How Dirty Data is Stopping your Business from Growing.

I've worked with hundreds of businesses. Different industries, different sizes, different founders, different problems on the surface. Underneath, it's the same disease every time.

Dirty data.

Not "we don't have data." Every business is drowning in data. Dirty data: the customer who exists three times under three spellings. The location that's "Store #4" in the POS, "Main St" in payroll, and "MAIN STREET LLC" in the accounting file. The vendor that got renamed in 2023 but not in the system that pays them.

Nobody puts this on a risk register. It should be at the top.

The math nobody runs

Gartner estimates poor data quality costs the average organization $12.9 million a year. Scale that down to a business doing $5M in revenue and the proportions hold. You're essentially paying the tax in hours, bad calls, and, most importantly, your hard-earned cash.

There's an old rule in data quality called the 1-10-100 rule: it costs $1 to verify a record when you capture it, $10 to fix it later, and $100 if you do nothing and let the error do its work downstream. I've never seen a small business operate at the $1 stage. Most live at $100 and call it "how we've always done it."

Here's what $100 looks like in practice. A multi-location operator I worked with was convinced their newest location was underperforming. The numbers said so. They were three weeks from cutting staff there when we found the problem: half that location's sales were being coded to a different entity. The "underperformer" was actually their second-best store. The data wasn't lying, exactly. It was in the wrong place.

Why it's a ceiling, not a nuisance

Dirty data doesn't hurt you evenly. It hurts you precisely when you try to grow.

At one location, the owner is the data model. They know Store #4 and Main St are the same place because they signed the lease. Every discrepancy gets resolved by the person who holds the whole business in their head.

Add a second location, a bookkeeper, a manager, a second entity for the new state, and the owner's head stops scaling. Now every report requires a translation layer, and the translation layer is a person doing weekend spreadsheet work. Growth doesn't expose the dirty data problem. Growth is the dirty data problem, multiplied by every new location, system, and hire.

That's why so many businesses plateau at two or three locations. It's not the market. It's that every expansion makes the numbers less trustworthy, and nobody expands confidently on numbers they don't trust.

What actually fixes it

Not a cleanup project. I've watched businesses pay for one-time data cleanups the way people buy gym memberships in January: sincere, expensive, and undone by March. Data gets dirty continuously, so it has to be kept clean continuously.

What works is structural:

1. One canonical list of your real-world entities: locations, people, vendors, accounts. Not per system. One.
2. Mapped aliases. Every system's name for a thing points at the canonical thing. "Store #4" and "MAIN STREET LLC" resolve to the same location, permanently.
3. Lineage. Every number in every report traces back to a source record. If you can't answer "where did this figure come from?" in one click, you don't have real reporting.
4. Rules that outlive people. The fix can't live in your bookkeeper's head. It has to run the same way the month after they leave.

You can build this discipline manually. Some businesses do, and it's genuinely better than nothing. But it's exactly the problem we built Sourcebook to solve. It connects your existing systems, maintains the canonical model, and keeps the data clean as it flows, instead of after it rots.

The takeaway

Your competition isn't the ceiling on your growth. The gap between what your systems say and what's actually true is. Every business I've seen break through a plateau fixed their data before they fixed anything else. Every business I've seen stall was sure they'd get to it next quarter.

The tax is compounding either way. The only question is whether you're paying $1 or $100.