How to run the selection
A structured process beats a feature comparison, every time. The aim is evidence rather than impressions, and being able to show how you reached the decision.
From longlist to shortlist
Build a longlist of eight to fifteen products from vendor research, peer recommendations, professional networks and your own market scan, then filter against your must-haves to a shortlist of three to five. More than five and you cannot give any of them proper attention. Fewer than three and you have no real comparison.
The RFP
A written request gives you comparable answers you can hold a vendor to later. Build it from your requirements list rather than a generic template, and ask for specifics rather than yes or no. "Describe how your system handles a mid-year move from Scotland to England" tells you far more than "do you support Scottish income tax". A payroll RFP template is a useful starting structure, and our RFP template includes more than 250 questions to cut down from.
Scripted demos: make every vendor run your scenarios
This is the single most valuable thing you can do, and almost nobody does it. A standard vendor demo shows a clean payroll with tidy data, chosen to show the product at its best. It tells you very little.
Instead, write a demo script, send it to every shortlisted vendor in advance, and use the same script for all of them. Include your awkward cases, not your easy ones:
- A mid-month starter on a non-standard pay element, with a backdated correction the following period.
- An employee moving from Scotland to England mid-year, showing the year to date position after the move.
- A variable-hours worker's payslip, showing hours split by rate.
- Someone with two court order deductions where earnings are too low to satisfy both.
- A pay award backdated three months across a closed period.
- The pension contribution file as your provider actually requires it.
- A pay run submitted and then found to be wrong.
- The month end journal, in the shape finance needs.
Ask to see each one performed in the system, not described, and note where a vendor has to come back to you. In our experience, the buyers who learn most from a demo are the ones who supply their own awkward scenarios in advance, because it moves the conversation from what the product can do in principle to what it does with your payroll.
Where a trial or sandbox is available, take it, and have the person who will run payroll use it rather than the person leading the project.
Reference calls and what to ask
Ask every shortlisted vendor for references from organisations of similar size and complexity, and speak to at least three: one recently implemented, one who has been with them longer.
What to ask: what did implementation involve from your side, and how did it compare to what you were told. What took longer than expected. How quickly are support queries resolved, and does the person answering understand UK payroll. How did the last April change go. What do you wish you had asked before signing.
Migration, and what to ask about it now
Migration belongs in the selection because it is where projects go wrong, and because the answers differ by vendor.
What to ask: what does the vendor do and what does your team do. How is year to date data brought across and validated. How many parallel runs do they recommend, and what does your team do during them. Can you go live mid-year, or is April the only sensible cutover. What happens to historical records in the old system. In our experience, the most common cause of overruns is incomplete or undocumented data in the outgoing system, worth investigating before you commit to a timetable rather than after. The detail of running the project sits in payroll implementation.
Scoring and making the decision
Score against your requirements, not your impressions, in a way that survives being questioned.
- Weight each requirement band before scoring anything: must-haves heaviest, should-haves roughly half that, nice-to-haves light enough to break a tie but not decide an outcome.
- Score each product against each requirement on a short scale: met as standard, met with configuration, met with custom work, not met. The configuration and custom work distinction matters, because custom work costs money and makes upgrades harder.
- Score the vendor separately from the product. Support, UK payroll expertise, implementation approach, roadmap and financial stability belong in their own section.
- Add total cost as a separate line, using the three to five year figure, so a cheap product cannot quietly win on price alone.
- Record why, not just the number. The written rationale is what you need when someone asks in two years why this product was chosen.
Settle the commercial terms before signature: contract length and notice period, how the annual uplift is calculated, what support includes and what is chargeable, what happens if headcount moves significantly either way, and what data you get back on exit and in what format.