but “who”: first get the right people on the bus, then decide where it is going. In M&A, the buyer inherits the bus already in motion. The cruel detail is that some important passengers can get off at the next stop, taking the map, the clients and the know-how with them. Classic Harvard Business Review studies estimate that most acquisitions fail to deliver the promised value, and the loss of talent ranks among the most cited causes (Harvard Business Review, “The Big Idea: The New M&A Playbook”).

To contain that risk, there are what I call the “staircase clauses”: three steps that protect the investment after closing. They are the non-compete (non-competition), the non-solicitation (no poaching) and the key-man retention (retention of key talent). The first two rely on prohibition. The third relies on incentive. Together, they mark the difference between buying a company and buying the right to watch it empty out.

Step 1: non-compete, the brake the law already knows

The non-compete is the oldest and best regulated of the three. The seller promises not to set up a competing business or work for a rival for a period. And here is news that surprises many people: in Brazil, this prohibition already exists by default, even without a written clause.

Article 1,147 of the Civil Code (Law 10,406/2002) states that, absent express authorization, whoever sells an establishment may not compete with the buyer for the following five years. In other words, the five-year term is the ceiling the legislator itself deemed reasonable. Brazil's antitrust authority, CADE, has followed the same line: under its settled guidance, non-compete agreements of up to five years are lawful, provided they are tied to protecting the goodwill of the business sold (CADE, Guide for the Analysis of Horizontal Mergers).

The rule of thumb, then, is simple. A short term passes without a scare: two or three years are rarely challenged. Five years is the limit. Go beyond that, or throw in “all of Brazil” without technical justification, and the odds of a judge trimming the excess rise. Brazil's Superior Court of Justice (STJ) has been clear: a clause with no time limit is void. In a recent ruling (REsp 2,185,015/SC), two former partners in children's clothing retail had carved up the market by garment size, with no term at all. The court declared the non-compete void precisely because it was perpetual, but preserved the rest of the agreement. Locking up the market forever does not fly.

There is also a detail that separates amateurs from professionals: compensation. When the restriction falls on an employee, the labor courts require financial compensation to validate the clause. Without payment, the Superior Labor Court (TST) tends to consider it abusive. It makes sense: if you want someone to sit still, pay for the silence. When the restricted party is the seller, however, the sale price usually already amounts to sufficient compensation. That is why an owner who pockets millions will hardly convince a judge that he was forced, for free, not to compete.

It is worth a look abroad. In the United States, the FTC tried, in 2024, to ban almost all non-compete clauses for employees, arguing that they lock up the labor market. The measure was ultimately struck down in court, but it signaled a trend: regulators worldwide are watching restrictions that are too broad. In Brazil, CADE's message has been the same for a long time. A clause that seems to exist only to lock up the market, rather than to protect the value bought, lands in the antitrust authority's crosshairs.

Step 2: non-solicitation, the pact not to hunt in someone else's backyard

The non-solicitation is the lighter cousin of the non-compete. It does not stop the seller from working; it stops them from going after employees, clients or suppliers of the company they sold. The former owner may even open another business, if there is no non-compete, but cannot call the old contact list and empty out the team overnight.

Because it is more surgical, this step tolerates less generalization. “Any potential client” is the kind of drafting that breeds disputes. The ideal is to list who is protected, or at least set an objective criterion (clients active in the last twelve months, for example) and tie it to a term. The more specific the clause, the lower the risk of annulment. Vagueness, in a contract, is an invitation to litigation.

And here is a detail many people forget: the LGPD, Brazil's General Data Protection Law (Law 13,709/2018). Policing a non-solicitation sometimes tempts the buyer to monitor emails, calls or business contacts. Be careful. The law requires collecting only the minimum necessary to verify compliance with the obligation (the principles of purpose and minimization) and discarding the data as soon as it is no longer useful. Surveillance without a legal basis turns a protective clause into a data-protection liability.

Step 3: key-man retention, when locking down becomes inviting

We reach the most elegant step. Key-man retention inverts the logic of the first two: instead of prohibiting, it invites. Stay bonuses, accelerated vesting, earn-outs and profit-sharing plans exist so that the team that keeps the wheel turning keeps pushing during the transition. No one holds on to talent by shouting. You hold it with a project and with money, in that order.

