Building a Mentor Network That Does Not Quietly Collapse
Recruiting forty mentors is easy. Keeping eight of them active is the hard part. How to structure a mentor network so it survives its first year.
Forty names, three people
Almost every incubator mentor network follows the same arc. Launch with an impressive list. Announce it. Six months later, three mentors have done nearly every session, a dozen have done one each, and the rest have never been contacted at all.
The list is still on the website. It is still accurate, in the sense that all forty agreed. It just describes something that does not happen.
The failure is not recruitment. It is that sign-up is visible and engagement is not, so nobody notices the gap until the three reliable people start declining.
Recruit for a defined ask
The reason most mentor commitments evaporate is that nobody ever said what the commitment was. "Would you be willing to mentor some startups?" is easy to say yes to and impossible to fail at, which means it is also impossible to succeed at.
Define the ask before you approach anyone:
- How many sessions, over what period. "Two one-hour sessions a month for six months" is a commitment somebody can check against their calendar and decline honestly.
- What kind of session. A one-off diagnostic, a recurring relationship, or being on call for a specific question.
- What they are being asked to help with. Pricing, hiring, a technical review, going to market.
- Whether it is paid, and what the arrangement is.
A specific ask gets fewer yeses and far more delivery. That trade is worth taking every time.
Match on the problem, not the sector
Sector matching is intuitive and mostly wrong. It produces conversations where both people know the industry and neither has done the thing the founder is stuck on.
What founders actually get stuck on is narrower and more portable: pricing, a first enterprise sale, a co-founder disagreement, a hiring decision, whether to rebuild. Somebody who has priced a product before is useful to a founder pricing a product, whatever the sector. Somebody who "knows fintech" is useful only if the problem is fintech-specific, and it usually is not.
So capture two things about each mentor: what they have actually done, and the problems they are willing to be asked about. Those are different from a job title and far more useful for matching.
The measure that keeps it honest
The only number worth reporting about a mentor network:
How many mentors delivered at least one session in the last quarter, and how many startups received at least one?
Not roster size. Not sign-ups. Sessions delivered, and coverage.
Most incubators, on measuring this the first time, find that a small minority carry it. That is not a scandal - it is normal, and it is fixable, but only if you can see it. The specific risks it warns you about are:
- Burnout of the reliable few, who will eventually stop
- Startups with no mentor at all, which is invisible from the mentor side
- A public roster that overstates what you provide, which becomes a credibility problem the first time an applicant checks
Payment models, and when each fits
There is no single right answer, but there is a wrong one: leaving it undefined.
| Model | Works when | Watch out for |
|---|---|---|
| Unpaid, goodwill | Occasional office hours, alumni giving back | Reliability drops once novelty fades |
| Honorarium per session | Scheduled commitment over months | Set it before recruiting, not after |
| Retainer | A mentor attached to a cohort throughout | Underuse - you pay whether or not sessions happen |
| Equity or success-linked | Deep, long engagements | Complicated, slow, and rarely worth it below a real cheque |
| Subsidised by the incubator | You want founders to value it without bearing full cost | Someone has to administer the subsidy |
The pattern that fails most often is starting unpaid and introducing payment later. Early mentors feel their goodwill is being reclassified, and it is a difficult conversation. Decide first.
Make scheduling somebody's problem other than the founder's
A surprising amount of mentor-network failure is calendar friction. A founder emails a mentor, waits four days, gets a slot that clashes, and the exchange dies. Neither party is at fault and the session never happens.
Anything that removes the negotiation helps disproportionately: published availability, a booking step rather than an email thread, and a confirmation both sides can see. The specific gain is not convenience - it is that a session which needs five emails to arrange usually does not get arranged.
Record that a session happened
Sessions that leave no trace cannot be reported, cannot be followed up, and cannot be counted when a funder asks what support you provided.
The minimum worth recording per session: who, with whom, when, roughly what about, and whether it happened. That last field is the one people leave out and the one that matters, because a booked session and a delivered session are different things and only one of them counts.
Do not over-engineer this into a form nobody fills in. Five fields, filled in most of the time, beats fifteen fields filled in never.
Attach mentors to programmes, not to everything
A mentor helping one cohort does not need to see your funding records, your other programmes or your full startup directory. Beyond the obvious privacy point, over-broad access creates a practical problem: nobody remembers to remove people, and two years later your access list is a history of everyone who ever helped.
The right shape is a mentor attached to a specific programme, seeing only what that requires, with the attachment ending when the programme does.
A short diagnostic
Answer these about your own network:
- How many mentors delivered a session in the last three months?
- How many startups in the current cohort have had at least one session?
- What is the ask you made when recruiting - can you state it in one sentence?
- Is payment defined and written down?
- Can a founder book a slot without negotiating by email?
- If a mentor left today, would their access end automatically?
Most incubators can answer the first two only by asking around. That is the finding.
Frequently asked questions
- How many mentors does an incubator actually need?
- Far fewer than most recruit. Ten to fifteen genuinely active mentors serve a cohort of twenty better than a list of sixty who mostly do not respond. A long roster is a marketing asset and an operational liability, because founders waste weeks discovering who is real.
- Should mentors be paid?
- It depends what you are asking for. Occasional office hours work unpaid. Structured, scheduled commitment over months usually does not, and a small honorarium changes reliability more than the amount would suggest. Decide before you recruit, because changing it later is awkward.
- What is the most common reason mentor networks fail?
- Nobody measures who is actually delivering sessions. Sign-up is visible and engagement is not, so the roster grows while the number of real conversations falls. By the time it is noticed, the same three people have done everything and are tired of it.
- How should mentors be matched to startups?
- By the specific problem, not by sector. A founder stuck on pricing needs someone who has priced something, whatever the industry. Sector matching feels right and produces a lot of pleasant, useless conversations.
