The Boom That Is More Dangerous Than the Slump

Why this matters

Ask an owner which year nearly finished them and a surprising number name the year after their best one. Going into a slowdown, everything you do is defensive and obvious: watch cash, hold price, keep the crew busy. Coming out of a boom, every decision that hurts you was made months earlier, felt correct at the time, was applauded by everyone around you, and cannot be unmade. The boom is not dangerous because it ends. It is dangerous because of what it persuades a shop to commit to while it lasts, and because the bill for those commitments arrives on a schedule that has nothing to do with when the work stopped.

Why the exit is the dangerous part

Three asymmetries do all the damage, and each one is invisible while demand is strong.

Commitments made in a boom are fixed; the revenue that justified them is not. A lease, a finance note, a salaried position and a larger premises are all promises to pay in every future month regardless of what that month brings. The demand that made them look affordable is the only variable in the arrangement.

A boom hides bad pricing and bad hiring. When everything sells, an underpriced job still sells. When you cannot find anyone, a marginal hire still gets made. The market stops giving you feedback exactly when you are making the most decisions.

A surge borrows from the future. Work pulled forward is work that will not be there next year. Replacement demand in particular is finite - the systems that failed in the heat wave are not going to fail again next summer - so the trough after a genuine surge sits below the baseline, not at it. This is the same mechanism a downturn runs in reverse, and Deferrable Versus Non-Deferrable Work and What a Downturn Touches derives it in full.

Growth also consumes cash before it produces it, which is a real and separate failure mode and is owned by The Busy but Broke Growth Trap. Read it alongside this one; the two describe different halves of the same year.

The four bills, ordered by when they arrive

The measure is how long after the peak each cost shows up. Note what falls out of sorting them this way: the bills that arrive first are the ones you can still do something about, and the one that arrives last is the one you cannot.

1. Fixed overhead, at the turn - zero to one month. The month revenue falls, every fixed cost added during the boom becomes a larger share of a smaller number. This one arrives immediately and needs no trigger.

2. Hiring quality, two to six months. Work done by rushed hires under rushed supervision comes back as callbacks, warranty work and the occasional claim, and the lag is roughly the time it takes a bad install to declare itself. You are paying for the boom's hiring standard with the trough's labour hours, which are the hours you most need for new work.

3. Pricing discipline, three to nine months. A shop that spent a boom not having to justify a price loses the ability to hold one, in the office and on the truck. The bill arrives the first time you need a firm number in a soft market, and it arrives as a discount rather than as a lost job, so it does not appear in any report you currently run.

4. Equipment and space, over the remaining term - years. The van notes, the lease, the larger shop. This bill is the easiest to overlook month to month and by far the largest in total, and it is the only one on the list you cannot renegotiate by deciding to.

Worked: three years through a surge

Index a normal year's revenue at 100 and use the same structure throughout: labour 32 percent of revenue, materials 25, other direct 8, fixed overhead 25, leaving net profit at 10 percent. Those sum to 100. The figures below are illustrative but the shape is the ordinary one.

Year one, the baseline. Revenue 100, net 10.

Year two, the surge. A regional heat event, ten hard weeks. Revenue comes in at 128. Overtime, three quick hires, and quotes that nobody negotiated pull net margin down to 7 percent, so net is 128 x 0.07, which is 8.96.

Stop there for a second, because that line is the article. The 28 percent bigger year produced about a tenth less net profit than the quiet one: 8.96 against 10, down 10.4 percent. Everyone in the shop experienced year two as the best year they had ever had.

Year three, the hole. The surge pulled replacement work forward, so revenue lands at 94, six percent below baseline. Materials and other direct costs scale with revenue: 33 percent of 94 is 31.02. Labour does not scale at all, so it has to be carried in points rather than as a rate. Baseline labour of 32 points paid six technicians, the nine now on the books less the three the surge added, so a technician is 32 over 6, or 5.33 points a year, and 2.67 a half. The shop keeps all nine through the first two quarters and then sheds the three surge hires: nine for half a year is 24.0, six for the other half is 16.0, so labour lands at 40.0 points, eight above the 32 a baseline year carries. Fixed overhead is no longer 25 points of a baseline year; the boom added six, so it is 31 in absolute terms. Total cost is 31.02 plus 40.0 plus 31, which is 102.02, against revenue of 94. Net is minus 8.02, a loss of about 8.5 percent of revenue.

