How small payment data errors snowball

A late payout, a misrecorded chargeback, a missing payment. At high volume, small data errors pile up until nobody trusts the revenue figure.

The Actuals teamActuals · 2 min read

A payout arrives a day later than expected. A chargeback is recorded against the wrong order. One payment never shows up in the bank file at all. On the day itself, none of it looks serious.

In a company that processes thousands or millions of orders, payments or subscriptions, these inconsistencies do not stay minor. They pile up until nobody fully trusts the revenue figure anymore.

Where the data drifts apart

A typical stack has an e-commerce, ticketing or subscription billing platform at the front. Behind it sit several payment processors, and at the end an ERP. Every handoff between those systems is a place where a transaction can go missing, arrive twice or change shape.

Unreliable data makes reconciliation hard, but that is only the first problem. Revenue gets misreported and forecasts drift. Compliance becomes a risk, and decisions get made on numbers that are slightly wrong.

What broken data costs

The research below is not specific to finance, but the pattern will feel familiar to any controller.

  • McKinsey found that knowledge workers spend about 19% of their time searching for and gathering information.
  • A 2018 IDC study found that data professionals lose about 50% of their week to broken data. That is 30% searching for, governing and preparing data, plus 20% duplicating work.
  • Gartner estimates that poor data quality costs organisations an average of $12.9 million a year.

For a finance team, that lost time is time spent fixing data instead of analysing it.

One subscription business, before and after

One client we worked with is a global company processing a high volume of subscriptions, across several payment providers and banks. Before they automated their data controls, the finance team spent weeks fixing discrepancies. Payments went missing, chargebacks were misrecorded, and delayed payouts led to errors in revenue.

After automating, the errors dropped and manual work was cut by more than half.

What reliable data takes

Reliable data does not come from checking harder at month end. It comes from keeping three things in sync every day.

  1. Orders, payments and payouts from every system come into one place.
  2. Each payment is matched to its order and each payout to its payments, so errors show up the day they happen.
  3. Accounting works from that matched data, so finance has one source of truth instead of several versions to reconcile.

Think back to the late payout, the misrecorded chargeback and the missing payment from the opening. With this in place, each one is visible within a day, with a reason attached.

A test you can run this week

Take last week’s payouts from one payment provider. Trace each one back to the orders behind it. If every payout ties out without a search, your data is in good shape. If it takes your team days, or some payouts never fully add up, you know where the snowball starts.

Sources

  • McKinsey, research on how knowledge workers spend their time, as cited in the original article
  • IDC (2018), study on time data professionals lose to broken data, as cited in the original article
  • Gartner, estimate of the annual cost of poor data quality, as cited in the original article
In practiceAdyen, Stripe, PayPal. Three report formats, one truth?

Written by

The Actuals teamActuals