Best payroll software for 20 employees in New Zealand

Compare seven New Zealand payroll providers

Who this shortlist is for

Our brief is 20 New Zealand employees, weekly pay, some variable hours, and a named payroll operator. This is an editorial shortlist based on published pricing and operational documentation checked on 2 October 2026. It is not a hands-on accuracy ranking.

For this buyer, the important split is between a monthly employee charge and a charge each time employees are paid. The other split is responsibility: software, payment handling and full payroll processing are different purchases. Decide whether you want to run payroll yourself, outsource payments or outsource the whole pay run.

1. PayHero: weekly payroll with timesheets

PayHero Super includes approvals, imports and unlimited employee capacity at NZD 39 plus NZD 6 per employee monthly, before GST. Its lower Origin plan stops at ten employees, so it is not the appropriate headline price for this brief. Universe adds features including API access and an audit log.

Our verdict: start with Super for weekly payroll run by your own staff. Its employee fee is monthly, so a fifth weekly pay run does not multiply that fee.

Check the billing population: an employee paid or recording timesheets during the month counts. A seasonal worker with timesheets but no payment is not automatically inactive. If rostering is required, price the separate Droppah addition.

Check that the quoted edition records the corrections and approvals your payroll operator needs. If an audit log is a requirement, compare the relevant edition rather than assuming every plan includes it.

2. Smartly: software or a managed payroll service

Smartly’s DIY plan charges NZD 2.60 per employee per pay run, plus NZD 51.50 monthly, excluding GST. Its separate managed service charges NZD 6.20 per employee per pay run with a NZD 365 monthly minimum. Bank fees may apply. People Management is another separately priced product.

Our verdict: consider Smartly if you may want a provider to process payroll. DIY still leaves the employer entering and checking data; the managed route changes who processes it.

Ask for the input deadline, approver, correction process and substitute operator in the proposal. The annual cost example below covers DIY payroll only.

3. Crystal: choose how much payroll work to outsource

Crystal describes four service levels: DIY, PAYE-only, Basic Bureau handling payments, and Full Bureau. You can therefore ask for help with tax or payments without outsourcing the whole pay run.

The pricing page lists Crystal Timesheets, Audit Trail and Timesheets Correction as additions. Its advertised small-employer starting price covers up to five people, so it is not a 20-person quote. Setup, training and historical-data work also need scope.

Our verdict: ask Crystal to price the same 20-person file at the two service levels you would consider. Agree who checks the submitted hours and pay rates, fixes errors and approves corrections. Payment handling alone is not a promise that someone has checked every input.

If historical migration is important, list the records to be imported and the totals to be checked in the proposal; Crystal says transfer options depend on the previous platform.

4. iPayroll: decide who will make PAYE payments

iPayroll’s Standard package lists net-pay processing, KiwiSaver, filing, leave, an employee app and support. Its PAYE intermediary service collects PAYE each run and manages payments where that service is selected. The DIY option leaves the employer making PAYE payments.

Our verdict: consider it if you want help making PAYE payments. Confirm who investigates a failed or incorrect payment under the quoted service.

The public calculator starts with a per-pay-run figure. We did not verify a 20-person calculation, so that starting figure is not used as a monthly quote here. Request the applicable frequency, headcount, setup and bank charges in writing.

A partner introduction for outsourced processing is separate from buying the standard software service.

5. PaySauce: promising seasonal fit; unresolved public price

PaySauce says it charges for unique employees paid during a month, allows unlimited pay runs and charges only for months when payroll is run. Standard includes mobile timesheets, banking and PAYE functionality; Premium adds functions including rostering and custom job costing.

Our verdict: move it up for seasonal staffing when that charging model fits. We would still require a dated written price: the page’s main rate cards, displayed calculations and older embedded figures do not agree. We therefore publish no 20-person total here.

Make the proposal settle base fee, employee definition, inactive months, optional modules and GST before comparing it with the cost table.

