Founding team
Founder agreements and vesting: equity, exit, what to write down
General information on the law in Germany (as of August 2026), not legal advice. Other countries regulate this differently.
Four points belong in writing before you start: who holds what and on what grounds, what happens to a share when someone stops, who decides what in a deadlock, and who owns the code, the brand and the customer list. The expensive one is the second. Without a vesting clause a founder who leaves after seven months keeps their full share and their full vote, while everyone else keeps working. The standard is four years with a twelve-month cliff. Under German law a clause only holds if it is time-limited, distinguishes good from bad leaver and pays appropriate compensation, and a non-compete against a departing shareholder ends after two years.
The most expensive minute of the project
Two people decide to build something together. Somewhere between the second coffee and the first prototype one of them says: fifty-fifty, obviously. The other agrees, because disagreeing at that moment would mean saying out loud that they do not consider the other person an equal. The whole exchange takes under a minute, and it decides more about the following four years than most decisions that get a proper meeting.
Noam Wasserman looked at how founding teams actually do this and found that 73 percent fix the split within a month of founding. The speed is the finding, not the number itself: a fast, equal split correlates with worse outcomes, because it skips the conversation about who brings what and what happens if that changes. In his sample, roughly two thirds of failures traced back to people rather than to product or market.
What the research says about teams
Teams that fix the split within one month
73 %
Failures traced back to people, not product or market
65 %
German start-ups where the team, not the idea, decided the outcome
60 %
None of these numbers is about the idea. All of them are about an agreement nobody wanted to have while things were still going well.
Split speed and the people share: Wasserman and Hellmann, "The First Deal: The Division of Founder Equity in New Ventures", and Wasserman, "The Founder's Dilemmas". The 60 percent is from a study by the German RKW Kompetenzzentrum covering 100 ICT start-ups in the German-speaking region. Different samples and definitions, so the figures sit next to each other, they do not add up.
The four points that belong in writing
Not a template and not forty pages. Four questions. Every one of them has a default answer that applies whether or not you ever discuss it, and in three of the four cases that default is the one you would not have chosen.
Who holds what, and on what grounds
What exactly is each share being given for?
The hard part is not the number, it is the reasoning underneath it. A team that can explain the split in terms of time, money, experience and risk brought in still has an answer two years later. A team that explains it with "there are two of us" does not.
If left open: Later arguments about contributions, not about numbers
What happens when someone stops
Does a founder keep their full share if they leave after seven months?
With nothing written down: yes. The share is theirs regardless of how long they stayed. That is precisely what vesting is built against: the share is earned over time instead of given away on day one. This is the point most agreements are silent on, and the reason most separations get expensive.
If left open: A passive shareholder with a full vote, permanently
Who decides what, and what breaks a tie
What happens at 50-50 with two different opinions?
Nothing. That is the problem. A deadlock blocks every decision that needs a majority, up to and including the ability to act at all. Write down which decisions need everyone, which need a simple majority, and how you break a tie.
If left open: Paralysis at the worst possible moment
Who owns the ideas, the code and the customer list
Does what you build belong to the company or to the person who built it?
Without an explicit transfer, copyright stays with the author. Whoever wrote the code or drew the brand takes it with them if it comes to that. The same goes for the customer list. One line in the agreement settles it for everyone.
If left open: The product belongs to a person, not to the company
Why splitting by headcount is rarely the fair split
Equal shares can be the right answer. They are almost never the right answer for the reason they get chosen, which is that nobody wanted to weigh anything. Five things actually differ between founders, and they differ by a lot more than most teams assume before they sit down and compare.
How much each contribution usually weighs
Full time rather than alongside
By far the biggest difference and the one most often ignored. Someone who quits their job and goes full time carries a different risk from someone contributing three hours in the evening. Over four years the gap is not gradual, it is a multiple.
Money put in
Capital is the only contribution that can be stated exactly, which is why it is rarely argued about. The one thing to settle is whether it is equity or a loan. One buys shares, the other gets paid back.
