Growth Signals

A small business has pricing power when replacing it would cost the customer more than the proposed increase, and the record already scores it: a high win rate on quoted work, a schedule booked out past competitors, real switching friction, and a last increase that produced grumbling but no measurable churn.

Pricing power is a specific, testable condition: your customer's cost of replacing you exceeds the increase you are proposing. Everything else written about it decorates that sentence.

For a $1M-$50M operator, the replacement cost has five common sources. Switching friction: the customer would have to requalify a vendor, retrain habits, or re-integrate systems, and each of those has a price in hours and risk. Verified differentiation: something the customer can check, response time, first-visit fix rate, documented uptime, rather than something you assert.

Scarcity: when every qualified competitor is booked out, the alternative to your price is a wait, and waits cost money. Relationship capital: years of showing up convert into tolerance for one more increase, provided the increase arrives honestly. And contract structure: indexation clauses, renewal terms, and surcharge language that move price without reopening the whole deal.

The record already scores your power; the audit is reading it. Four entries matter. What happened after your last increase: measurable churn, or grumbling followed by renewals. Your win rate on quoted work at the current price. How often price is the stated reason in your loss log, with the caution that buyers name price when the real reason is harder to say. And your lead time: a schedule booked out six weeks is your market's bid for your capacity, entered on your behalf.

Return to the HVAC owner. He wins eight of ten bids while both competitors are booked six weeks out, and he has held his install prices for two years because raising them felt risky.

On the audit above he holds clear pricing power and has been donating it: the win rate says the market clears above his number, the lead times say the alternative to him is a wait, and the two flat years say nobody has tested any of it. His risk sits inverted from where he feels it: the dangerous move is the one he is already making.

The opposite profile deserves equal honesty. A machine shop feeding one OEM through a procurement portal, quoting against three qualified suppliers on identical spec sheets, holds little power at the moment of quote, and pretending otherwise burns relationships. The moves there run longer: niche depth the spec sheet cannot capture, tooling the competitors would have to build, delivery reliability procurement learns to price in. Power of that kind is constructed over years and spent in minutes, which argues for spending it deliberately.

The test itself has mechanics. Pick a segment where the downside is bounded: new quotes only, one service line, one region, never the whole book at once. Move enough to read; a 1% test drowns in noise, while a move around 5% prints a signal.

Set the read window before you move: two months of close-rate data against the prior quarter, gauged by the pipeline numbers from the first section. And write down in advance what response sends you back. An honest test names its own retreat before it starts, which is what separates testing from drifting.

One decay law ties the two profiles together. In a cost regime like the current one, with producer prices up 4.66% on the year (July 2026, FRED, chart in the next section), a price held flat is a real price cut that never got announced. Power unexercised in that environment does what unexercised options do. Test small, test often, read the response.

Power you never test quietly expires.

Pass cost increases promptly, visibly, and ranked by source. Externally documented costs, tariffs and fuel and indexed commodities, pass as named, dated line items tied to a source both sides can check, with notice and an effective date. Internally sourced increases, your own wages and margin repair, pass on the strength of your position.

The chart above is this page's thesis drawn from public data. The upper line is what producers receive (PPI final demand); the lower line is what consumers pay (CPI). Producer prices have run ahead of consumer prices in every monthly print since November 2025, and the gap has held above a full point since April (FRED, July 2026 prints).

Between those two lines sits every operator who buys upstream and sells downstream, which is to say every operator. The vertical distance is margin leaving somebody's business each month, and the pass-through decision is the question of whose.

Watch how a cost increase actually travels, because each link has a delay that flatters inaction. The vendor letter arrives, and your unit cost steps up on a date certain. Your quote either moves or it does not; that decision is yours alone and usually gets made by default.

If the quote moves, your win rate responds over the following weeks, at whatever your market's true elasticity turns out to be. And if the win rate moves, your mix shifts, because the price-sensitive slice of your book leaves first, which quietly changes the average customer you keep. Absorbed increases feel free for a quarter; they are on payment terms.

The DTC brand eating a freight increase is the clean case of the invisible version. The carrier reprices, landed cost per unit steps up, and the retail price stays fixed because no invoice the customer ever sees has changed. Nothing announces the compression; it surfaces at quarter close as a gross margin percentage nobody chose. The instrument for that brand is landed cost per unit, tracked monthly against selling price, so the pass-through decision gets made on a date instead of discovered on a statement.

