Your Cash Reserve Number Is Made Up — Here's the Real One

October 2008. Two builder accounts collapsed in the same week. Not slow-paid. Gone. I sat at the kitchen table on a Sunday night trying to figure out how long I could keep paying twelve men. I had a rough sense. I had a savings account I hadn't been in for a while. And I had a number in my head — "we're fine for a while" — that turned out to be fiction I'd told myself so I could sleep.

"A while" was six weeks. That's not a cushion. That's a countdown.

The "three months of expenses" rule didn't help me, because I didn't actually know what three months of expenses was. I knew what was in the account on Thursday. I knew what the payroll run looked like on Friday. What I didn't have was my fixed weekly floor — the number that doesn't move no matter what. The number you pay whether you booked one job that week or twelve.

That number is what a reserve is actually protecting. If you don't know it cold, your reserve target is a guess.

---

The 90-Day Rule Has a Denominator You're Ignoring

Ninety days. Not because it's a round number. Because in 2008, 90 days was what I needed and didn't have.

Most people misapply it. They multiply 90 days by average monthly overhead — materials, fuel, subscriptions, everything. Wrong number. Materials stop when work stops. Subscriptions you can kill. What doesn't flex is payroll.

By spring 2008, I had twelve guys. Weekly payroll burn on wages alone — not benefits, not workers' comp, just wages — was roughly $14,000 a week. I know that now because I had to figure it out by hand in November, at the kitchen table with Diane, working backwards from what was left in the account and counting how many Fridays I could cover.

I should have known it in March.

Your 90-day number is 13 weeks times your fixed weekly payroll burn. That's it. Everything else is negotiable when things get hard. Payroll isn't — not if you want to keep the people you need.

Shops with a working foreman — a guy who runs a small crew and still gets dirty — know their fixed floor without thinking about it. That foreman is load-bearing. He's the last man you'd cut. If you know what it costs to keep him for 90 days, you know your floor. Sign your next truck note after you've got that number covered. Not before.

---

Building Your Real Number

The "three months of expenses" rule is borrowed the same way off-the-shelf flat-rate pricing books are borrowed from some other shop's cost structure. Useful for about a year. After that, a way to stop thinking.

Your number comes from three things you already know.

Your longest receivable stretch in the last three years. Pull your AR aging. Find the gap between when the invoice was due and when it actually cleared — not the average, the worst one. For me in 2008, receivables went from 45-day terms to 110 days on the back end, and some of that money never came at all. That 65-day gap is what your reserve has to cover. Not some abstract monthly baseline. The actual distance between what you expected and what happened.

Your fixed weekly payroll burn. Write it down. If you have to look it up, you don't know it well enough. That number should be as automatic as your home address.

The cost of your last unplanned equipment failure. Not a scheduled service. A real surprise — compressor went, press tool quit, something on the work truck. Whatever it cost, add it as a hard line. Because it will happen again when the work is already slow and the margin is already thin. That's not bad luck. That's just how the timing works.

Your floor is 13 weeks of fixed payroll. Your ceiling adds the stretch cost — longest receivable gap in days, times weekly payroll divided by seven — plus one equipment emergency. That's the range you're building toward.

It's a bigger number than "three months of expenses" gave you. Good. It's supposed to be.

---

What 110 Days Actually Looks Like

It doesn't come at you all at once. It's a sequence.

You stop drawing your own salary first. Tell yourself it's temporary. Tell Diane it's a couple weeks. Then you start timing vendor payments — you know which ones have a grace period and which ones will put you on COD if you're five days late. You're playing the float in ways you never had to before. Then it's Friday morning, you're calling to confirm the ACH is going to clear, and you already know it should, but you're not sure it will.

That was November 2008. Smith Mechanical. Receivables that had been running at 45 days stretched to 60, 75, 90, and two of them stopped moving entirely. I laid off four men. Hired every one of them personally. Kept three. The math on which three stayed was done on paper at the kitchen table. Being wrong was going to cost somebody something.

