The Spreadsheet Jockey: Artisanal Excel, Please

Dave: "I need the list of active facilities with their regional codes."

Jen: "Pull it from the asset management system."

Dave: "I did. It has 412 facilities. The budget system says 397. The HR system says 408."

Jen: "Which one is right?"

Dave: "None of them. Ralph in Building 4 has the real number. He keeps a spreadsheet."

Every organization has a Ralph. [not his real name. but you already knew that.]

He is the person who quietly built a spreadsheet years ago because the official systems couldn't answer a basic question. Now he maintains it. He's the only one who understands the formulas. Three VPs depend on his Tuesday morning update. Two dashboards pull from his file. And one entire compliance submission is just his spreadsheet reformatted into a PDF.

Who is the Matador? He or she is the Spreadsheet Jockey. And "The Matador" is the most important person in your data operation that nobody will ever put on an org chart. The Matador often gets applause which sounds like seals at the waterpark.

Who is the Crowd? The managers who need numbers by Tuesday.

Who is the Bull? Your live operational data, about to flip the matador and his "vlookup" upside down.

I chose this metaphor because it works for what we all observe daily. The Four Words explain why he exists.

[Catching up? The Four Words: Data Governance, Data Standards, Data Management, Data Exchange. Same order. Every time. Parts 1 to 3 covered why you can't skip the first two.]

Ralph exists because Word 1 (Data Governance) was never done. Nobody documented the authoritative source for "active facility." Nobody assigned ownership of that definition. Three systems were bought, each with its own interpretation, and nobody reconciled them. So Ralph did it himself.

That's not initiative. That's organizational failure wearing a lanyard.

The human middleware problem

Here's what happens in practice. An executive asks a question. "How many active contracts do we have?" The answer should come from a system. Instead, it triggers a phone tree. Someone emails Ralph. Ralph opens his spreadsheet. Ralph cross-references two databases, adjusts for the records he knows are duplicates, removes the ones that should have been archived last quarter, and sends back a number.

That number is accurate. I want to be clear about that. Ralph's spreadsheet is usually the most accurate source in the building. For a brief moment in time, that's the problem.

When your most accurate data source is a person with a .xlsx file on a shared drive, you have built human middleware. You have turned a GS-12 into an API. And you are one retirement, one transfer, or one bad Tuesday away from losing your entire data pipeline.

[side note: I have personally watched an agency treat a massive excel sheet as "live" inventory which then drives leadership dashboards and downstream tools provided to support contractors. Human middleware is a true story. The backup plan was "someday we should try really hard to modernize data integrity."]

[Dad Joke] Why did the spreadsheet break up with the dashboard? It was tired of doing all the work and getting none of the credit.

Why this keeps happening

Ralph didn't create this problem. The org did. The pattern is always the same:

Leadership buys a platform. The platform is a Data Lake with a proprietary ontology which does not match reality, and now it includes AI features without governance. What could possibly go wrong?! Now the large tools platform leads your government data strategy instead of following. Is your vendor-provided ontology portable? We all know the answer is... doubtful.

If you cannot take your ontology and model with you, you do not own it, you are RENTING it. No agency should outsource or rent its own data governance.

Nobody wants to do the governance work to reconcile the model with the business definitions. So the platform launches with "good enough" data. Two quarters later, someone notices the numbers don't match. Ralph builds a spreadsheet to "bridge the gap." Five years later, that bridge IS the infrastructure.

This is what happens when you skip Word 1 and Word 2. Your tools (Word 3) get deployed without governed definitions or standards. Your integrations (Word 4) move data between systems that don't agree on what the data means. And somewhere in the middle, a human being holds it all together with VLOOKUP and conditional formatting.

The Spreadsheet Jockey audit

Want to know how big this problem is in your org? Ask four questions:

1. For your top 10 business entities (customer, contract, facility, employee, asset), can you name the single authoritative system of record for each?
2. When the numbers from two systems don't match, who do people call?
3. How many spreadsheets in your org exist because a system couldn't answer a basic question? Manager consensus: "well it's all we have really."
4. Are your dashboards lying to leadership? Be honest.

If the answer to #2 is a person's name and not a system, you have a Spreadsheet Jockey. You probably have several. And your data governance program should start by documenting what they know, not by buying another platform.

[Save those four questions. Use them in your next data strategy meeting. The room will get very quiet.]

What 1st Class governance does for Ralph

Here's the good news. When governance is done right, Ralph gets his life back.

Word 1: document the authoritative source for every key business entity. Publish it. Make it official. No ambiguity.

Word 2: define the standards. "Active facility" means this. Not that. Not "it depends." Every system that stores facility data conforms to this definition or gets flagged.

Word 3: configure the tools to enforce those definitions. The platform doesn't accept records that violate the standard. The dashboard pulls from the governed source.

Word 4: integrations move governed data between systems that share the same definitions.

Good news: this is the moment that $$$ AI platform returns real, measurable value. Not before. When governance leads, tools follow.

And Ralph? Ralph goes back to his actual job. The one he was hired to do before the org turned him into a data janitor.

The path forward

Your Spreadsheet Jockeys are not the problem. They are the symptom. They exist because governance didn't happen first. Every spreadsheet workaround is a postcard from a system that failed to answer a basic question.

Four words. Same order. Same answer.

If you want to retire the spreadsheet, govern the source.

Part 4 of The Governance Series by daitagov. First published on LinkedIn.