Now the worked math, with round numbers built for reruns. An illustration, and only that: a distributor sells at $100 a unit, carries $60 of unit cost, and moves 1,000 units a month, so contribution runs $40,000. The vendor letter announces 10%. New cost, $66.

Absorb it entirely and the margin drops to $34: contribution falls to $34,000, meaning a 10% cost move becamea 15% contribution cut. That gearing is the part operators consistently underweight: the increase lands on the whole cost base and comes out of the thin margin slice.

Passing it through has its own trap. Six dollars on a $100 price is a 6% increase. Suppose the owner's honest guess, and it can only ever be a guess, is that each 1% of price costs half a percent of volume. At 6%, volume slips 3% to 970 units, each earning the restored $40 margin: $38,800.

Passing every dollar still ends $1,200 a month short of the old contribution, because the lost units were earning margin too. On these numbers, restoring the old dollars takes about a 7.6% increase, above the dollar match. The bar chart below carries the three scenarios; the calculator in the next section runs the same arithmetic on yours.

Tariff costs deserve their own paragraph because they are the easiest pass-through you will ever run, and operators still fumble them by hiding them. A tariff is external, documented, and checkable, which makes it surchargeable: name it as its own line item, cite the schedule it comes from, state the effective date, honor quotes already outstanding to their expiry, and commit in writing to moving it both directions if the tariff changes.

The fuel-surcharge structure in the Operator Economy Watch is the template: a public baseline, stated bands, a published index both sides can verify. A silent list-price rise spends your credibility. A named surcharge spends the tariff's.

The same logic ranks all your increases. Externally sourced, documented costs (tariffs, fuel, a commodity input with a public price) pass with the least friction, so pass them promptly and visibly. Internally sourced increases (your own wage decisions, your own margin repair) pass on the strength of your position, which is what the previous section audited. Operators who blend the two into one vague annual number give up the easy half to protect the hard half.

One more reading habit after any reprice: watch mix alongside volume. Departures after a fair increase cluster at the price-sensitive end of the book, so the average account you keep improves even as the count dips. The accounts that remain tend to quote easier, pay faster, and stay longer, and the capacity the departures free up gets resold at the new rate. Volume is the number that stings in the week after the letter. Mix is the number that decides the year.

Costs arrive by letter. Margins leave in silence.

Five inputs you can pull from your books tonight: price, unit cost, the increase from the vendor letter, monthly volume, and your own elasticity estimate, labeled as exactly that. The outputs are the decision: what absorbing costs you per month, the break-even increase that restores today's dollars once volume responds, contribution under no, half, and full pass-through, and where the contribution-maximizing price sits under your estimate.

Three findings recur when operators run it. The absorber discovers the bleed already started: the monthly shortfall at zero pass-through has been accruing since the vendor's effective date, and the calculator only names it. The matcher discovers that passing the exact dollars still loses money once volume responds, which is why the break-even output sits above the dollar match.

And the operator with a strong position discovers headroom: at a low honest elasticity, the contribution-maximizing increase runs well past the cost-recovery number, which turns a defensive reprice into a pricing decision worth making on its own schedule.

What to do with the output: the break-even number goes into the reprice letter, and the letter goes out on the timeline the triggers section lays out. The elasticity field is the weakest input by design, so run it at half your guess, your guess, and double it, and let the number that survives all three carry the decision.

Raise prices when a trigger fires, and on a calendar even when none does. The triggers worth writing down: a vendor increase you cannot offset inside a quarter, a quarter of closing well over half your quoted work, a schedule booked out past six weeks, and twelve months since the last review.

Most operators reprice on pain: the quarter margin visibly breaks, the increase goes out in a hurry, and the size of it startles customers who heard nothing for three years. The alternative is a standing trigger list, checked monthly in the same sitting as the pipeline numbers from the first section. When a trigger fires, the review happens; the review, run through the calculator above, decides whether the letter goes out.

