Find the pattern behind the complaints.

Texas shows 487,000 published narratives. Georgia shows 286,000. New York shows 221,000. Behind every one of those numbers is a person who took time out of their day to write down what happened.

Reading a cluster is not reading a statistic. It is reading hundreds of people, in their own words, describing the same thing.

The two largest companies in the database sit at 817,000 and 775,000 published narratives. A mid-sized one carries 66,000. Size is the reason for the gap, not the point of it.

Raw count never says what is wrong. What matters is whether the complaints look the same, whether the same problem keeps showing up, and how the company responds when it does. The tool answers all three, and it treats an 817,000-narrative company the same way it treats a 66,000-narrative one.

Some complaints get a refund. Some get an explanation. Some are still waiting for either. 2.5 million complaints closed with an explanation. 1.9 million closed with monetary relief. 3,700 remain in progress.

Response types are attached to every complaint in the data, so you can see whether a company engages or dismisses, and whether the pattern of complaints correlates with the pattern of responses.

Of every issue category in the database, three account for more than half of the total. The single largest issue alone covers 1.2 million complaints, a number so broad that a filter treats them all the same way.

The tool splits that number into the specific complaints underneath, so a category becomes a set of real practices.

3.8 million complaints got a timely response. 43,000 were late. The ratio tells you whether a company takes the CFPB response window seriously, and the late ones cluster around a handful of large institutions.

Most companies in the database have a few hundred complaints. A small number have more than 50,000. Both groups are treated the same way by the tool: one company, one pattern set.

Size changes how many complaints there are, not what the complaints say.

Not every complaint comes with a written narrative. On the CFPB side, publishing a story is voluntary and separate from filing. That is why the numbers here are smaller than the raw complaint count (4 million narratives versus 37 million total complaints), and why every narrative the tool reads is a real person who chose to tell their story.

Banks, credit reporting agencies, debt collectors, card services, and mortgage companies make up almost the entire database. Credit reporting alone is 64 percent of the total.

Different industries, different complaint patterns, same clustering tool. The pattern of a bank complaint and the pattern of a credit-bureau complaint look nothing alike.