- What is capital allocation deciding in a business this size?
- Why the hurdle rate is the wrong first filter
- The order that works
- Capacity ahead of demand
- Flexibility instead of capacity
- Working capital is capital
- Buying, repaying, and returning
- Why standing still is also an allocation
- How to run the review
- What Berkshire taught me about pace
The question usually arrives in the form of a hurdle rate. What return does a project need to clear before we fund it? It is a reasonable question and it is the second one. The first is narrower and far more useful: which dollar are we talking about, and what does the business physically do with it?
In a mid-market industrial business, those two questions produce different answers often enough that the order matters. A project can clear every hurdle rate you set and still be the wrong place for the next dollar, because the money went somewhere the business was not constrained. And a project that looks marginal on paper can be the best allocation available, because it removes the one thing limiting everything else.
This is written for the person who actually decides: an owner, a general manager with a capital budget, a searcher who just bought a plant, or an operating partner deciding which of several requests to fund. It sits under how to run year one after an acquisition, because the first serious allocation decisions usually land in the first twelve months of new ownership.
What is capital allocation deciding in a business this size?
It is deciding which of four destinations gets the next dollar: physical capacity, working capital, an acquisition, or the owners and lenders. In a business under a few hundred million in revenue, those four compete directly, because there is rarely enough cash to fund more than one of them well.
The four destinations
Capacity means anything that changes what the plant can physically do: machines, tooling, space, automation, or the people to run a second shift. Working capital means inventory and receivables, which absorb cash quietly and give it back slowly. Acquisitions buy capability, capacity, or customers you would otherwise have to build. And the fourth destination is simply out: debt repaid, or distributions to the owners.
Why size changes the problem
A large company can fund several of these at once and let the portfolio sort it out. A mid-market industrial business usually cannot. One capital decision a year is often the whole story, which means the cost of putting it in the wrong destination is not a lower return. It is a year.
That is why the sequencing question outranks the arithmetic question at this size. Getting a decent project in the right destination beats getting an excellent project in the wrong one, because the excellent project leaves the constraint exactly where it was.
Watch the companion video on Debt vs Equity: The Capital Structure Theory.
Why the hurdle rate is the wrong first filter
Because a hurdle rate ranks projects against each other without asking what limits the business. Two projects can show the same return while one of them removes a constraint and the other adds capability behind a constraint that has not moved.
What an IRR calculation assumes
Every return calculation on a capital project assumes that the incremental volume it enables will actually be sold and actually be produced. Both halves of that assumption sit outside the spreadsheet. If the plant is already full where that volume would land, the calculation is describing a machine that will make product the business cannot ship on time.
What the constraint does to the same calculation
Run the same arithmetic on the constrained operation and the numbers change without any new assumptions. The incremental volume is real because there is demand waiting on it, and the incremental cost is lower because you are not building a second version of something that already exists.
So the first filter is not a rate. It is a question: where does this business stop being able to say yes? Fund that, then apply the hurdle rate to the candidates inside it.
The order that works
Ask five questions in sequence, and stop at the first yes. Is there a constraint limiting profitable volume? Can flexibility remove it more cheaply than capacity? Is working capital the constraint instead? Does an acquisition buy something you cannot build? Only then do repayment and distributions compete.
Two notes on using it. The sequence assumes you have actually established the constraint rather than accepted the version of it that circulates internally. Every function in a business nominates its own constraint, and the nomination usually matches whatever that function would like funded. The test is whether removing it tomorrow would immediately produce more profit, and whether the evidence behind it came from the floor and the order book rather than from a monthly report.
The second note is that the answer changes. A business that was capacity-constrained last year and had the capacity built is not capacity-constrained now, and continuing to fund the destination that worked last time is the most common way a good allocation habit turns into a bad one. Rerun the sequence from the top each time rather than picking up where the last decision left off.
The sequence is not a ranking of importance. It is a ranking of dependency. Capacity behind a lead-time constraint does nothing. An acquisition bolted onto a plant that cannot serve its own customers on time makes two unhappy customer bases instead of one. Each step only becomes the right answer once the step above it is genuinely a no.
Capacity ahead of demand
Capacity has to be built before the revenue that justifies it, which is why it is the hardest destination to defend and the one that most often pays. If you wait for the orders, you lose them on the lead time that the missing capacity creates.
The grocery scale case
In a grocery scale business I worked on, a new scale had launched two years earlier and was not selling. It cost more, and it only paid off for a store that replaced all of its scales at once, while stores replaced them one at a time as they broke. Part of the fix was commercial, pitching thousands of scales at once on total cost of ownership rather than on break and fix. The other part was capital: new-scale capacity was tripled before the demand existed to fill it, old-scale production was cut, and the old scale’s price was raised hard.
