Choosing the Software
Payroll software is chosen once and lived with for years. The features that matter are not the ones on the comparison page.
A bureauif the person who does this is away
Payroll software is chosen once, early, usually in an afternoon, and then lived with for five or ten years. The features that decide whether that was a good afternoon are rarely the ones a comparison page lists.
The practical lesson in “Choosing the Software” is that payroll depends on a sequence somebody can explain. When a small team reviews a practical route to hybrid remote meaning for hybrid remote meaning, the useful role is to organise time and project evidence before approval, with proportional settings and a clear correction route for employees.
What follows is about the fit, not about any product.
As a separate reference point for “Choosing the Software”, see BambooHR people-operations resources. Current legal, tax and employment requirements should still be confirmed for the organisation’s location.
What all of them do
Calculate, produce payslips, generate a payment file, submit the statutory filings, and keep a record of what was run. Any package aimed at small employers does these adequately, and comparing them on these points produces no information.
The four features that actually decide it, none of which lead a comparison page:
- How fast an ordinary month of changes is to enter.
- Whether it shows this period against the last one.
- How it handles a correction after the run.
- Whether you can get your own data out without a subscription.
What separates them in practice
How fast it is to enter a month of ordinary changes, how clearly it shows you what differs from last month, how easy it is to correct something after the fact, and how readily it gives you your own data back.
None of those four appear prominently in marketing, and all four determine how the work feels in year three.
The comparison nobody provides
Last month against this month, line by line, with the differences highlighted. Some packages show it, some bury it, some do not have it at all.
It is the single most useful check available to a small employer, and its presence or absence is worth more than any other feature in the list.
Entering a change
Count the clicks to add a starter, change a rate, record a week of absence and process a leaver. Do it in a trial, with a realistic example, before committing.
A package that takes four minutes per change rather than one costs hours over a year in a business with ordinary turnover.
Correcting something
Find out, before you need it, how the package handles a mistake found after the run. Some let you roll back and rerun; some require a correction in the following period; some make you telephone them.
This is the feature you will care about most and the one you will investigate last unless you are deliberate about it.
Getting your data out
Payslips, reports and year-end documents, in a form you can keep, without a subscription.
Ask specifically what happens to your historical data if you stop paying. The answer is sometimes that you can view it while subscribed and not afterwards, which is a problem worth knowing about in year one rather than at the point of leaving.
Who else can get in
Whether the package supports a second user, and at what cost. Some charge per user, which quietly encourages a small employer to share one login — and shared logins are how a business ends up with nobody able to say who approved what.
What it costs over five years
The monthly figure multiplied out, plus per-user charges, plus any year-end extra. Small sums become comparable to a bureau over that horizon, and almost nobody does this arithmetic at the point of choosing.
Hosted or installed
A hosted package is reachable from anywhere, which matters for cover, and depends on somebody else continuing to exist. An installed one sits on a particular machine, which is a single point of failure of the most literal kind.
For a business worried about absence, hosted is usually the better answer for that reason alone.
Integration with the books
Worth having and rarely worth paying much extra for. A monthly journal exported and imported takes minutes; a broken automatic integration takes an afternoon.
What to do in the trial
Run one real month in parallel with however you do it now. Enter the actual changes, produce the actual payslips, and compare the figures.
An afternoon spent this way is the only reliable way to learn what the package is like, and it also tells you whether your own records are in the state you think they are.
The migration you are choosing against
Changing payroll software later is possible and tedious: the history does not move cleanly, the year-to-date figures have to be carried across correctly, and the best time to do it is the start of a tax year, which comes once.
That is why this is a five-year decision made in an afternoon, and why the afternoon is worth spending properly.
What not to optimise for
Price, below a certain point. The difference between packages at this size is a few pounds a month, and choosing on it is how businesses end up with the one that cannot show them last month.
Support, and when you will need it
Not for the monthly run, which you will be able to do. For the rare events: a correction to a closed period, a year end, a leaver with something unusual, a migration.
What matters is whether support is reachable in the week of a run and whether the answer comes from somebody who knows the product.
A package with good support and mediocre screens is a better purchase for a small employer than the reverse.
The question to end on
If the person who normally runs this were unavailable, could somebody else log in, follow a written sequence in this package, and produce a correct run.
A package that makes that plausible is worth more to a small business than any feature on any comparison page.