Cadence beats amplitude. Small and regular clears the same ground as rare and large at a fraction of the relationship cost, because customers price the surprise as much as the number. The catch-up trap is the usual failure: three flat years while your costs ran with the producer-price line on the chart above, then an 18% correction that reads as betrayal.

The same ground covered at 5% or 6% a year reads as weather. Every year you skip is a loan from your own margin at compounding interest, and the repayment letter is always larger than the increases you avoided.

The letter itself has a discipline. State the change, the effective date, and the thanks; skip the apology paragraph, because a letter that argues invites an argument. Give notice that matches your customers' own planning cycle, thirty days for transactional work, longer where budgets get set annually.

Honor outstanding quotes to their stated expiry, which is a reason quotes should carry expiry dates at all. And sequence by segment: new customers first, since they have no reference price; the middle of the book at renewal; the anchor accounts by conversation, because that is where the dollars and the risk both concentrate.

Two cross-checks close the loop. A capacity trigger is also a hiring question: booked out six weeks can mean the price is low or the team is small, and the arithmetic for the second answer lives in the People & Ops guide. And a margin defended is a covenant defended: contribution is the numerator of the coverage math your bank runs, so a well-timed reprice is also balance-sheet maintenance, per the Capital for Operators guide. Pricing decisions rarely stay inside the pricing column.

The mistake, stated first: operators price discounts against revenue, and discounts come out of margin. Revenue is the big number; margin is the thin slice that keeps the lights on, and the discount takes its whole bite from the slice.

The arithmetic, as a labeled illustration with round numbers. At a 30% gross margin, a $100 sale carries $30 of margin. A 10% discount takes the price to $90 and the margin to $20: one tenth off the price removed one third of the margin. For contribution to hold flat, unit volume has to rise 50%, the ratio of 30 to 20.

At a 40% margin, the same discount needs 33% more volume; at a 20% margin, it needs 100%. The general form is m over (m minus d), old margin over discounted margin, and running your own margin through it is the fastest cure for round-number discounting on record.

Now the asymmetry that makes pricing courage cheap. The same arithmetic runs the other way: a 10% increase at a 30% margin holds contribution down to a 25% volume loss, p over (m plus p). So the increase tolerates losing a quarter of the book while the discount demands finding half a book of new volume, and yet the discount feels safe and the increase feels reckless. The feelings have it backwards, and the pipeline numbers from the first section are how you check which way your market actually responds.

A discount is a purchase. If you cannot name what it bought, it was a gift.

The standing rule of the discount ledger

The B2B discipline follows from the rule. Every discount buys something specific, named in writing: a volume commitment with a clawback if the volume misses, payment inside ten days, a multi-year term, a referenceable logo, delivery moved into your slow season.

Unconditional discounts train procurement, whose whole job is to get last year's exception written into this year's baseline. Quarter-end discounting deserves its own warning: close enough deals at a discount in the last week of the quarter and your best customers learn the calendar, at which point your deadline has become their discount.

The ladder needs a floor and an owner. Who can concede what, written down: the salesperson's five points, the manager's five more, the owner's final say past that, because stacked small concessions become the 20% nobody approved. The floor is the number below which the work does not carry its own cost, computed from the margin math above, and the floor holds even at quarter end, when it matters most.

Watch the anchor-account trap, where all of this compounds. The oldest, largest customer, priced two regimes ago, often consumes a third of capacity at the book's worst margin, and the relationship makes the numbers feel impolite to run. Run them anyway: volume at the real margin, against what the same capacity earns at the current book rate. The fix is a dated glide path agreed in conversation, and the arithmetic is the preparation for it, because a specific number moves a negotiation that a feeling cannot.

And when a customer needs the number to fall, concede scope before rate. Same price for a reduced package keeps your rate intact and your reference price unbroken; a reduced price for the same package resets both. The customer gets a smaller bill either way. Only one version costs you the next negotiation.

Terms are discounts wearing a calendar. Net-60 granted casually is a price cut equal to the cost of financing sixty days of your customer's working capital on your own line, and it never appears in the discount report.

Price the calendar like money, because it is money: a customer asking for longer terms is asking for a loan, and a customer paying in ten days has earned a real concession. The operators squeezed hardest by the pass-through gap on the chart above are usually financing their customers on both ends, price and terms at once.

Every discount is a price customers remember.