How we measure UK payment behaviour: our method

Adfin's payment figures come from its own records of invoices that UK businesses raise and collect through Adfin, measured directly from the due date and the date each payment arrived. They aren't a survey or a panel. This page explains what the figures in the UK Getting Paid Report and our articles cover, what counts as late, and why Adfin's published late-payment rates differ.

In this article

The short version

• The figures come from invoices raised through Adfin by UK businesses, mostly small businesses and accountancy practices billing modest recurring fees.
• Late means paid after the due date set on the invoice, compared as calendar dates.
• Direct debit, invoices customers pay themselves and payments recorded from outside Adfin are measured separately, never blended.
• The Getting Paid Report's 34.3% late rate covers only invoices customers paid themselves; the 8.7% and 9% rates cover every way customers paid, including direct debit.
• Rates are published as measured, sample sizes as rounded thresholds, and nothing is based on fewer than 200 payments.

What the figures cover

Each figure is built from payment requests: one invoice or fee a business raised through Adfin for one customer to pay. Every request carries its due date, the date it was paid and how it was paid, so nothing is estimated, weighted or self-reported.

The UK Getting Paid Report covers more than 450,000 paid invoices raised through Adfin by more than 1,000 UK businesses between June 2024 and August 2026. More than 55,000 of them were paid by the customer themselves, by card, Apple Pay, Google Pay, pay by bank or bank transfer, and those are the invoices behind every on-time and days-late figure. Direct debit figures cover more than 250,000 collections first attempted between July 2024 and July 2026.

A few findings use a narrower window, and the article quoting them says so. The comparison between invoices set to collect automatically and invoices paid on request covers more than 150,000 payment requests created between July 2025 and January 2026.

Test and internal accounts are left out, as are deleted and draft requests, and every invoice is in pounds. Voided invoices are left out of any figure about whether money was collected, since a cancelled invoice isn't a failed collection.

Keep the population in mind when you compare. Adfin's customers lean towards small accountancy practices, and the typical invoice a customer paid themselves was £180. If you send large invoices on long terms, the patterns may hold for you, but the levels probably won't.

Three groups of paid invoices, measured separately

Paid invoices fall into three groups, and each one measures something different, so adding them together would give a figure that describes none of them.

A direct debit is collected on a schedule you set and normally lands a few days after the due date because of the Bacs cycle, even when nothing has gone wrong. Measured against the due date, almost every direct debit would read as late, so it's measured on whether the money arrives instead: 97% of collections succeed first time and 98.7% are collected in the end. Those figures use collections first attempted at least four weeks earlier, so a collection still going through the cycle isn't counted as a failure.

Payments recorded against the invoice from outside Adfin are dated by the accounting system, which can lag the day the money arrived, so their timing isn't compared with the other two groups.

What counts as late

The reference point is the due date set on the invoice, not the invoice date or a statutory default. A payment on or before that date is on time, and one after it is late. Days late is the number of whole days between the two, compared as calendar dates, so a payment at 11pm on the due date counts as on time.

Lateness is only measured on invoices that were eventually paid, because an unpaid invoice has no payment date. Invoices that are never paid tend to be larger, so every late rate here is on the low side, more so for bigger invoices.

Payment method is taken from how the invoice was actually paid, not how it was offered. So "card invoices" means invoices that ended up paid by card. Where retries or part payments mean an invoice was paid more than once, it's counted once, from its first payment.

Recent months are left out of trends. Invoices raised in the last couple of months that are still unpaid can't be counted as late yet, which makes the newest months look better than they'll turn out.

Why Adfin publishes 34.3%, 8.7% and 9%

Adfin has published three late-payment rates that look as though they contradict each other. They measure different groups of invoices over different periods, and the biggest difference is whether direct debit is included.

The 34.3% describes the invoices where your customer decides when to pay, which are about 1 in 8 paid invoices on Adfin. Most invoices on Adfin are collected by direct debit on a set date, and very few of those go overdue. A rate across every invoice is therefore pulled well below a rate for customer-paid invoices alone, and 8.7% and 34.3% can both be right at once.

The 8.7% and 9% both cover every way customers paid, so they're close to each other and far from 34.3%. They come from different periods: the Barometer covers ten months to July 2026, while 9% is a rounded benchmark set in February 2026, so the two aren't expected to match exactly.

If you're quoting Adfin, use 34.3% for how often customers who choose when to pay are late, and the Barometer for the share of all invoices on Adfin paid late in its period. Before you set either against a national survey, it's worth checking that the survey counts the same invoices the same way.

What the data can't tell you

Adfin's records don't include a history of chasing for this period, so nothing here says which reminder, channel or time of day works, or whether payments that cluster on the due date followed a reminder or your customer's own routine.

Most differences between groups reflect who is in them. Compare payment methods inside a single business's own invoices and the on-time gaps fall below 1.5%, so a pooled comparison mostly describes which businesses and customers use each method. The same goes for invoice size, sector and terms, and no figure here shows cause.

Records hold an invoice's current status rather than its full history, payment terms are inferred from the gap between invoice date and due date, and a business's sector label can change. None of it is a UK national statistic: it describes Adfin's customers.

Rules for publishing a figure

Every figure passes the same checks before it appears on a page:

• Nothing is published from fewer than 200 payments.
• Nothing is published where one business supplies a large share of the result, however many payments are behind it.
• No individual business, customer or invoice can be identified.
• Rates are given as measured, sample sizes as rounded thresholds, and Adfin never publishes revenue or payment value.
• Each figure says which group of invoices and which period it covers, and restated figures are labelled.

Related guides

• The UK Getting Paid Report: how UK businesses pay and get paid
• Direct debit, card or bank transfer: which gets paid on time?
• Do clients pay faster with Apple Pay or Google Pay?
• How to forecast cash flow when clients pay late
• Credit control: a complete guide

Common questions

Is Adfin's payment data from a survey? No. It comes from Adfin's own records of invoices UK businesses raised through Adfin, with the due date, payment date and payment method attached. Nothing is modelled, weighted or self-reported.

How many invoices are behind the UK Getting Paid Report? More than 450,000 paid invoices from more than 1,000 UK businesses, raised between June 2024 and August 2026. On-time and days-late figures use the more than 55,000 that customers paid themselves.

What counts as paid on time? A payment on or before the due date set on the invoice, compared as calendar dates, so a payment late in the evening on the due date counts as on time.

Why are Adfin's late payment figures 34.3%, 8.7% and 9%? They count different invoices. 34.3% covers only invoices customers paid themselves, while 8.7% for October 2025 to July 2026 and the 9% benchmark set in February 2026 cover every payment method, including direct debit, where very few invoices go overdue.

Why is direct debit measured separately? A direct debit normally lands a few days after the due date because of the Bacs cycle, so a due-date test would read almost every collection as late. It's measured on success instead: 97% first time and 98.7% in the end.

What can't this data tell me? Which reminders or channels work, since chasing isn't recorded for this period, and anything about cause. The figures describe patterns among Adfin's customers, not the whole UK economy.

Sources

• Adfin - The UK late payments report (accurate as of November 2026) https://www.adfin.com/reports/late-payments-report
• Bacs - The little Bacs guide to Direct Debit (accurate as of November 2026) https://www.bacs.co.uk/media/ezwlg0dn/little_bacs_guide_to_direct_debit-v2.pdf
• Adfin - Next-generation direct debit (accurate as of November 2026) https://www.adfin.com/direct-debit-payments