What already existed
A working prototype, a live customer relationship, a registered trademark. Something the others would otherwise have had to build or buy. It counts once, not forever.
Experience and access
Someone who knows the market saves the team months. The contribution is real and still routinely overvalued, because it cannot be measured in hours. Usable as a premium, rarely as the main argument.
The original idea
The most overvalued item on the list. The idea decides what everyone talks about, but it does no work. A premium for it is normal and fair; a majority for it almost never is.
The bars illustrate how much weight each contribution usually carries in a founding team, they are not measured data. The order matters more than the exact width, and your own case may reorder it.
One practical suggestion: write down each person's answer to all five points separately, then compare. The disagreement almost never sits in the final percentage, it sits in what each of you thought the other was contributing. That is the conversation, and it is a great deal cheaper now than after the first investor asks for a cap table.
What a vesting clause is worth, as a number
Vesting means the share is earned month by month instead of being owned on day one. Before the cliff nothing is earned; after it the share accrues to the end of the period. The default below is a founder leaving in month seven, because that is inside the cliff and the two numbers are furthest apart there.
Try it yourself
Move the last slider across the cliff and watch the second number appear. Nothing is stored or transmitted.
An equal split between two founders is 50 percent each. That is the most common starting point and rarely the fairest one.
Four years is the market standard, and investors will expect it whether or not you have already agreed something else.
Nothing is earned before this date. Twelve months is standard: long enough to see whether someone really stays.
Set this to the month you would least like to think about. That is the one the clause is written for.
The clock
What the departing founder keeps
The exit falls inside the cliff. Without a clause the company gives away 50 percent for 7 months of work. With one, 50 percent stays with the people who are still there.
The model is the standard one: nothing before the cliff, linear afterwards. Real agreements often add quarterly steps, acceleration on a sale and separate good and bad leaver treatment. The result is an order of magnitude to think with, not a clause.
Good leaver, bad leaver, and no leaver clause at all
Vesting says how much of the share has been earned. A leaver clause says what happens to it, and at what price, depending on why the person is going. The third row is what applies to most founding teams, because they never wrote a first or second row.
| Case | Typical example | What happens to the share | At what price |
|---|---|---|---|
| Good leaver | Illness, relocation, an amicable split with no fault | The vested part stays, the unvested part returns | Market value, often with a discount of around 20 percent |
| Bad leaver | Serious breach, dismissal for cause | Even the vested part can return | Nominal value or acquisition cost, sometimes with a steep discount |
| Nothing agreed | The normal case when friends found together | The full share stays, whatever the duration | No buy-back right, only a voluntary sale at the seller's price |
Discount ranges are what German practice commonly reports, not fixed rules. What is appropriate in a specific agreement depends on the case and belongs in front of a lawyer.
What applies if you write nothing
The common belief is that a team without an agreement has no agreement. It has one, it just did not choose it. Two people pursuing a common purpose form a partnership under German law without paperwork and without intending to, and the statutory rules apply from that moment.
| Form | The agreement itself | If someone leaves | The share |
|---|---|---|---|
| Partnership (GbR) | No form required, valid orally, comes into being without intent | Since the 2024 reform the partnership continues and the leaver exits | A compensation claim against the partnership, amount often disputed |
| Limited company (UG, GmbH) | Notarial deed mandatory, without it the company does not exist | No right to walk out at will, the share stays with the shareholder | Transfer only by notarial deed, so never settled in passing |
One change is worth knowing about. Until the end of 2023, notice by one partner dissolved the partnership. Since the reform of German partnership law took effect on 1 January 2024, the partnership continues and the person giving notice simply leaves, with a claim to compensation against the business. That is better than dissolution and still leaves the one question that causes the argument: how much the claim is worth.
General orientation on German law, not legal advice. The document itself belongs with a notary for a limited company, and with a lawyer in every case where more than one person holds a share.
