If your technical co-founder isn't delivering, diagnose with evidence first, protect delivery second and touch equity last. Read your vesting and leaver clauses before any conversation, because they decide which options you actually have.
This guide is for a non-technical founder who can't judge the code and has to decide anyway. It sits within our startup CTO guides. It's general information, not legal advice.
How do you tell a slow co-founder from a stalled one?
Start a dated log today. Feelings don't survive a hard conversation, and a log does. VentureBeat's 2016 piece on struggling technical co-founders lists ten signs. Most of them are things you can write down: missed commitments, evasive answers when you ask about progress, jargon where a plain answer should be, poor communication on deadlines, doing only what's asked, treating meetings as progress, calling hard things "easy", dismissing input, team departures, and your own instincts. Its advice on what to do is short: document objectively, talk in person, act quickly.
Treat the last sign as a prompt to start logging, not as evidence. The rest you can record with a date.
The table is our judgment, not a validated test. It separates a co-founder who is slow but sound from one who has stalled.
| Signal | What to log | Slow but sound | Stalled |
|---|---|---|---|
| Missed commitments | Date promised, date delivered, reason given | Misses are rare, flagged early, with a new date | Repeated misses, surfaced on the day, new dates also missed |
| Evasive answers | The question you asked and what came back | Gives a straight answer, even a bad one | Answers change, or move to a different topic |
| Jargon | Moments you couldn't follow a status update | Will re-explain in plain words when asked | Gets irritated when asked, or explains less each time |
| Deadline communication | Who raised each slip, and when | They tell you first | You find out from a customer or a teammate |
| Output | Something you can see or click each week | Small, visible progress | Meetings and plans, nothing shippable |
One bad week proves nothing. In our judgment, a pattern across six to eight weeks, logged by date, is something you can put in front of them.
The test that cuts through most of this is a written milestone. "Ship the checkout flow by the 14th" can be met or missed. "Be more responsive" can't. If you've never written expectations down, the first honest step is to do that, because a co-founder can't fail a standard you never stated.
Who can check the work if you can't read the code?
Hire a senior engineer you trust, with no stake in the outcome, to review the work for a few days. This is the same move as step 3 of the co-founder vetting and vesting checklist, applied after the fact instead of before the deal.
Scope it tightly, and ask for one short written report. We'd give the reviewer access to the repository, the deployment setup, and the last eight weeks of commits and technical decisions. The eight weeks is our judgment, long enough to show a pattern and short enough to read in a day or two.
Ask the reviewer three things in plain words:
- Is real, working software being produced, and is it going out to users?
- Are the decisions sensible for a company this size, or is the stack more complicated than the product needs?
- If this person left tomorrow, could someone else pick the code up?
You're looking for a difference between "slow because the problem is hard" and "slow because little is happening". A reviewer can tell you which. You can't, and it isn't a failing to say so.
Tell your co-founder the review is happening. Doing it behind their back turns a delivery question into a trust question, and you'll need their cooperation for access anyway. If they refuse a neutral review of company work, that refusal goes in the log.
What does a 30-day recovery plan look like?
The 30 days are our judgment, not a standard. It's long enough to ship something real and short enough that you aren't funding drift. Define "met" and "not met" before day one, in writing, so nobody argues about it on day thirty.
Week 1: have the direct conversation, with the log
Talk in person, as VentureBeat advises. Lead with the pattern, not the person: three dates, three commitments, what actually happened. Then ask what's in the way. The answer tells you whether you're dealing with capacity (they're overloaded or still holding a day job), skill (the work is beyond them right now) or commitment (they've mentally left). Each calls for a different fix, so listen before you propose one.
Redefining the role is a legitimate outcome. A co-founder who is strong on product and customers but weak at shipping may do better with a narrower remit and an engineer to carry the build. That's no substitute for the milestones below.
Weeks 1-2: agree three shippable milestones
Each milestone needs three things: a result you can see, a date, and a named owner. "Payments live for ten pilot customers by the 21st, owned by the co-founder" works. Three is enough. More gives everyone somewhere to hide.
If you have your outside reviewer, ask them to sanity-check that the milestones are achievable. An impossible target tests your estimating, not their delivery.
Weeks 3-4: review against the milestones
Hold a review at the end of week four, with the reviewer if you can. For each milestone: met, partly met, or missed, judged against the definition you wrote on day one. Two met and one partly met is a recovery. Three missed, with the same pattern as your log, is an answer.
What do your vesting and leaver terms say?
Open the documents before the conversation, not after. Which options you have depends on your shareholder agreement, your articles of association, and your jurisdiction. Get a startup lawyer to read them, because this is a summary of what three published sources describe, not a reading of your contract. The vetting checklist linked above has the fuller vesting explainer.
Y Combinator's Michael Seibel says a typical setup is four years of vesting with a one-year cliff, and that if you leave or get fired within a year you walk away with nothing. He frames the purpose directly: if you fear what will happen if you have to break up with a co-founder, make sure you have a proper vesting schedule.
Orrick, a law firm, says German founders' shares can be subject to vesting, but only through a separate contractual agreement, and it describes a four-year schedule with a twelve-month cliff as the typical German schedule. Under reverse vesting the founder stays a registered shareholder with voting rights, and the shares may be subject to a call option if they leave early. Orrick says bad leaver events typically include fraud, gross misconduct and similar bad faith actions. Orrick describes good and bad leavers, and some deals add a grey tier.
