Payroll software measurement mistakes England: audit guide

Payroll software measurement mistakes England reviewers may encounter can make an HMRC filing-success claim look stronger than its evidence. A buyer, reviewer or analyst checking an existing claim about PAYE submissions can trace a reported result through Real Time Information (RTI) records, distinguish an FPS from an EPS, and check how failures and later changes were handled. The aim is to decide what a particular claim proves, not to design a general payroll dashboard or compare products.

What an HMRC filing-success claim means

For employers operating PAYE in England, HM Revenue & Customs (HMRC) receives payroll information through RTI. The relevant legal framework includes the Income Tax (Pay As You Earn) Regulations 2003.

A Full Payment Submission (FPS) reports pay and deductions; an Employer Payment Summary (EPS) is used for other PAYE information, including adjustments where applicable. They are different submission types and should not be treated as interchangeable records.

A supplier's word "successful" needs examination.

A file being prepared or transmitted is not, by itself, evidence that HMRC accepted it. Equally, an HMRC response to a submission does not establish that every payroll calculation was correct or that employees received the right pay. HMRC's processing of a return is not an endorsement of the software that created it.

Keep the claim narrow enough to test. If a product statement combines HMRC filings with pension-provider files, employee payments or general payroll completion, separate those outcomes before judging the HMRC part. Where a supplier handles payroll on an employer's behalf, establish whose submission activity the claim describes; business-model differences can affect what the software itself is responsible for.

Capture the claim before testing it

Save the exact wording, chart or report, including its date and any explanation beside it. Record which product, service arrangement, reporting period and employer group it refers to, if disclosed.

A claim about "successful HMRC submissions" is not specific enough to check until the supplier explains what event it counts as successful.

Ask for the claim's unit in practical terms: is it counting individual submissions, employer-pay-periods, or something else? Ask which types of RTI submission are included, and whether the stated result covers only first attempts or also later submissions.

These questions expose whether a headline compresses unlike events into one label without requiring you to invent a replacement key metric.

Keep the original version of the claim if the supplier later updates it. Without that record, a revised result can be mistaken for the number originally presented, making it difficult to judge what evidence supported the claim at the time.

Trace a submission to its HMRC evidence

Use a small, authorised sample from the period covered by the claim. For each selected case, match the payroll system's event to the submission type, employer and pay period.

Then look for the corresponding HMRC response or other controlled evidence available to the people authorised to access it. Ask the supplier to explain how its own status labels map to those records; a label in a product log is not self-explanatory proof of the underlying outcome.

Include an ordinary case and an exception, such as a rejected submission or one that needed correction. Check whether the evidence supports the status shown in the claim and whether the record can be followed through any later action. Avoid requesting employee names or detailed pay data when anonymised or otherwise minimised records are sufficient for the test.

Timing matters to what a claim says. An FPS is normally reported on or before payday, subject to applicable exceptions.

If the claim refers to timely filing, compare its stated timing rule with the pay date and the submission evidence; a submission timestamp alone may not demonstrate that the rule used in the claim was met.

Check what the count leaves out

A headline can be misleading if it counts only records that reached a favourable end state while omitting relevant failed or incomplete attempts. Ask how retries, resubmissions and corrected records are treated.

Several attempts relating to one payroll period should not quietly appear to be several independent successes if the claim is presented as describing payroll periods. Attribution methods set out how incidents and corrections are credited.

Ask for the exclusions and any changes to them over time. Removing a difficult employer group, a submission type or records with missing responses can change the apparent result without any improvement in the service. If the supplier cannot explain a change in the population behind the claim, treat comparisons across the affected periods cautiously.

Check that FPS and EPS cases have not been blended without a clear reason. Their purposes differ, so a combined headline may conceal which kind of filing is driving the result.

Where the supplier's claim covers a mixture of direct employer filing and bureau-managed work, identify those arrangements rather than assuming the same workflow applies to each. The comparison should follow the claim's actual scope, not imply that the result represents every payroll service model.

Look for late responses and revised records

Ask what happens when an HMRC response arrives after the result was first reported, or when a later correction changes an earlier period.

Check whether the previous result is retained with a revision history, marked as provisional, or replaced without an explanation. A changing result is not automatically a measurement mistake; an unexplained change that obscures what readers saw before is a reason to qualify the claim.

Official statistics show the same pattern: the ONS user guide to PAYE RTI statistics explains that its earliest estimate imputes missing returns and is revised as real data replaces them.

If there is an example of a corrected case, trace it through the original submission, the subsequent action and the current record. Note whether the claim's published value changed and whether its explanation makes that change understandable. If the supplier cannot provide an example, ask it to describe the process and what evidence would show that it operated as stated.

Decide what the claim supports

Record a conclusion in one of three forms: the claim is supported for the scope stated, supported only with a qualification, or not reproducible from the evidence provided.

Name the evidence examined and the largest limitation, such as missing HMRC responses, unclear treatment of retries or a population that changed. Do not turn an unresolved gap into an assumption that the supplier's claim is false; state what could and could not be verified.

A reviewer should make that limitation visible wherever the claim is discussed. A disclosure policy can set how corrections are shown. A payroll software review methodology can help make the evidence trail and judgement inspectable. The useful result is not a more impressive success label, but a clear account of what the available RTI records establish, and what they do not.