Writing an Incubator Annual Report People Read
Who the report is really for, the numbers that belong in it, and how to write one that helps renewal rather than just recording the year.
Three audiences, one document
An incubator annual report is usually written as a record of the year. It is actually read by three people with different questions:
- The funder, deciding whether to renew. Question: is this a good use of money, and is this centre well run?
- The institution, justifying the centre internally. Question: what did we get, in terms my colleagues recognise?
- The prospective applicant, deciding whether to apply. Question: who went through this, and what happened to them?
You do not need three documents. You do need to know which parts serve which reader, because a report written for none of them - a chronological account of activities - satisfies nobody.
Open with two pages that stand alone
Most readers read the first two pages. Write them last, and write them to work in isolation:
- What the centre is for, in two sentences
- The year in numbers: applications, admissions, companies supported, funding disbursed, outcomes
- The two things that went best
- The one thing that did not, and what changed as a result
- What is planned next year
That fourth item is the one people leave out and the one evaluators notice. A report where everything went well reads as marketing; a report that names a problem and what was done about it reads as management.
The numbers that belong in it
Grouped so a reader can find what they came for:
Activity - applications received, admitted, and the ratio. Companies supported, split by engagement level. Sessions delivered and companies reached. Events held.
Operations - space allocated, occupied and collected. Funding committed, released and remaining. Mentors active, not merely enrolled.
Outcomes - survival at 12 and 24 months against the entering cohort. Stage progression. Companies that raised anything, with median and total. Jobs, with the definition stated. Closures, with reasons.
Consistency - each of the above alongside the previous two years, using the same definitions.
That last group is the one that most affects renewal. A single year's numbers require trust; three years of consistent numbers show a direction, and a direction is something an evaluator can defend internally.
Say what the definitions are
In an appendix, briefly: what counts as a company supported, how headcount is measured, whether survival is against entrants or graduates, what "active mentor" means.
This looks like pedantry and it does the opposite of what people fear. A report that states its definitions reads as more credible, not less, because the reader can tell the numbers mean something specific. It also protects you next year, when someone compares the two and the methodology has not changed.
Include failures, counted
A cohort where every company is still trading is either a very unusual year or incomplete tracking. Evaluators know this.
Report closures with categories - founders returned to employment, could not find a market, co-founder split, acquired, no contact. It takes three lines and it does two things: it demonstrates you track companies after they leave, and it makes your survival figure believable.
Use real examples, with permission
The part prospective applicants read.
Two or three companies, half a page each: what they were doing at entry, what happened during the programme, where they are now. Named, with permission, and specific.
What does not work is a montage of logos. A logo says a company existed; a paragraph says what the programme did. And the second is what a founder deciding whether to apply is actually looking for.
Make it assemblable
The reason annual reports are painful is the same reason board packs are painful: the data is reconstructed rather than maintained.
If applications keep their scores, disbursements keep their milestones and approvers, updates are collected on a schedule with a record of who did not answer, and closures are recorded with dates and reasons, then the numbers section of the annual report is a query and the writing is the only work left.
If none of that is true, the report is three weeks and the numbers are approximations that will not match next year's.
What to leave out
- A month-by-month chronology of events
- Photographs without captions saying who and why
- Aspirational language about ecosystems and innovation
- Any number you cannot say the definition of
- Total portfolio funding presented as if the centre caused it
That last one is worth resisting even though everyone does it. "Companies in our portfolio raised ₹X" is true and defensible. "We helped companies raise ₹X" is a causal claim, and an evaluator who spots one unsupportable claim reads the rest differently.
The test before you publish
Read the first two pages as though you were the person signing the renewal. Can they tell whether the centre is well run, whether it is improving, and what it will do next year?
If the answer requires reading page fourteen, the summary needs rewriting.
Frequently asked questions
- Who is an incubator annual report actually for?
- Usually three audiences at once: the funder deciding on renewal, the institution justifying the centre, and prospective applicants judging whether to apply. They want different things, which is why a single undifferentiated document satisfies none of them.
- How long should it be?
- Twenty pages is plenty, with a two-page summary that stands alone. The most-read part of any annual report is the first two pages and the pictures; everything else is reference for the small number who go looking.
- Should failures be included?
- Yes, counted and stated. A report with no failures reads as either selection for safety or incomplete tracking, and an evaluator will assume whichever is worse. Reporting closures with reasons is a credibility gain, not a loss.
- What makes a report useful for renewal?
- Consistency across years more than impressive numbers in one. An evaluator can act on a trend. A single large figure with no baseline and a methodology that changed is something they have to take on trust, and taking things on trust is what they are paid not to do.
