Research · Mortgage · regulated data

Four spreadsheets merged, every NMLS number verified

Four overlapping mortgage-industry contact files with no shared identifier, merged into one list where every licence number was checked against the official register.

4→1Files merged
7Columns standardised
NMLSEvery number verified

The problem

The client had four separate spreadsheets of mortgage professionals, built at different times by different people. They overlapped, they disagreed, and there was no shared unique identifier to join them on.

In a regulated industry that is worse than it sounds. A mortgage loan originator carries an NMLS number issued through the Nationwide Multistate Licensing System. If the number is wrong, out of date, or attached to the wrong person, the record is not merely untidy — it is unusable for compliance purposes.

What we did

  1. Agree the merge key before merging

    NMLS number where present, falling back to a name-plus-company match for rows missing one. Deciding this first is what stops two people at the same firm collapsing into one record.

  2. Standardise all four files to one schema

    Seven columns, same order, same formats, same conventions for company names and phone numbers, before anything was combined.

  3. Verify every NMLS number at source

    Each licence checked against NMLS Consumer Access, the official public register. Not a lookup table, not an assumption — the regulator's own record.

  4. Cross-check employment currency

    A valid licence attached to a firm the person left two years ago is still a wrong record. Company emails were cross-checked so the employment reflected on the row was current.

  5. Merge, keeping the most complete value

    Where two rows described the same person, the survivor took the best value from each field rather than simply the newest row.

The result

One consolidated file, seven standardised columns, every NMLS number verified against the official register, and employment currency confirmed. The client could use the list for regulated outreach without a separate compliance pass.

Regulated data has a different definition of “clean”. In most projects, verification means the email will not bounce. Here it meant the licence exists, is current, and belongs to the person named on the row.

Last reviewed 28 July 2026. Figures are taken from the delivered project record. Client names are withheld where no permission to name has been given.

Questions

About this project

What is an NMLS number?

A unique identifier issued through the Nationwide Multistate Licensing System to mortgage loan originators and the companies they work for in the United States. It is publicly checkable through NMLS Consumer Access, which is where we verified every number on this list.

Do you work with other regulated industries?

Yes - banking, fintech, insurance and public sector. In each case we verify against the official register for that sector rather than a third-party database, and we record where each value came from.

How do you merge files with no shared identifier?

By agreeing a merge key in writing first, then standardising every file to one schema before combining. Where no identifier exists we match on a combination of fields and flag rather than auto-merge anything the rule is not confident about.

Can you verify licences at scale?

Yes, though verification against an official register is checked record by record rather than in bulk, so it takes longer than email verification. We quote it separately so you can see what that step costs.

Have a similar problem?

Send a sample of your data. We will tell you what is wrong with it, what it would take to fix, and what it costs — before you pay anything.

Get a free assessment WhatsApp