Diligence was a test you sat once. Ownership is a reporting obligation that repeats monthly for the length of the hold.
Most teams underestimate that shift. The data room got assembled by hand, over six weeks, by people working evenings, and it passed. Then the first board pack is due in three weeks, and the same manual assembly has to happen again, and again after that, with a sponsor who now compares each month to the last one.
The first 100 days decide whether that becomes a system or a permanent fire drill.
Day 1 to 30: agree the definitions before you fix anything
The instinct is to buy a better system. Resist it for a month. A new HRIS configured against undefined terms produces clean, fast, contested numbers.
Agree what a head is. Contractors, part-time staff, people on notice, unfilled accepted offers, employees of an acquired entity not yet integrated. Every one of those is a judgement call, and the sponsor, the CFO and the HR lead frequently hold three different answers. Write the definition down and get it countersigned. Nearly every "the numbers don't tie" conversation in year one traces back to this being skipped.
Name the system of record. One system answers each question. Not the HRIS for some fields and a spreadsheet for the rest, and never a system whose data is maintained by whoever needed it last.
Read the diligence findings as a work list. Whatever the buy-side flagged about people data has not gone away because the deal closed. It is now your open item, and it will be raised again at exit.
Establish the reporting calendar. When the pack is due, what is in it, who signs it. Set this before month one closes, because the first late pack sets the expectation for every one after it.
Day 31 to 60: fix the mechanics that produce the numbers
Now the systems work.
Permissions. Founder-era access models are almost always too broad, and post-close is when that becomes visible. New investors, new board members, new reporting lines, and often a new finance team who should see cost but not individual compensation. Rebuild the model rather than patching it. The permission audit is the method.
Payroll controls. The single most common post-close finding we are asked to remediate is payroll approved by the person who prepared it. Separate preparation from approval, name both roles, and produce the audit trail automatically rather than on request.
Reconciliation between HR and finance. Headcount in the HRIS, payroll cost in the ledger, and the board pack should agree without anyone rekeying. If they do not reconcile now, they will not reconcile under time pressure at exit.
Close the manual spreadsheet. Any number in the board pack assembled by hand is a number that breaks when the person who assembles it goes on holiday.
Day 61 to 100: make it repeatable without you
The test is not whether this month's pack is right. It is whether next month's is right when nobody heroic is available.
- Every recurring task has a named owner and a date, not a habit.
- The monthly pack generates from the system, with commentary added rather than figures assembled.
- Someone other than the original builder can run the cycle end to end.
- The definitions agreed in month one are documented where a new hire will find them.
That last point matters more than it sounds. People ops leadership turns over during a hold, and undocumented definitions leave with whoever held them.
What the sponsor will actually ask for
Four artifacts, repeatedly, for the length of the hold:
- Headcount by department and location, current and trailing, on the agreed definition.
- Fully loaded people cost against plan.
- Attrition, with leavers classified consistently.
- Open roles against approved plan, with start dates.
If producing those takes more than an afternoon, the problem is the system rather than the analyst.
Three things that go wrong
Buying the platform in week two. The system gets selected against requirements nobody has written, by people still learning what the sponsor wants. Wait until the definitions and the reporting calendar exist. Then the selection is straightforward, and the demo questions do most of the work.
Treating it as an HR project. People cost is usually the largest line in the P&L at this size. Finance owns the number, HR owns the record, and the reconciliation between them is the deliverable. Run it as a joint project or run it twice.
Assuming the exit is far away. The reporting spine you build now is the same one a buyer tests in three to five years. Building it properly once is cheaper than assembling a second data room by hand.
The 100 days on one page
| Window | The objective | The output |
|---|---|---|
| Day 1 to 30 | Agree definitions and ownership | Signed headcount definition, named system of record, reporting calendar |
| Day 31 to 60 | Fix the mechanics | Rebuilt permissions, separated payroll approval, HR to finance reconciliation |
| Day 61 to 100 | Make it repeatable | Named owners, system-generated pack, documented definitions |
We run this as a fixed-scope engagement for VC and PE-backed companies between 25 and 250 employees, either as the implementation itself or as the operating layer afterwards. If a pack is due and the numbers do not tie yet, book a 20-minute call and we will tell you which of the three windows you are actually in.
Download this guide
Take the whole thing with you.
The whole guide as one PDF you can send to your team. We email it over.
One email, with the PDF attached. Nothing else.
Common questions
Should we change HRIS in the first 100 days?
Rarely in the first month. Agree the headcount definition, the system of record and the reporting calendar first, then select against written requirements. A platform chosen before those exist gets configured twice.
Who owns people reporting after close, finance or HR?
Both, on different halves. Finance owns the number that reaches the board, HR owns the record it comes from, and the reconciliation between them is the actual deliverable. Assigning it to one function alone is how the figures diverge.
What if the portfolio company has no HR lead?
Common at this size, and it changes the sequence rather than the work. The recurring cycle needs a named owner from day one even when that owner is fractional or external, because the failure mode is a system with nobody running it.