Revenue went from $42 million to $67 million and profit from $6 million to $15 million. The capacity came first. It had to, because the commercial pitch was a promise to replace an entire fleet at once, and that promise is not credible from a plant that cannot build the fleet.
Why the order of operations matters here
Capacity that arrives before the commercial change it depends on is idle. The commercial change that arrives before the capacity is worse, because it is a promise the plant cannot keep, and a missed promise costs more than a late one. In the scale business the two were run deliberately in parallel, with the capacity slightly ahead, and the price of the old product raised at the same time so that the choice the business wanted customers to make was also the obvious one.
What it costs to be wrong
Unused capacity looks like waste on every report an owner reads, and sometimes it is. The honest version of this decision names what would have to be true for the capacity to fill, and states when you will know. If the commercial change it depends on has not moved within two or three quarters, you have learned something, and that is a far better position than never having built it and never knowing.
Flexibility instead of capacity
Before funding new capacity, check whether the real problem is that existing capacity can only make one thing. Flexibility is usually an order of magnitude cheaper than capacity and it removes a kind of constraint that capacity does not touch.
Two hundred thousand dollars versus two million
At a shopping cart manufacturer inside Berkshire Hathaway, there were eight automatic lines. A few ran far over capacity, many ran far under, and half of all products had no automatic path at all, because each line had been automated for exactly one cart variation. Read as a capacity problem, the answer was a new automatic line at about $2 million. Read as a flexibility problem, the answer was $200,000 of end-of-line machinery that let an existing line run a second, third, and fourth product. Volume on automatic lines went from under half to nearly ninety percent.
The follow-on is the part I think about more. A later win of about $50 million came on a cart designed from the start to run on a line that was sitting vacant. Flexibility did not just fix the existing mix. It changed what the commercial team could go and sell.
The same logic in reverse
Flexibility also cuts the cost of being wrong about demand. A line that can run four products is a bet on a category. A line that can run one is a bet on a customer. The second bet can be correct and it carries a risk that rarely appears in the capital request, which is that the asset’s value depends entirely on a relationship the business does not control.
When flexibility is the cheaper answer
Look for it whenever utilization is uneven across similar assets, whenever changeover time is the reason a line looks full, and whenever a single customer’s loss would idle an asset. Those three conditions all describe a business that has bought capacity and not bought optionality, and the second is usually available at a fraction of the price of the first.
Working capital is capital
Inventory is an investment decision that most industrial businesses make by default rather than on purpose. When the binding constraint is lead time rather than output, inventory is often the highest-return allocation available, and it will look like the worst one on the balance sheet.
Inventory as an investment
In the tank liner business I owned, I raised inventory so lead time could come down. Win rates rose, and the business became the default vendor on projects of $100,000 and up. Nothing about that decision looks good in a working capital review. Everything about it looks good in the order book, which is where it was supposed to show up.
Buying long-lead parts ahead of orders
The same logic scales. In a beverage filling equipment business, the longest lead item in the build was forgings, and buying them ahead of customer orders cut twenty weeks out of the quoted lead time. That is cash spent before any customer has committed, which is exactly the spend a cash-focused review is designed to stop.
The test that separates a good inventory investment from a bad one is whether the inventory removes a constraint that is costing you orders. If it does, it is capital allocation. If it is there because nobody set the reorder points properly, it is a working capital problem and should be treated as one.
Buying, repaying, and returning
Acquisitions, debt repayment, and distributions are the right destination when the operating constraints have genuinely been answered. Each is defensible. Each is also where money goes when nobody has done the work of finding the constraint.
Buying capability you cannot build
An acquisition earns the next dollar when it buys something the business cannot create in a reasonable time: a capability, a customer position, a geography, or capacity that would take years to permit and build. It is a weak answer when the same capability could be added to the existing plant for a fraction of the price, which is more often the case than deal pipelines suggest.
When repaying debt wins
Repayment wins when the cost of the debt is genuinely above the return on the best available operating project, and it wins for a second reason that does not show up in a rate comparison. Debt constrains what the operating plan can do in a bad quarter. In a cyclical industrial business, paying some of it down buys the ability to keep funding capacity when demand dips, which is precisely when capacity is cheapest to build.
What an acquisition costs that the model leaves out
Management attention, for at least a year, and usually the attention of the same people who would have run the operating changes. That is the real competition between destinations in a business this size. A bolt-on and a plant reorganization do not compete for cash so much as they compete for the two or three people capable of leading either one, and the cash model never shows that constraint.