Ashfords, a UK law firm, defines a good leaver as typically someone leaving for reasons beyond their control, such as illness, death, redundancy or retirement, and a bad leaver as typically someone dismissed for cause. Unvested shares are typically forfeited whichever category applies. Vested shares of a bad leaver may be repurchased at nominal value. Ashfords also notes that under the latest BVCA model articles, voluntary resignation is no longer treated as an automatic bad leaver event.
Two things follow, both our judgment. First, a co-founder's removal is not a matter of one sentence in a chat. Whether and how you can remove someone depends on those documents and on local law, so don't assume you can, and don't say it to them as a threat.
Second, if there's no vesting at all, the shares are probably already theirs, and changing that needs their consent. That makes the conversation a negotiation, and a recovery plan with a fresh vesting schedule attached becomes your best offer.
How do you keep shipping while you decide?
Don't let the product wait on a decision about a person. Do two things in parallel.
First, check who owns the keys. The repository, cloud accounts, domain registrar, app store accounts and secrets should sit in company-owned accounts with a second admin, not on one person's personal login. If any don't, move them this week and say plainly that it's housekeeping. Our guide on a 30-day plan for a leaving CTO has the access and secret-rotation steps, and they apply here too, since you want them in place whether or not your co-founder stays.
Second, add outside cover. There are two shapes. A senior engineer takes some of the build load, which helps if the issue is capacity. A fractional CTO takes technical direction and works with your co-founder, which helps if the issue is decisions or priorities. Your review will show which gap is bigger.
Bringing in cover while the co-founder remains is not a verdict on them. Frame it as protecting the milestones.
When do you restructure the equity and when do you part ways?
There are four paths. The table is judgment, built on the mechanics above.
| Path | When it fits | What it needs |
|---|---|---|
| Coach and re-scope | Capacity or role problem, they're committed, the review shows real output | Written milestones, a narrower role, possibly an engineer to carry the build |
| Re-cut with new vesting | They're committed and want to stay, but no vesting exists or it's already fully vested | Their agreement and a lawyer; expect it to be a negotiation |
| Fractional CTO over them | Judgment gap: output is real but the decisions are weak | Clear decision rights, buy-in from them, a defined end date for the arrangement |
| Part ways via leaver terms | Commitment is gone, or the 30-day review shows no recovery | A lawyer reading your agreement, articles and local law before you act |
A decision test: if the review is positive and milestones are met, stay with paths one or two. If the output is real but the decisions are poor, path three. If you've documented a pattern, they've had a fair 30 days and nothing changed, plan the exit. Do it through whatever your documents provide, not around them.
If you do part ways, you'll be looking for a replacement, and what stake to offer is its own question. Sizing a replacement co-founder's stake covers it.
What this means for your team decision
Your decision is the shape of the technical team, and you should make it inside the 30-day window: keep your co-founder in place, re-scope their role, add interim cover, or exit.
Use two inputs. The first is the outside review, which tells you whether the output is real. The second is your vesting position, which tells you what each option costs. If you have good evidence and a clear position, choose one path by the day-30 review and tell your co-founder which. If you have neither, buy the missing piece before you decide: in our judgment the review is a few days of a senior engineer's time and the lawyer a few hours, and both are small next to a cap-table mistake.
Don't wait for the 30 days before covering delivery. Your customers and your next funding conversation depend on the product moving. To compare the cover options and what each costs, read our guide to interim cover options and cost.
Getting senior engineering cover through HighCircl
HighCircl matches companies with senior engineers from seven EU countries and sends a shortlist of three to five candidates within 72 hours. About 1 in 10 senior engineer applicants passes a four-stage vetting process run by senior engineers. There's no minimum hour commitment, so you can add delivery cover now and change it after your day-30 review. Senior engineer rates run €45-105/hr ($50-115/hr), with a 20% capped margin on top of what the engineer earns. If an engagement isn't working, HighCircl replaces the engineer at no additional recruitment cost. Leadership cover is a separate question; see HighCircl's fractional CTO hiring page.
FAQ
Can you fire a technical co-founder?
It depends on your shareholder agreement, your articles, any vesting and leaver terms, and your jurisdiction. In our judgment, a co-founder who holds shares is a different case from an employee, so removal isn't like ending a contract. Y Combinator notes that with a one-year cliff, someone who leaves or is fired within a year walks away with nothing. In our judgment, that only helps if the vesting is documented. Take the documents to a lawyer before you act.
What does technical co-founder mean?
It's the co-founder who owns the product's technology: the architecture, the build, and the engineers. They're usually a shareholder from the start rather than an employee hired later. If you can't judge that work yourself, an outside reviewer can.
Should co-founders split equity 50/50 or 51/49?
Y Combinator recommends equal or near-equal splits. Whether 50/50 or 51/49, we'd say the split matters less than the vesting under it, and that's our judgment. An equal split with proper vesting protects both people. A 51/49 split with no vesting protects neither.
How long should you give an underperforming co-founder to improve?
Thirty days against three written milestones is our suggestion, and it's judgment rather than a rule. Less than that and you haven't given them a fair test. More and you're funding drift. If the first week's conversation shows commitment is already gone, you don't need the full 30 days to know.
What if my co-founder refuses an outside review?
Log it. A neutral technical review of company work is a reasonable request, and a flat refusal tells you something about the commitment question. Offer a different reviewer or a narrower scope before you read too much into it.