Buy when the work has outgrown the method, not when headcount hits a number. The reliable triggers are operational: payroll running across multiple states, statutory filings assembled by hand, an audit question you cannot answer from records, managers messaging HR for leave balances, or an exit where system access stayed live. There is no headcount threshold.
Because it measures the wrong variable. A single-site team on identical contracts with one leave policy generates far less administrative load than a smaller group spread across locations, entities and shift patterns. Complexity, not size, is what breaks a manual method: the number of policies, the number of exceptions those policies produce, the number of jurisdictions involved and the number of approvals each transaction needs. Two organisations with the same employee count can sit a long way apart on that curve. Any published headcount threshold also ignores your tolerance for risk and the seniority of the person doing the work by hand. Watch the operational signals instead, since they describe your situation rather than an average of somebody else's.
Six are worth writing on a wall. Payroll now runs across more than one state or legal entity. Statutory filings are assembled by copying between sheets. Somebody asked a question about a past period and the answer took days to construct. Managers ping HR for leave balances instead of looking them up. An employee left and something they had access to stayed open. And the cycle now depends on one person being available, so their leave becomes an organisational risk. Any single one is worth a conversation. Two or more appearing in the same quarter means the manual method has already failed and you are simply absorbing the cost of it in people's evenings.
The number of distinct rule sets you must track multiplies, and they do not move in step. Registrations, contribution schemes, professional tax treatment, filing calendars and record-keeping duties all differ by state, and each carries its own portal, its own format and its own timing. A manual process copes with one set through familiarity; it fails with several because nobody memorises them all and the differences are exactly where errors hide. Because the specifics vary by state and are revised from time to time, confirm the current position with a qualified advisor or the relevant state authority before building any of it into a process. This is usually where consolidated [payroll administration](/payroll-management) stops being optional.
That you have documents rather than records. The question is rarely exotic - how many people were employed on a given date, what this person was paid in a past period, when this policy changed and who approved it. If answering means reconstructing the past from emails and file versions, you do not have a history, you have artefacts. That gap matters beyond audits: it shows up in due diligence, in disputes, and any time leadership asks for a trend rather than a snapshot. Systems earn their keep here by recording changes as dated events rather than overwriting a cell, so the past stays answerable without anyone needing to remember it.
Every one of those messages is a transaction that bypassed the record. A manager who asks HR whether someone has leave available is not being difficult; they are working around a system that will not tell them. The cost is doubled: HR time answering, and a decision made on a number nobody verified. Worse, once approvals routinely happen in chat, the record stops reflecting reality and the balance you eventually settle at exit becomes a matter of opinion. Putting balances and requests where managers and employees can see them directly - through [leave management](/leave-management-software) rather than an inbox - removes the chasing and, more importantly, keeps the underlying data honest.
That offboarding is a memory rather than a process. When an account, a device, a shared drive or a building pass stays live after someone leaves, the cause is almost never negligence - it is that the leaving event existed in one person's head and never triggered anything. This is the signal most likely to be dismissed as a one-off, and the one with the largest downside, since the exposure sits with data rather than with money. It also tends to travel with a second failure: final settlement calculated from incomplete leave and attendance figures. If an exit produced both, the record and the process are already too loose to be run by hand.
Document the process you actually run, not the one you would describe in a meeting. Walk a full cycle and write down each task, who does it, what it depends on and where the data comes from. Note every exception you handle, since exceptions are what configuration has to absorb. Clean the employee data while there is no deadline attached, because the same work is far harder once a project plan is counting on it. Then match that documented reality against candidate [HR systems](/hr-software) rather than against a feature list. Buyers who skip this step end up configuring a vendor's default process and discovering the mismatch during the first live cycle.
Pitch N Hire is an applicant tracking system built for recruiters and hiring teams. If this answer described something you want to run properly, the ATS is where it lives.
Free for 1 user · No credit card · Talk to a real hiring expert
Get a personalized walkthrough of Pitch N Hire on your own roles and workflow. No slides, no obligation.
Prefer to talk? Book a demo · Talk to sales · View pricing
Free 1-user plan · No credit card · Talk to a real hiring expert
See your true cost-per-hire and how much Pitch N Hire could save you — our free Recruitment ROI Calculator gives you the numbers in under a minute. No signup required.
Open the free ROI calculatorPrefer a tailored walkthrough on your real roles? Drop your work email:
★ Free 1-user plan · No spam · Talk to a real hiring expert
