How We Found $25 Million Hidden in a Supply Chain

Nobody thought $25 million was sitting in the supply chain.

The company I was working with was a well-run operation — experienced leadership, solid margins, decades in business. They weren't making obvious mistakes. They were buying from suppliers they'd worked with for years, paying prices that felt fair, and getting product that mostly arrived on time.

But mostly is doing a lot of work in that sentence.

The problem wasn't that they were making bad decisions. The problem was that they were making decisions without information. Every supplier selection was a combination of relationship, habit, and instinct. Nobody had ever built a system to ask the obvious question: are we choosing the right supplier for this job?

The Question Nobody Was Asking

When I started looking at their procurement data, the first thing I noticed was that the same decision was being made hundreds of times a year — which supplier to use for a given project — with almost no consistency in how that decision got made.

Sometimes price won. Sometimes it was whoever called first. Sometimes it was whoever the buyer liked. Sometimes it was inertia — we used them last time, so we'll use them again.

The data told a different story. When I looked at the historical outcomes — not just the prices paid, but what actually happened after the contract was signed — a pattern emerged:

• Some suppliers consistently delivered on time. Others consistently didn't, and the cost of that lateness was never showing up in anyone's price comparison.
• Some suppliers had quality issues that generated returns and rework. Those costs were buried in other budget lines, invisible to the people making buying decisions.
• Some supplier relationships were generating real value. Others were generating comfortable familiarity and hidden costs.

The company had 50,000+ historical project records. Every one of them was a data point. Nobody had ever looked at them as a system.

What We Built — And Why It Worked

The solution wasn't complicated in concept, even if the execution required some work. We built a vendor scoring system that looked at every supplier across three dimensions simultaneously:

1. Cost — not just the quoted price, but the total cost including returns, rework, and any costs generated downstream by their work
2. Quality — defect rates, return rates, issue frequency, and issue severity over time
3. Delivery — on-time performance, communication during delays, and the operational cost of late delivery

Each supplier got a composite score on every project type. The scores updated automatically as new project data came in. When a buyer needed to select a vendor, instead of relying on instinct, they had a ranked list of suppliers for that specific job type, weighted by what actually mattered.

The system didn't replace human judgment. Buyers could still choose a lower-ranked supplier if they had a good reason. But now there was a default — and the default was based on evidence rather than habit.

The result was $25 million in annual procurement savings. Not from renegotiating contracts. Not from cutting suppliers. Not from paying people less. From making better decisions with information that already existed.

What This Means For a $10M Business

The company I was working with is significantly larger than $10 million in revenue. But the principle scales down — and often the ROI is actually better at smaller companies, because the gap between current practice and best practice is wider.

Here's what the equivalent looks like at a smaller scale:

You probably have the same problem with a different label

Maybe it's not procurement. Maybe it's marketing spend — you're running campaigns based on what worked two years ago or what your agency recommends. Maybe it's inventory — you're ordering based on last year's sales or gut feel about next season. Maybe it's staffing — you're scheduling based on experience rather than what your transaction data actually shows about when you're busy.

Every one of those is the same problem: a recurring decision being made without a system, using information that already exists but hasn't been organized.

You don't need 50,000 data points to start

The company I worked with had 50,000 historical records. You might have 500. That's enough to start seeing patterns. The question isn't whether you have enough data — it's whether anyone has ever organized it into a form that lets you learn from it.

The money is already in your business

This is the part I want to be direct about. Most of the value I find for clients isn't new revenue. It's revenue they're already generating that's leaking through decisions made without information. Suppliers who are more expensive than they look. Customers who are less profitable than they appear. Marketing channels that seem to work but don't. Inventory that ties up cash that could be working harder.

The $25 million wasn't created by the system. It was already there — it was just invisible.

Three Questions To Ask About Your Own Business

If you want to know whether there's a similar opportunity in your company, start here:

1. What decisions do you make repeatedly that could be improved with better information? Procurement, marketing, inventory, pricing, hiring, scheduling — any recurring decision is a candidate.
2. Do you have historical data on the outcomes of those decisions? You probably do, even if it's scattered across your accounting system, your CRM, and someone's spreadsheet.
3. Is anyone currently connecting that data to the decision? In most companies I talk to, the answer is no. The data exists. The decisions happen. Nobody has built the bridge.

If you answered "no" to question three, there's almost certainly money sitting in your business that a data system would find.

The Honest Caveat

I want to be clear about something: $25 million is not a typical result. That number reflects a large organization with significant procurement volume and a gap between current practice and best practice that happened to be particularly wide.

What is typical is finding meaningful, measurable savings or revenue in the first 90 days of a serious data engagement. For a $10 million business, that might be $200,000. For a $25 million business, it might be $800,000. The ratio is often similar even if the absolute number is smaller.

What's also typical is that those results compound. A better supplier scoring system gets better as more data flows through it. A customer retention model improves as more customer behavior is captured. A demand forecast gets more accurate as more seasonal cycles are observed.

The first year is usually the worst year in terms of return. Which means the question isn't really "is there a payoff?" The question is "how long are you willing to wait before you start?"