The 6 HR reports a founder should get every month without asking

Ask a founder how many people work in the business and you will often get a pause, then a number, then a qualification about whether that includes the contract staff at the second site. Ask what attrition looks like and you get an impression. Neither is carelessness — the numbers exist, but they are assembled on demand, by hand, by whoever is asked.

The fix is not a dashboard project. It is six numbers, defined once, produced on the same date every month whether anybody asks or not.

Define before you measure

Every one of these reports is useless if the definition moves. Settle each one in writing before the first report, because the most common failure is not a missing number — it is two people quoting the same number computed differently.

1. Headcount and movement

Opening headcount, joiners, exits, closing headcount, split by department and by employment type. The arithmetic must reconcile: opening plus joiners minus exits equals closing. If it does not, stop and find out why, because every other report downstream is built on this one.

• Define whether contract, trainee and probationary staff are included, and keep the answer the same every month.
• Split by department so the movement is visible, not just the total.
• Show vacancies alongside, because headcount without open positions is half a picture.

2. Employee cost

Total employee cost for the month, split by department, with the previous month beside it. Include what actually leaves the business — salaries, employer contributions, overtime, incentives and contract manpower — because a cost report that excludes the overtime bill is the one that surprises you at year end.

The useful cut is cost per department against headcount per department. That is where a department that has quietly grown its overtime or its contract spend becomes visible without anybody having to raise it.

3. Attendance and absence

Absenteeism rate, late arrivals, overtime hours, leave taken and leave balance outstanding. For shift businesses add coverage against plan, because that is the number that predicts whether next month's output is achievable.

Leave balance outstanding deserves particular attention: it is a liability that accumulates quietly, and most businesses first look at it when somebody leaves and the encashment calculation surprises them. It comes directly out of attendance, shift and leave discipline — if the balance is not trustworthy, neither is this report.

4. Hiring

Open positions, how long each has been open, offers made, offers accepted, and joiners against those acceptances. Four of those five are usually tracked somewhere. The fifth — acceptance to joining — is almost never tracked and is frequently where the problem is.

• Ageing of open positions matters more than the count. Three positions open for two months is a different problem from twelve opened last week.
• Acceptance-to-joining rate tells you whether you are losing people in the notice period, which is invisible in every other report.
• Time to hire, measured consistently — from requisition approval, not from when someone started looking.

5. Attrition

Exits in the month and rolling twelve months, by department and by tenure band, with a regretted and non-regretted split. The tenure cut is the one that earns its place: people leaving inside six months point at hiring or onboarding, people leaving at two to three years point at progression, and the two need completely different responses.

Resist the urge to add a reason chart in month one. Reasons drawn from exit forms are usually unreliable until the interview process has been fixed, and a confident chart built on "better opportunity" is worse than no chart.

6. Compliance and documentation status

What falls due this month and next, what is outstanding, and whose it is. Plus documentation completeness: how many employee files are missing something, and which.

One page, same shape, every month

MONTH: ______            PREPARED BY: ______   DATE: ______
-----------------------------------------------------------------
1 HEADCOUNT    opening ___ + joiners ___ - exits ___ = ___
               by department: ______________________________
               open positions: ___

2 COST         total ______  prev month ______  change ___%
               by department: ______________________________

3 ATTENDANCE   absenteeism ___%  late ___  overtime ___ hrs
               leave balance outstanding: ___ days

4 HIRING       open ___  oldest ___ days  offers ___
               accepted ___  joined ___  (accept-to-join ___%)

5 ATTRITION    this month ___  rolling 12m ___%
               regretted ___ / non-regretted ___
               by tenure: <6m ___  6-24m ___  >24m ___

6 COMPLIANCE   due this month ___  outstanding ___  owner ___
               employee files incomplete: ___

THREE THINGS NEEDING A DECISION:
  1. ______________________________________________
  2. ______________________________________________
  3. ______________________________________________

The last block is the one that makes the pack get read. Six tables with no ask is a document; six tables and three decisions is a meeting agenda.

What to do in the first three months

1. 1

   Month one — define and produce, badly

   Write the definitions, produce the pack from whatever data exists, and accept that two or three numbers will be wrong or unavailable. Mark them as such rather than leaving them out.

2. 2

   Month two — fix what month one exposed

   The gaps are always in the same places: headcount that does not reconcile, overtime not captured, leave balances nobody trusts. Fix the source, not the report.

3. 3

   Month three — hold the review

   Same date, same format, with the three decisions. By the third pack the trend line starts being more interesting than the level, which is when the exercise begins to pay.

Do not wait for the data to be clean before starting. The pack is what exposes which data is dirty, and a business that waits for perfect inputs produces its first report never.

Where this goes next

Once six numbers are reliable and produced monthly, two things follow on their own. Management starts asking better questions, because a trend invites a question in a way that an impression does not. And the effort of producing it by hand becomes obvious, which is usually when a business looks at getting the numbers out of a system instead — the point of HR reporting and analytics work, and of the reports and analytics module if you already hold the data in one place.

The order matters though. Define the numbers first and automate second. Automating an undefined number produces a wrong answer faster, and with more authority.

The two numbers founders most often miss

Two figures appear in this pack that rarely exist anywhere else in a growing business, and both tend to produce a reaction the first time they are seen.

The first is leave balance outstanding. It accrues quietly, it is a real liability, and most businesses meet it for the first time in a settlement calculation that is larger than anyone expected. Put it on the pack from month one, even if the underlying balances need correcting — the correction is the point.

The second is acceptance-to-joining rate. Every business tracks offers made. Almost none tracks how many of those acceptances actually walked in on day one. Where that rate is falling, the problem sits in the silent notice period rather than in recruitment, and no other report in the pack would have shown it.

Neither number requires a system to produce. Both require somebody to have decided they are worth producing, which is the whole of what HR reporting and analytics work does at the start — settle what each figure means, trace it to a source that can be checked, and then insist it appears on the same date every month.