Four limits that make a clause worthless
Most founder agreements circulating in German start-up circles are translated from US templates. Several of the clauses in them do not survive a German court, and an unenforceable clause is worse than none: it creates confidence that turns out to be unfounded exactly when it is needed.
Unlimited duration does not bind
Vesting periods of three to four years are accepted. Tying a founder permanently to an unvested share, or being able to remove them at any time without cause, risks the clause being void as contrary to public policy.
Compensation of zero is the most common trap
Unvested shares returning at nominal value is standard practice. Losing already vested shares for nothing is not. The further a clause sits from market value, the less likely it survives review.
Good and bad leaver have to be distinguished
A clause that treats every exit alike reads as an unreasonable disadvantage. Someone who falls ill is not the same case as someone who damages the company.
Non-competes end after two years
The German Federal Court of Justice voided a customer-protection clause against a departing shareholder because it exceeded the two years that are permissible as a rule (judgment of 20 January 2015, II ZR 369/13). Longer periods are the exception and need justification.
Case reference: German Federal Court of Justice, 20 January 2015, II ZR 369/13. General orientation, not legal advice.
Five mistakes that show up years later
Every one of them is comfortable at the time it is made. That is what they have in common, and it is why they are so hard to avoid.
Fixing the split in week one and never looking at it again. At that point nobody knows who actually delivers. Revisiting the split after six months is not weakness, it is the only moment at which it can still be corrected without a fight.
Choosing equal shares to avoid the conversation. Splitting by headcount is fast, looks fair and only postpones the real question. It gets asked later, when one of you is working full time and the other has stopped replying.
Skipping vesting because you trust each other. Vesting distrusts nobody, it only describes what should happen when someone's life changes. The moment you are able to agree it is exactly the moment nobody thinks it is necessary.
Copying an English-language template. Bad-leaver clauses from US agreements regularly reach too far under German law and are then unenforceable, which turns the supposed safety net into its opposite.
Settling everything verbally and considering it handled. A partnership comes into being with no form and no intent as soon as two people pursue a common purpose. Writing nothing down does not leave you without an agreement, it leaves you with the statutory one.
Frequently asked questions
Does a founding team really need a written agreement?
For a German limited company it is mandatory and has to be notarised, otherwise the company does not come into existence at all. For a partnership no form is required, and that is the trap: the partnership exists anyway, but on statutory terms. A written agreement does not replace trust, it only describes what applies once the situation changes.
What is vesting and why should founders agree to it?
Vesting means a share is earned over time instead of being owned on day one. The standard is four years with a twelve-month cliff: leave before the cliff and you keep nothing, after it the earned share accrues monthly. Without vesting, a founder who leaves after seven months keeps their full share and keeps voting, while the others do the work.
Are equal shares a good idea for founders?
Sometimes, but rarely for the reason they get chosen. Research by Noam Wasserman found that 73 percent of teams fix the split within a month of founding, and that the fast equal split in particular goes together with worse outcomes. Not because equal is wrong, but because the conversation about time, money and risk was skipped.
What happens if a founder leaves and nothing was agreed?
Their share stays with them. In a German partnership, notice no longer dissolves the business since the 2024 reform: it continues, and the leaver gets a compensation claim whose amount is regularly disputed when nothing was written down. In a limited company the share stays with the shareholder in any case, and a transfer requires a notarial deed.
Can a clause simply remove a founder?
No. Clauses that let you exclude a shareholder at any time without cause, or take already vested shares without compensation, are open to challenge under German law. What is accepted is time-limited vesting, a distinction between good and bad leaver, and appropriate compensation. Post-contractual non-competes are limited to two years as a rule.
Read next
Is YOUR idea worth the agreement?
Before you split anything, find out what there is to split. 4 to 12 AI experts assess market, financials and risks of your specific idea, in minutes, from € 5.99.
Assess my idea →