Moving HR off spreadsheets without signing up for a subscription

Every organization runs HR on spreadsheets at the start, and it works. A sheet for staff details, another for leave, a third the finance person rebuilds each month for payroll. For fifteen people that is not a problem worth solving.

It stops working gradually, then all at once. Here is what actually breaks, in the order it usually breaks, and what to do about it.

What breaks first: nobody knows which file is current

The first failure is not data loss. It is ambiguity.

Someone downloads Staff List.xlsx to work on at home. Someone else edits the copy in shared storage. Two weeks later there are two versions with different phone numbers for the same person, and no way to tell which is right except asking them.

Cloud spreadsheets help, but only partly. They fix simultaneous editing; they do not fix the copy someone exported for a bank form in March and has been quietly updating since.

The underlying problem is that a spreadsheet has no idea what a record is. It is a grid. Nothing stops a row being duplicated, or a leaving date being typed into the wrong column, because nothing in the file knows those cells mean anything.

What breaks second: the audit question

At some point somebody external asks a question you cannot answer.

An auditor asks who approved a particular leave request and when. A former employee disputes a final payment. A grant application wants headcount by department at a specific date last year.

Spreadsheets do not keep history. Version history in Google Sheets or SharePoint tells you a cell changed; it does not tell you that Sunita approved Bikash's five days of unpaid leave on 12 March, which is what the question was actually about.

This is usually the moment HR software goes on the budget, because the cost of not being able to answer suddenly becomes concrete.

What breaks third: payroll takes a week

Payroll on spreadsheets is a monthly reconstruction. Someone opens the attendance sheet, cross-references the leave sheet, works out who was absent without approved leave, applies that to the salary sheet, recalculates tax, and produces payslips.

Every step is manual, which means every step is a chance to be wrong. And the errors are the expensive kind: an underpayment is a grievance, an overpayment is awkward to claw back, and a tax miscalculation is a problem with the revenue authority rather than with your staff.

The tell that you have crossed this threshold is simple. If one person's absence would delay payroll, payroll is no longer a process. It is a person.

What to fix first

You do not have to solve all of this at once, and you should not try. In rough order of return:

1. One employee record. A single authoritative place for each person's details, so there is no question which copy is current. This is the foundation everything else sits on.
2. Leave with an approval trail. Not a form that emails a manager — a request with a recorded decision, a decision-maker and a timestamp. This is what answers the audit question.
3. Attendance that employees enter themselves. Self-service removes the person who transcribes timesheets, and removes the transcription errors with them.
4. Payroll that reads the other three. Once leave and attendance are real records rather than sheets, the payroll run can derive unpaid days instead of someone working them out.

Notice that the sequence matters. Payroll automation is worthless if the attendance data feeding it is still typed in by hand from a paper sheet.

The subscription problem

Here is where most organizations get stuck, and it is a legitimate stuck.

The obvious next step is a hosted HR platform. Most charge per employee per month. At thirty people and a fairly typical eight dollars per employee, that is roughly $2,880 a year — for software replacing four spreadsheets.

That would be fine if it stopped there. It does not. The bill scales with the thing you are trying to grow. At a hundred people the same product costs $9,600 a year, and you have not asked it to do anything new.

There is a second cost that is easier to miss. Once your employee records, salary history and leave balances live inside a vendor's system, moving is genuinely difficult. An export is a spreadsheet, and you have just spent two years escaping spreadsheets. The switching cost is what lets the price rise.

The other option

Software you install on your own server and pay for once.

This is not exotic. It is how business software was sold for decades before subscriptions became the default, and it is still how a lot of infrastructure gets bought. You pay a licence fee, you run it on a machine you control, and the cost stops when you have paid it.

For HR specifically it has a property that matters more than the money: the data stays in a database you own. Not exported from a vendor on request — held by you, backed up by you, queryable by you. The audit question becomes something you can answer without asking permission.

The honest trade-off is that somebody has to run the server. If your organization has nobody who can maintain a Linux box and a PostgreSQL database, that is a real cost, and you should count it. It is usually a few hours a month rather than a role, and installation is something you can pay someone to do once. But it is not zero, and anyone who tells you it is zero is selling.

Where to start this week

If you do nothing else, do this: pick one authoritative employee list, put it somewhere with access control, and delete the other copies. Not archive — delete. As long as the old copies exist, someone will use one.

That single act solves the ambiguity problem, which is the one quietly corrupting everything downstream. Everything else can wait until you have decided what you are buying.

---

Ace HR is self-hosted HR software licensed for a one-time fee — employees, attendance, leave and payroll on your own infrastructure. See what a licence includes, or look at the screens.