Experian's CFPB Relief Rate Fell to 0.44% in 2025

One bureau resolved 1 in 227 complaints. Another resolved 2 in 3.

ProPublica got here first. Joel Jacobs published it on March 10, 2026, and CNN picked it up the next day. We are not going to pretend we found it.

What ProPublica published was percentages. What nobody has published is the arithmetic underneath them, and what happened after their data stopped.

So here it is. In 2025, Experian resolved 0.44% of its closed CFPB complaints with relief. That’s 6,130 out of 1,388,821. One in 227. Equifax resolved 63.44% of its own complaints, and TransUnion 64.45%. Same year, same process, same regulator. A 144x gap between Equifax and Experian.

Every number on this page comes from one public file, and you can re-run it yourself.

That file is also how we know we have zero complaints in it. We checked ourselves first. Before writing a word about anyone else. Credit Booster AI appears 0 times across all 16,934,825 rows.

What we measured

The CFPB publishes every complaint it routes to a company, and it publishes how that company said the complaint closed. There are four outcomes that matter: closed with monetary relief, closed with non-monetary relief, closed with explanation, and still in progress.

Relief rate = (monetary + non-monetary relief) / all closed complaints.

That’s a within-company rate. A company measured against itself. This matters more than it sounds.

The obvious way to rank bureaus is complaints per consumer file. We couldn’t do it honestly, so we didn’t do it at all. Experian claims 220 million-plus US consumers. Equifax publishes a 400 million GLOBAL number. TransUnion publishes 1 billion, also global. None of those are audited, US-scoped, or comparable to each other. Dividing by them would look rigorous and be fiction. A rate of a company against itself needs no such denominator.

Here’s the concession that has to ship with the headline, and it’s a real one: this measures what each bureau reports about itself. The CFPB says so in its own December 2025 report: it “uses the term ‘relief’ for actions the NCRAs report they have taken in response to CFPB complaints.”

So a collapsing relief rate can mean a bureau granted less relief. It can also mean the bureau kept doing the same things and started filing them under a different label. Both fit this data. We can’t tell you which, and neither can anyone else working from the public file.

Write “Experian’s reported relief rate collapsed to 0.44%.” Don’t write “Experian stopped helping consumers.” The first is checkable. The second isn’t ours to say.

The numbers nobody printed

The regulator saw this before ProPublica did. The CFPB’s December 2025 report says, in prose: “Experian’s volume of complaints identified as resulting in relief has decreased and its use of administrative responses has increased.”

Then it printed the evidence as a chart with no values on it.

That’s the gap this page fills. Here’s Experian, quarter by quarter. Every one of these quarters is final: 0.0% of them were still open when we pulled the file.

Quarter	Relief	Closed	Rate
2024-Q1	44,036	122,841	35.85%
2024-Q2	46,432	164,692	28.19%
2024-Q3	27,880	214,227	13.01%
2024-Q4	22,306	253,078	8.81%
2025-Q1	4,843	307,617	1.57%
2025-Q2	606	323,676	0.19%
2025-Q3	214	372,977	0.06%
2025-Q4	467	384,551	0.12%

Look at 2025-Q3. Experian closed 372,977 complaints and granted relief on 214 of them. That’s 1 in 1,743.

It isn’t a blip. It’s held for six straight quarters across roughly 1.4 million complaints.

TransUnion is not a clean comparison either

This is the part where our own headline needs a correction, so we’re making it before someone else does.

The 144x gap is Equifax against Experian, and it’s exact. But putting TransUnion’s 64.45% next to Equifax’s 63.44% makes TransUnion look steady. It wasn’t. ProPublica said so, and our own quarterly data agrees with them.

TransUnion ran at 82.44% in 2025-Q2. Then 49.28% in Q3. It never went back. The full-year 64.45% is just a high first half and a low second half averaged into one misleading number.

So the honest framing is not “Experian is the outlier and the other two are fine.” It’s: Experian fell two orders of magnitude, TransUnion roughly halved, and Equifax was the one that held. For a while.

And Equifax has an asterisk of its own, sitting right there in the chart above. In 2023-Q1, Equifax reported relief on 22 of 70,982 closed complaints. That’s 0.03%, which is lower than Experian has ever gone in any quarter, including its 2025-Q3 floor of 0.06%. Nine months later Equifax was at 62.15%.