When to take it out
Distributions are the honest answer when the plant is not constrained, the price list is right, the lead time is competitive, and no acquisition beats the alternatives. That combination exists. It is just rarer than the number of businesses that distribute would suggest, and the way to earn the right to that answer is to have looked.
Why standing still is also an allocation
Giving every part of the business the same share of capital it got last year is a decision, not the absence of one. The evidence suggests it is a common one and an expensive one.
McKinsey’s analysis of 1,508 US-listed companies over the twenty years to 2010 found that most multibusiness companies gave each business an unchanging share of total corporate capital year after year. Comparing the top third of reallocators with the bottom third, the annual gap in total shareholder returns widened to 3.9 percentage points, and the survival gap between the two groups reached 22 percent. The same research found that the 2007 to 2010 downturn had almost no effect on how actively companies reallocated, even though tight capital is exactly when reallocation matters most.
Giving every part of the business the same share of capital it got last year is a decision, not the absence of one.
Two details in that work are worth carrying into a much smaller business. The first is that the high reallocators moved existing resources as readily as new ones, while the low reallocators mostly only redirected new money. Taking capital away from something that already has it is the harder half, and it is the half that separates the two groups. The second is that the returns went to companies making repeated, stepwise shifts year after year rather than one or two dramatic swings, which is an argument for reviewing allocation on a rhythm rather than in a crisis.
The mid-market version of the same inertia is a capital budget built by asking each plant what it needs. That process produces a list shaped by who asked well, not by where the business is constrained.
The version of this that applies to a single plant
If you run one site rather than several, the same inertia shows up as a maintenance and replacement budget that repeats last year’s pattern with an inflation factor on it. The reallocation question at that level is which line gets the money this year, and the answer is almost never the line that has always received it. The line that has always received it is usually the one whose supervisor is most practiced at asking.
How to run the review
Run it twice a year, start from the constraint rather than the request list, and require every proposal to name what the business will be able to do that it cannot do today. Two hours with the right question beats a quarterly process with the wrong one.
What to put on the table
Three things. Where the business currently stops being able to say yes, with evidence from the floor and the order book rather than a report. The list of requests, each rewritten as the capability it creates. And the honest cash position, including what is already committed.
What to decide
One destination, not five. Name the constraint the next dollar is going to remove, fund it properly rather than partially, and write down what should be visibly different by when. Partial funding of three constraints is the most common failure in mid-market allocation, because it produces three unfinished changes and no removed constraint.
What to write down
The decision, the evidence behind it, and the date you will check. The value of that record is not governance. It is that six months later you can tell the difference between a project that failed and an assumption that was wrong, and those two require completely different responses.
What Berkshire taught me about pace
The most useful thing I took from operating inside Berkshire Hathaway was not a technique. It was a tolerance for spending money on things that will not show up for a while, and a matching unwillingness to spend it on things that will never show up at all.
Those two sound like opposites and they are the same discipline. A business that will not fund capacity ahead of demand, or inventory ahead of orders, is not being careful. It is choosing a slower version of itself and calling it discipline. A business that funds every project with a good return is not being aggressive. It is avoiding the harder job of deciding which constraint actually limits it.
The allocation decisions I have seen pay best were unglamorous and early: flexibility rather than a new line, capacity before the commercial promise that depended on it, parts bought before a customer had committed. None of them looked impressive in the quarter they were made. All of them changed what the business could physically offer, which is the only thing capital can really buy.
If you take one thing from this, take the order. Constraint first, flexibility before capacity, working capital before acquisition, and the other three destinations only once those questions have been genuinely answered rather than skipped.
About the Stagnation Assassin
Todd Hagopian is a Fortune 500 transformation executive who has generated $3B+ shareholder value across five turnarounds at Berkshire Hathaway, Illinois Tool Works, and Whirlpool. He is VP of Product Strategy and Marketing. Known as The Stagnation Assassin, he is the author of two published books: The Unfair Advantage: Weaponizing the Hypomanic Toolbox and Stagnation Assassin: The Anti-Consultant Manifesto. His blog is published in 15+ languages and read by operators worldwide. Bring him to your stage via his speaking page or connect with him on LinkedIn. His newest book is 10-Minute Transformation (February 23, 2027).
Running a portfolio leadership meeting or an industry conference? Operators who have to defend a capital request, and owners who have to choose between them, tend to argue about rates when the argument is about constraints. My speaking page has the keynote and workshop formats that walk a room through the order above using their own capital list.

