Data sources & methodology

Where the data comes from

• AFCA (Australian Financial Complaints Authority) publishes a complaint-by-firm Datacube covering insurance, banking, and other financial-services complaints, updated monthly. Firms with fewer than 4 complaints in a 12-month period are not published by AFCA and do not appear here either.
• State energy and water ombudsman schemes (EWON in NSW, EWOV in Victoria, EWOQ in Queensland) publish periodic reports covering energy-sector complaints. Where a scheme publishes company-level figures in a format we can verify, we include it; where it only publishes aggregate, industry-wide figures, we do not invent a company-level breakdown that was not actually published.
• CFPB (Consumer Financial Protection Bureau), for the US leaderboard. The CFPB publishes a Consumer Complaint Database as a machine-readable public API, updated daily. Every complaint in it was forwarded to the company named, which was required to respond, and both the complaint and the response category are part of the public record. We take complaint counts by company by quarter directly from that API.
• Scope of the CFPB data. It covers consumer finance only — banks, credit cards, credit reporting, debt collection, mortgages, and student, auto and payday loans. The CFPB has no jurisdiction over telecom, energy, retail, travel or insurance, so companies in those industries do not appear on the US leaderboard. Their absence means the regulator does not cover them, not that they attract no complaints.

What "per 10,000 customers" means, and why it's often missing

A raw complaint count is misleading on its own: a company with ten million customers will naturally generate more complaints in absolute terms than one with ten thousand, even if it treats each customer identically. Normalising complaints against a customer- count denominator (per 10,000 customers) makes company sizes comparable. We only show this figure where the source itself publishes, or we can otherwise verify, a customer count for that company and period. We do not estimate or infer a customer count to manufacture this figure. Where it is not shown, that is a gap in what was published, not something we chose to omit.

What "resolution rate" means

Where a source publishes what proportion of complaints against a firm were resolved (in the consumer's favour, or at all, depending on how that source defines it), we show it as that source defines it, alongside a link to the source so you can check the exact definition used for that period.

For CFPB rows specifically, this figure is the share of complaints the company closed with monetary or non-monetary relief, as categorised by the company itself in its response to the CFPB. Complaints closed with an explanation, or closed without relief, are not counted as resolved. Complaints still in progress are excluded from the denominator entirely, so a company is not penalised for the recency of its complaints. That is the CFPB's own categorisation, not our assessment of whether the outcome was fair.

Do not compare this figure between companies.Because each company categorises its own responses, two firms handling complaints similarly can report very different relief rates purely through labelling — one filing an outcome as "closed with non-monetary relief" where another files the same outcome as "closed with explanation". We have observed exactly that pattern between credit bureaus in the same quarter. The figure is meaningful as a company's trend over time against itself; it is not a league table of who treats customers better, and we do not present it as one.

Update cadence and the review gate

AFCA data is checked for updates monthly. Energy ombudsman data is added as new reports are published, which is typically annual or half-yearly depending on the scheme. Every new row, whether pulled in automatically or entered manually from a downloaded report, is saved in a "pending" state first and is reviewed by a person before it becomes visible on this page. Nothing appears here automatically the moment it is ingested.

What this page is not

This is not a ranking of company quality, customer service, or trustworthiness, and nothing on this page should be read as an accusation, characterisation, or opinion about any company. Any ranking shown is explicitly labelled by the specific metric it measures (for example "most AFCA complaints, insurance, 2026-01"), never as a general judgment. A high complaint count can reflect company size, a difficult market period, a single systemic issue, changed reporting practices, or genuine service problems — the regulator's own published data and reports are the place to look for that context, and we link to them on every figure.

Corrections

If a source has updated or corrected a figure shown here, or you believe a figure has been misattributed or transcribed incorrectly from the cited source, contact us via the details on our homepage and we will review it promptly.