When a Startup in Your Cohort Stops
Spotting it early, having the conversation, settling the money and space, and recording it honestly. The part of incubation nobody plans for.
Nobody plans for it, and it happens every cohort
Programme handbooks cover selection, mentoring, milestones and graduation. Almost none of them cover what to do when a company simply stops - the founders go quiet, the desk is empty, and there is undisbursed money sitting against their name.
It happens in most cohorts. Handling it well takes an afternoon; handling it badly leaves committed funds locked up, a space you cannot reallocate, and outcome numbers you cannot explain.
Notice it early
The signal is almost never an announcement. It is a pattern:
- Two consecutive missed reporting cycles
- No response to direct contact over two or three weeks
- The desk unused
- A milestone deadline passing with no request for an extension
Any one of those is normal. Three together means something has changed.
The reason founders do not tell you is straightforward: it is embarrassing, and they hope to have better news before the conversation happens. Waiting for them to raise it means waiting a long time.
This is also the practical argument for tracking who has not filed an update rather than only reading the ones that arrive. A missing update is invisible unless something is holding the expected list against the received list.
Have the conversation early and make it easy
The conversation you want is not an interrogation. It is an offer of a clean exit.
Something close to: we have noticed things have gone quiet, that is completely normal and it happens in every cohort, we would rather know where things stand than not, and there is a straightforward way to close this out that does not involve anyone being embarrassed.
Three outcomes, and it is worth naming all three so the founder can pick:
- Paused. The team intends to continue but something has interrupted - a job, health, a co-founder leaving. Agree a specific date to revisit, and pause commitments rather than closing them.
- Pivoting. Still working, on something different. This needs a decision: is the new thing still eligible for the programme and the funding? Sometimes yes, sometimes not, and it needs a recorded answer either way.
- Stopped. The company is not continuing. Close it out properly.
A founder who is given three named options usually picks one honestly. A founder asked "what is happening?" says "we are figuring some things out", which is not information.
Close out the money properly
This is the part with consequences beyond the one company.
- Release undisbursed commitments back to the source. Formally, with a date and a reason. Money left committed to a stopped company is money the programme cannot commit to anyone else, and because the totals still reconcile, nobody notices for months.
- Record what was disbursed and what it bought. You will still have to account for it. A tranche released against a milestone that was met is a legitimate use of funds even if the company later stopped - but only if the milestone and evidence were recorded at the time.
- Check the recovery position. Some schemes require unspent disbursed funds to be returned. Read the sanction terms rather than assuming, and if recovery applies, start it while the founders are still responsive.
- Note it in the next utilisation report rather than letting an auditor find it. A stopped company explained in advance is administration; the same fact discovered independently is a finding.
Close out the space
If they had a desk or a cabin:
- Agree a vacating date in the same conversation
- Settle the deposit against the written conditions - this is where having written them down pays
- Collect keys and access cards, and revoke building access
- Reallocate the space, and update the record so it does not sit "allocated" to a company that has gone
The access revocation is the one most often forgotten, and it is the one a security review will find.
Record it honestly
Mark the startup closed, with a date and a reason. Keep everything else.
Do not delete the record. Three reasons: your outcome measurement depends on knowing what happened to every company that entered, an audit that finds disbursement to a company with no record is a serious problem, and survival rates measured against graduates rather than entrants are the most common way incubator numbers become meaningless.
Useful reason categories, kept consistent so they can be counted:
- Founders returned to employment
- Co-founder split
- Could not find a market
- Ran out of money
- Pivoted out of scope
- Acquired or merged
- Unknown, no contact
That last one is legitimate and worth having. Some companies simply stop responding, and recording "unknown" honestly is better than guessing.
What to tell the rest of the cohort
Say something, briefly and without detail. A cohort notices an empty desk, and silence produces speculation that is usually less kind than the truth.
"X has decided not to continue with the programme, we wish them well" is enough. What matters more is what it signals: that stopping is a normal outcome, not a disgrace. Cohorts where failure is unspeakable are cohorts where struggling founders hide it for months.
Do not treat it as a selection failure by default
Some failures are selection failures and it is worth knowing which. But a programme with no failures has almost certainly selected for safety - companies that were going to be fine anyway - and that is a worse problem, because it means the programme is not doing anything.
The useful review question is not "did any fail" but: did we see it coming, and did the programme have anything to offer at the point it started going wrong? Sometimes the honest answer is that the company needed capital and you offer mentoring, or needed a technical co-founder and you had none in the network. Those are findings about the programme rather than about the founder.
A short close-out checklist
- Conversation held, outcome agreed - paused, pivoting or stopped
- Undisbursed commitment formally released back to source
- Disbursed funds accounted for, with milestone evidence
- Recovery position checked against the sanction terms
- Space vacated, deposit settled, access revoked
- Record marked closed with date and reason, not deleted
- Noted in the next utilisation report
- Cohort told, briefly
- Reviewed at cohort end: did we see it coming?
Frequently asked questions
- How do you know a startup has stopped rather than gone quiet?
- Missed updates are the signal, not the proof. Two missed reporting cycles, no response to direct contact, and an empty desk together mean it is time for a conversation. Founders rarely announce that they have stopped; they simply stop replying, because the conversation is embarrassing.
- What happens to undisbursed grant money?
- It should be formally released back to the source so it can be reallocated. The common failure is leaving it committed indefinitely, which quietly reduces what the programme can commit to anyone else while the totals still look correct.
- Should a failed startup be removed from your records?
- No. Mark it closed with a date and a reason and keep everything. Deleting it breaks your own outcome measurement, and an audit that finds a disbursement to a company with no record is a much bigger problem than a company that did not work.
- Does a failure count against the programme?
- Only if you hide it. A cohort with no failures usually means the programme selected for safety rather than ambition. What funders judge badly is not failure but a programme that cannot say how many companies stopped, or when.