6. Xero: the cost of adding payroll to an existing subscription

Xero NZ Ultimate’s regular price is NZD 135 monthly before GST, including payroll for ten people; additional payroll employees cost NZD 2 each monthly. Twenty therefore means NZD 155 monthly for that accounting-and-payroll subscription. A business already paying for the same Ultimate plan faces a different decision: the extra ten payroll employees add NZD 20 monthly.

Our verdict: move Xero near the top if the current accounting workflow works, payroll requirements fit and the people operating it want to keep it.

Compare the full accounting-and-payroll subscription separately from the cost of adding employees to an existing subscription. The extra payroll charge is not the total cost for a new customer. Retain the accounting baseline in both proposals and price only the changes you actually need.

7. Employment Hero: payroll as part of a wider HR purchase

Employment Hero lists payroll separately from HR and time-and-attendance additions, with contractual user minimums. Compare it if you also plan to replace HR or time tracking. Those modules are priced separately from payroll.

Historic gross earnings must be imported before parallel payroll so relevant leave calculations have the necessary history. The documented cutover sequence imports after the second-to-last old-system pay run, leaving the final run for parallel processing.

Regional anniversary holidays also depend on assigning the appropriate location; having a holiday calendar does not establish that the relevant pay rules have been activated and assigned.

Our verdict: consider it when you also need new HR or attendance software. Ask the migration supplier to import historic earnings and demonstrate a correct leave-pay calculation. Include any required HR and attendance modules in the total setup cost.

Annual subscription examples for 20 employees

Assume all 20 employees are chargeable throughout the year, are paid in every scheduled weekly run, and there are exactly 52 runs. All figures are NZD before GST, using regular published rates checked on 2 October 2026. These are subscription illustrations, not complete project quotes.

Sources: PayHero, Smartly and Xero NZ. The last two rows describe different buying situations, not different prices for the same new purchase.

Smartly’s four-run month is NZD 259.50; a five-run month is NZD 311.50. Multiplying a four-run bill by twelve misses four weekly runs. Check the actual payroll calendar, extra runs and which employees are included before budgeting.

The examples exclude migration, training, retained software, payroll labour and any separately applicable fees. Smartly identifies possible bank charges. Crystal and iPayroll require a scoped quote here; PaySauce’s inconsistent public figures remain unresolved; Employment Hero needs the contracted module and minimum-user schedule.

Test leave pay using imported earnings

Employment New Zealand explains that an entitled annual-holiday week is generally paid at the higher of ordinary weekly pay and average weekly earnings. The timing, gross-earnings inputs and individual circumstances matter. Its calculation guidance also distinguishes payments included in gross earnings from excluded payments. A field named “gross” is not enough to establish the correct input.

This is why a successful name, bank-account and leave-balance import does not settle payroll acceptance. The replacement needs the appropriate historical earnings as well. Employment Hero’s AWE import instructions use a pay-frequency-specific template and historical periods. Other vendors need their own approved import process; do not paste one vendor’s spreadsheet into another product.

Have the payroll specialist approve the expected result for each fictional case before the demonstration. Special circumstances—including leave in advance, parental leave and unpaid leave—need their own cases where relevant. The simple example below does not decide those cases.

Four checks before accepting the payroll migration

Use fictional records in a demonstration environment. No bank payment or statutory filing is needed. Retain the source file, mapping, imported result, expected result, discrepancy and named owner for each case.

A vendor can pass with a documented manual control if the actual operator demonstrates and accepts it. The test is not a demand that every product block every error automatically. The buyer needs a process that detects the missing period, preserves the right history and gives someone responsibility for correcting it.

Which suppliers should you test?

For the stated weekly-pay employer, demonstrate PayHero and Smartly against the same history file, then ask Crystal to price the precise work you would hand over. If Xero already runs reliably, put it through the same tests before commissioning a replacement.

Agree the price and who will run each payroll task. Before accepting the migration, require the supplier to pass all four tests and show how imported earnings produce the leave-pay result.

Questions buyers ask