Inventory Is A Spectrum
A Spectrum, Not a Switch
Ask ten manufacturers how they handle inventory and most will give you one of two answers: we make to order, or we make to stock. In practice, nobody actually lives at either end. Everybody is somewhere on a spectrum, and the honest question isn’t which camp you’re in — it’s whether you know where each of your products sits and why.
We’ve spent the last year getting deliberate about that. Today a little over a third of our catalog is stocked to a min and a max. The rest gets built when it’s ordered. That isn’t a compromise; it’s a decision we re-make on a schedule, using what the demand data actually says instead of what we assume it says.
The hard part is that we’re doing this while growing. Every new distributor changes the math. A single container order can look like a runaway bestseller until you pull the thread and find out it was one shipment to one customer. Transitioning your inventory approach while the volume underneath it is moving is a lot like changing a tire without stopping the truck — it can be done, but only if you’re watching the road!
Almost All of Us Are Small
Here’s a number worth sitting with. According to the SBA Office of Advocacy, there are 36.2 million small businesses in the United States — 99.9% of all U.S. businesses — employing 62.3 million people, or 45.9% of the private-sector workforce.
That statistic gets quoted a lot, usually as a feel-good line. That means something a bit different to me. It means the overwhelming majority of the companies we sell to, buy from, and compete against are working with the same constraints we are: a finite building, a finite crew, and no room to carry six months of the wrong thing. When a car wash operator calls us, they’re not a procurement department. They’re usually the owner, and they’re calling because a bay is down or a drum is almost empty.
So, the bar isn’t “be as sophisticated as a global chemical company.” The bar is: be the supplier that a small business can actually plan around.
Four Weeks Minimum, Six Weeks Maximum
Our rule is simple enough to say in one breath: four weeks of demand on hand at the low end, six weeks at the high end. Below four we’re exposed to a single large order clearing the shelf. Above six we’re paying to warehouse a guess.
For context, U.S. manufacturers as a whole carried an inventories-to-sales ratio of 1.47 in July 2026 — right around six and a half weeks of supply (U.S. Census Bureau, via FRED). So our ceiling is roughly where the average manufacturer’s normal is, and our floor is meaningfully leaner. That’s intentional. In a business where a formula can be built in a shift, carrying six-plus weeks across the board is a storage bill, not a service level.
The band only works if the lead time underneath it is real. We measure order issuance-to-shipment on every sales order — the days between the order being issued and the product going out the door — and we track it monthly against a hold-the-line target, split between deliveries we run ourselves and everything else. That number is what sets the minimum. When it moves, the min/max moves with it. Our reorder points also carry a floor equal to the largest single order a product has seen, because an average tells you nothing about the customer who buys a quarter’s worth in one Tuesday.
Same-day and next-day is not a slogan here. It’s a measured number we watch, and it’s the reason four weeks is enough.
Where AI Actually Fits
We’re a polymer chemical manufacturer in Ohio, not a software company, so I’ll be plain about what AI does for us. It doesn’t formulate our products and it doesn’t run our plant. What it does is take work that used to sit in a queue — reconciling a purchase order against a confirmation, pulling a year of customer purchase history into a seasonality curve, turning six raw system exports into one operations scorecard — and turn it around in an afternoon.
That matters because the constraint on a company our size has never been ideas. It’s hours. Every hour we don’t spend rebuilding the same spreadsheet is an hour spent on the floor or on the phone with a customer. We’ve put a company-wide framework around it — which tools handle internal data, which handle public research, who’s trained, what gets signed before anyone starts — because the fastest way to lose the benefit is to be careless with it.
The competitive read is straightforward. Larger competitors have had this kind of analysis for years; they just paid a department to do it. The tooling has closed that gap. What it hasn’t done is close the gap on knowing your customers — and that’s still ours to win.
The Small Things Customers Actually Feel
The biggest shift hasn’t been technical. It’s been a mindset change, and it’s mostly made up of things that sound too small to write down:
- The date on the order confirmation is a promise. Not an estimate, not a best case. If we put it in writing, it ships that day.
- Acknowledge the order the moment it comes in. Nobody should have to call and ask whether we got it. Silence after a purchase order is the cheapest way to lose trust.
- Manufacture to stock wherever we responsibly can. Every SKU we move off made-to-order is a customer who stops waiting on a batch.
None of those show up on a spec sheet. All of them are what an operator remembers when they’re deciding who to call next time.
And all of it runs on one idea: the better we know our customers, the better we can predict them. Knowing that a product moves ahead of the first hard freeze, or that a particular account orders in multiples of twelve, or that one account’s volume is a scheduled container rather than steady demand — that’s what lets us hold four weeks instead of eight, and still have it on the shelf when the call comes.
That’s the work. Not a bigger warehouse. A better-informed one.
Yours In The Sudz,
Jacob Miller