Three years: 10 plus 8.96 minus 8.02 is 10.94, against 30 for three ordinary years standing still. The boom cost this shop 63.5 percent of the profit it would have made by doing nothing at all - 10.94 is 36.5 percent of 30.

Where the six points of overhead went, and which ones you can take back

Split the six points added in the surge by how long they are locked:

Two of the six points can be removed by a decision. Four cannot, and on their own those four are half of year three's loss: 102.02 less 4 is 98.02 against revenue of 94, so the year loses 4.02 instead of 8.02. Set the crew beside them. Cutting back to six technicians at the turn rather than two quarters into the hole takes labour from 40.0 points to 32 and total cost to 94.02, against revenue of 94, which is a year that breaks even. The crew is worth more than the four points, and the crew is still a decision you get to make in year three.

That is the reason the ordering above matters. The four locked points stopped being a decision the day they were signed, and they are the half of the loss that will still be there next year and the year after. The boom decision that felt smallest at the time - financing two vans over five years rather than renting for a season - is the one still setting the outcome three years later.

The surge shape: a year of demand in six weeks

The storm, the heat wave, the deep freeze. Demand arrives at several times normal for a short window and the trade-specific version of the trap has two extra edges.

The peak is not a market, it is a queue. You cannot serve all of it, so the work you take is a choice, and taking whatever calls first is a decision to let the queue choose for you.

The surge and the hole are the same demand. A regional event that replaces a thousand systems in six weeks has removed those thousand systems from the next several years of replacement demand in your area. The hole is proportional to how much of the surge you served, which means a shop that chased the surge hardest has the deepest trough. That is worth sitting with before hiring for it.

How much to chase, and the ethics of the pricing question inside a declared emergency, is its own decision. See related: Storm and Disaster Demand: How Much to Chase.

Demand you should serve, and demand you should decline

Declining work during a boom feels absurd and it is the single highest-return discipline in this card. Three kinds are worth declining on sight:

• Work outside your normal scope, taken because the phone is ringing. This is where the callbacks in bill number two concentrate, because nobody in the shop has done it a hundred times.
• Work that only closes at a price your own arithmetic does not support, on the reasoning that volume will make it up. Volume does not make up a margin gap, it multiplies it.
• Work that requires a commitment you would not make in a normal year - a lease, a note, a salaried head. If demand has to be permanent for the decision to work, and the demand is a weather event, the decision does not work.

The work worth serving hard is your existing customers, your existing service area, and your existing scope, at your normal price, for as many hours as the crew can sustain. That is the boom's real prize and it is entirely reversible.

The limits worth setting before the phone rings

Write these down now, while nothing is happening, because they are unwritable during a surge.

Fixed cost follows sustained revenue, not peak revenue. A workable default: no new fixed commitment until the higher revenue has held for two consecutive quarters. Tune the number, but hold the shape.

Cap the hiring rate at your supervision capacity, not your demand. One unproven technician per two experienced ones working at any time is a defensible starting point, and the binding constraint is who checks their work, not who answers the phone.

Rent, lease short, or subcontract the surge capacity. Paying a premium for flexible capacity is not a failure to plan. It is buying the right to stop, and in the worked example above the four points the shop could not stop paying are half of year three's loss.

Price holds. A boom is the easiest time to raise a rate and the worst time to lower one, and a discount granted at the peak becomes the price you are held to in the trough.

What a boom being handled well looks like

Check the figure, not the feeling:

• Net profit rose at least in proportion to revenue. In the worked case it did not: revenue plus 28 percent, net minus 10.4 percent. If your best revenue year is not also your best net year, the growth is costing you something and the report will not name it.
• New fixed cost is stated as points of BASELINE revenue, not of peak revenue. Six points of 100, not of 128. Measured against the peak it looks like less than five, which is precisely the arithmetic that makes the commitment feel affordable.
• Callback rate is being tracked through the surge, not after it. Bill number two is invisible for two to six months, so the only way to see it early is to measure it while the work is being done.
• Someone can state the trough assumption out loud. If nobody in the shop has an opinion about what next year looks like after the surge, the shop is planning on the peak continuing.

References

• U.S. Small Business Administration, managing business growth and working capital
• See related: The Busy but Broke Growth Trap (owns the cash-conversion side of growth), Deferrable Versus Non-Deferrable Work and What a Downturn Touches
• See related: Storm and Disaster Demand: How Much to Chase, A Recession Playbook for a Small Shop
• See related: Building a Shop That Survives a Bad Year