Sit with that. No operational change takes a company from 0.03% to 62% in three quarters. What moves that fast is which button gets clicked on a closure form.

That’s the strongest evidence on this page, and it argues against our own headline being read too hard. These rates track labeling behavior at least as much as they track what any consumer actually received. It’s why the word “reported” keeps appearing here, and it’s why we won’t tell you what Experian is doing behind the label. The CFPB saw the same thing across all three bureaus earlier: “Beginning in 2021, all three NCRAs began predominantly providing administrative responses to complaints. In 2022, all three NCRAs changed course.”

What surprised us: Equifax is falling too

ProPublica’s data ran through February 23, 2026, and at that point Equifax was the control group. Their piece says Equifax “did not show a similar decline.” That was right.

We pulled the file on July 16, 2026. Equifax is no longer holding.

Bureau	2025 full year	2026-Q1	Still open in Q1
Equifax	63.44%	49.89%	1.2%
TransUnion	64.45%	44.88%	1.0%
Experian	0.44%	0.19%	1.8%

Equifax granted relief on 228,716 of 458,452 closed complaints in 2026-Q1. That’s a drop of about 13 points from its own 2025 average.

Now the caveat, and we ran the check before believing our own finding. 2026-Q1 is near-final, not final: 1.2% of Equifax’s complaints from that quarter were still open at pull time. So we bounded it. If every single open complaint eventually closed WITH relief, Equifax’s Q1 rate would be 50.48%. If every one closed without, 49.30%. The finding survives both ends. It isn’t a counting artifact.

We’re not going to tell you why any of this is happening. We didn’t measure that.

What this means for your file

Almost nothing, and that’s the honest answer.

A relief rate is a fact about a company’s reporting behavior. It’s not a prediction about your report. It won’t tell you which bureau will fix your file. It can’t. The closure label is chosen by the company being complained about.

Here’s the part that costs us something to say: no app can make a bureau grant relief. Not ours, not anyone’s. If a tool tells you it can force an outcome at Experian, that tool is lying to you. What you actually control is narrow. You can know what’s on your three reports, catch the errors that are genuinely wrong, and keep your own records straight.

That’s the boring work, and it’s most of the job. Credit Booster AI does the reading and the tracking part across all three bureaus, on a 7-day free trial. It won’t change anybody’s relief rate.

Method

Source. CFPB Consumer Complaint Database, bulk CSV, https://files.consumerfinance.gov/ccdb/complaints.csv.zip. Pulled 2026-07-16, upstream Last-Modified: Thu, 16 Jul 2026 09:33:37 GMT. SHA256 9746821f7bfffb09947cab3e1d9acb3400edabb78388c8fbe70ff71f35aaaf48.

N. 16,934,825 complaints, 2011-12-01 to 2026-07-16. Every row read. Nothing sampled.

Bureau matching. The CFPB normalizes each bureau to exactly one company string: EQUIFAX, INC., Experian Information Solutions Inc., TRANSUNION INTERMEDIATE HOLDINGS, INC.. No subsidiaries swept in.

Denominator. Closed complaints only. In-progress and blank are excluded, because an unresolved complaint has no outcome to rate.

Censoring check. We measured the open share of every quarter before citing it. 2024-Q1 through 2025-Q4: 0.0% open, fully mature. 2026-Q1: 1.0% to 1.8% open, near-final, bounded above. 2026-Q2 is excluded entirely (52.7% open for Experian). 2026-Q3 is excluded (two weeks of data).

Validation. Our pipeline was checked against the CFPB’s own published figures in its December 2025 report, and every headline here was recounted from the raw file by a second, independent script that ignores the first one’s output. Both agree exactly.

What we could not compute. Days to respond, by bureau. The public file has a received date and a sent-to-company date, and no company response date. The CFPB’s own figure on this uses internal data.

Prior work. Joel Jacobs, “Experian and TransUnion Are Leaving More Mistakes on Credit Reports”, ProPublica, March 10, 2026. Read it first. It broke the story, and in May 2026 lawmakers wrote to the bureaus about it.