Then 2011. Whitman Builders out of Marlborough. I was owed $61,000 across multiple jobs. Got "next Friday" for months. Kept working because I didn't want to burn a bridge in a small market. He paid me 38 cents on the dollar in the bankruptcy.

Both situations had the same cause underneath them. I was treating builder receivables as assets when they were contingent liabilities with nothing behind them. A builder's word and a pay-when-paid clause I'd glossed over when I signed.

---

Builder Receivables Are Not Assets

This is the biggest flaw I see in how small shops think about their reserve. They look at the checking account, look at open AR, add them together, and call that the cushion.

That's optimism, not math.

A GC who pays in 60-plus days is using your labor as a free credit line. He's stretching payables as long as his subs will let him. You're letting him. Every time you keep working without collecting, you're extending that line at zero interest with no collateral.

Pay-when-paid clauses make this legal in Massachusetts and most other states — though you want to read your specific subcontracts, because the language varies. If there's a pay-when-paid clause, what you actually have is a conditional receivable. You get paid when the owner pays the GC. If the owner doesn't, you might not get paid at all. Counting that in your reserve calculation is counting money that might not exist.

There's also the lien clock. In Massachusetts you've got 90 days from last work to file a mechanics lien on most residential jobs. Miss that window and the receivable is worth whatever the builder decides to pay you. In a bankruptcy, that might be 38 cents on the dollar. Might be zero.

Strip your open builder receivables out of your cushion calculation. Just take them out and look at what's left. If that number makes you feel sick, that's useful. You needed to feel sick before the next slow October, not during it.

---

Monday Morning

Pull your last three years of AR aging. Find your longest stretch — not the average, the worst one. Multiply that number in days by your fixed weekly payroll burn, divide by seven. That's your stretch cost. Add it to your 90-day payroll floor. Write both numbers down.

Write down your last unplanned equipment failure. Actual cost. Add it as a third line. Keep it separate so you can see it.

Strip the builder receivables from your cushion number. Checking account only. Maybe an insured savings account where the money is actually sitting. That's your real cushion. If you're below your floor, you know what you're working toward. If you're way below it, you have a conversation to have with yourself about the next truck note and whether the builder account you've been nursing is worth the exposure.

Know your fixed weekly payroll burn without looking it up. If you can't, go find it before anything else.

No software required. No consultant. A calculator and an honest afternoon.

---

FAQ

Before I hire my next person, what's the minimum reserve I should have?

Enough to cover that person's wages for 13 weeks with zero new revenue. Not their first paycheck — thirteen weeks of it. If that cash isn't sitting somewhere already, you're betting the work that justified the hire doesn't slow down. Hire slow, and hire with money you've accumulated — not revenue you're expecting the new person to generate.

Every dollar coming in is already spoken for. How do I build anything?

Pick a fixed percentage — five percent of every deposit, three if things are genuinely tight — and move it to a separate account the day it hits. Don't wait to see what's left at month's end. There's never anything left. Keep it at a different bank than your operating account. Out of sight matters. If you can see it easily, you'll spend it.

Does my reserve target change based on how much builder work I'm doing?

Yes. Heavy builder exposure means longer worst-case receivable stretches and more contingent money on your books. A shop doing mostly direct residential — homeowners, card on file, collect at the door — has a tighter, more predictable cycle. If more than 40 percent of your revenue is coming from GCs or builders, your floor should be higher than the baseline. Builder concentration is risk. Price it like risk.

Does my line of credit count as part of my reserve?

No. A line of credit is a borrowing tool. In a real crunch — biggest builder goes dark, two guys need to be paid Friday — banks get nervous at exactly the wrong moment. Lines get reviewed, reduced, frozen. I watched it happen to a shop in my market in 2009. Your reserve has to be money you actually own, sitting somewhere you control, not a promise from a bank that conditions haven't changed.

How often should I recalculate the number?

Any time your crew changes by more than one person. Any time your work mix shifts — you pick up a builder account representing more than 20 percent of projected revenue, or you drop one. And once a year regardless, because payroll changes, equipment ages, and your worst receivable in the three-year window eventually rolls out and gets replaced by something new. Look at the actual numbers once a year. It takes less time than you think.