The contrast between WhatsApp and Instagram, both bought by Facebook, illustrates the difference well. While Acton slammed the door, Kevin Systrom and Mike Krieger, of Instagram, stayed six years, an enormous stretch for founders after a sale. Instagram's retention worked because, for a good while, there was autonomy and purpose, not just vesting. When those two fuels ran out, the founders left anyway. The lesson fits an HR proverb: money holds the body; purpose holds the soul.

The clause is entirely valid in Brazil, as long as it does not offend common sense. Demanding a disproportionate penalty from someone who decides to leave, or confiscating already-vested rights, is the shortest path to nullity. Key-man retention should be an incentive, not a leash. Talent kept in place by fear performs worse than talent that stays by choice.

Because this step usually goes hand in hand with the earn-out, the risk is familiar: payment tied to future performance becomes a source of conflict when the targets are poorly designed. That was precisely the subject of the first part of this pair of articles, which I recommend keeping in mind as you read this one: Earn-out in M&A: the mechanism that ties the price to the future.

The Linx-Stone case: the three steps under the spotlight

There is a Brazilian case that brings together all three steps and even became a governance lesson: Stone's purchase of Linx in 2020, in a deal of around R$ 6.7 billion.

In the package, beyond the price, Stone and Linx negotiated non-compete and non-solicitation terms with the founding shareholders (Nércio Fernandes, Alberto Menache and Alon Dayan) and a services agreement with the company's then-president, Menache. A perfect portrait of the staircase: non-compete, non-solicitation and a retention tie, all in the same SPA.

The problem is that the non-compete was valued at around R$ 185 million, earmarked precisely for the controlling shareholders. The technical staff of the securities regulator (CVM) flagged the issue: was that payment for the clause, or a disguised premium the minority shareholders would never see a cent of? The legal question was whether it constituted a “particular benefit,” which would bar the founders from voting at the shareholders' meeting. The partners appealed to CVM's board and, by majority, the view that the contracts were not a particular benefit prevailed. The deal was approved by the shareholders and, later, by CADE itself (Non-compete clauses and fierce competition in delivery).

The Linx-Stone case teaches the main point: a poorly calibrated staircase clause either dies or makes headlines. The value of the non-compete must be proportionate to the real sacrifice, or it risks looking like a way to reward the controlling shareholder at the minority's expense. Transparency, on this terrain, is not frills. It is armor.

How to build the staircase without tripping

An interesting feature of these three steps is that case law and CADE talk to each other: their benchmarks of reasonableness are close. After more than twenty years watching M&A contracts cross my desk, I have gathered the recommendations that most reliably prevent headaches:

• Be specific. Define the activities, the regions and the clients covered. The less generic the clause, the lower the risk of annulment.
• Use short terms. On the seller's side, start by defending two years and only stretch it with a clear technical reason (long-maturation sectors, for example). The buyer's lawyer tends to open at five. Settling at three is usually a civilized middle ground.
• Pay for the sacrifice. Whoever asks for silence offers something in return: salary, a lump sum, a bonus. An employee non-compete without compensation is void under the case law. For the seller, the company's price generally already settles it.
• Explain the why. Put the economic rationale for the restriction in writing. If the courts or CADE ask for explanations, the paperwork will already be ready.
• Mind the data. Collect only the minimum necessary to police compliance and discard the rest. The LGPD is here to stay.
• Include a plan B. A severability clause lets a judge cut the excess and keep the rest standing, instead of bringing down the entire contract.

It is worth recalling where these steps appear on the timeline of a sale process. Talks begin far back, in the NDA or the LOI, take shape during due diligence, and flow into the SPA, the purchase and sale agreement, where the staircase clauses actually live. Good M&A advisory designs these steps alongside the valuation, not as an appendix stitched on the eve of closing: Due diligence: what the buyer really investigates.

In the end, the three clauses answer the same question: how do you ensure the value you bought does not evaporate the next day? With a dose of common sense, you can protect the investment without suffocating former partners, employees and the market. The balance between security and freedom is what sustains a transaction long after the signature.

Buying a company is easier than keeping it from leaving with the people who built it.