The Quote That Costs Twice: Why Slow Pricing Loses Jobs — and Wins Them Too Cheap

It’s 9pm and a builder is at the kitchen table pricing an extension. The enquiry came in on Monday, the site visit was Wednesday, and it’s now Thursday night because this is the only quiet hour there’s been all week. The quote runs to one page, the numbers are mostly carried over from a similar job last year, and a few lines are frankly a guess. It will go out tomorrow, four days after the homeowner first made contact. Two other builders quoted on Tuesday.

There are two ways this evening’s work loses money, and most firms only ever notice one of them.

The first is the job that never comes back. Four days is a long time to a homeowner with a project they’re excited about, and the firm that quoted first, while they were still keen, has a real head start. The late quote often loses to nothing more than timing. The second loss is quieter and arguably worse: the quote that wins. Priced from memory rather than worked out, it’s frequently too low, and once the job starts the extras creep in — the bit of extra steel, the wall that turns out to need underpinning, the spec the client nudges upward — and most of them are never re-priced or billed. The firm wins the work and gives away the profit on it.

The job is lost before the quote even arrives

Quoting is the part of the week that gets done last because it can only be done by the people who are busiest. The estimator is also the project manager. The director who prices the work is also the one running the current job. So the quote waits until the evening, or the weekend, or the gap between site visits that never quite appears.

Meanwhile the buyer is making a decision. Someone commissioning bespoke work — an extension, a fit-out, a design, a cleaning contract — is rarely getting one price. They’re getting three, and the longer a quote takes to arrive, the more likely the decision is already leaning toward whoever turned theirs around fastest. The same speed advantage that decides a cold enquiry decides a quote — we wrote about that race in The Enquiry That Went Cold — and it compounds the longer everyone else takes. A slow quote doesn’t just risk the job. It often hands a competitor the advantage of being the only number on the table for two or three days, which is plenty of time for a homeowner to talk themselves into it.

None of this shows up as a loss the firm can see. A quote that’s never returned looks identical to a quote that was simply too expensive. The firm files it under “didn’t win that one” and moves on, never knowing the work was lost to a calendar rather than a price.

And the job that’s won is often won too cheap

The second loss hides inside the wins, which is why it’s so persistent. A quote built from memory and last year’s figures carries two errors that both point the same way. It underprices the work at the outset, because the easiest number to reach for is the one that feels competitive rather than the one that protects the margin. And it has no mechanism for catching what changes once the job is live.

That second part is where the real money goes. On almost any bespoke job, the scope moves. The client asks for an upgrade, the site throws up a surprise, a small “while you’re here” turns into half a day’s work. Each of these is a legitimate variation that should be priced and added to the invoice. In a small firm running flat out, most of them are absorbed instead — agreed verbally on site, never written down, never charged. By the time the final invoice goes out, nobody can reconstruct what was quoted versus what was actually done, so the safe thing is to bill the original figure and eat the difference.

A firm can be busy, well-reviewed, and winning plenty of work, and still be quietly running on a margin several points thinner than it thinks, because the quote was low and the extras were free.

Four versions of the same leak

The trade changes, the shape doesn’t.

Different trades, the same leak. The work that wins and feeds the business gets priced by the people with the least time to price it, in a rush, with no dependable record of what was promised against what was actually delivered.

Why it costs twice

The phrase that fits is “costs twice” because the two losses compound. The slow, manual quoting process loses a share of winnable jobs outright. Then, on the jobs it does win, the same haste and the same lack of a paper trail give away a slice of the margin. A firm fighting on both fronts is working harder than its competitors to stand still: winning fewer of the jobs it quotes, and earning less on each one it wins. It belongs to the same family as the Admin Tax — a cost real enough to shape the year-end figure, yet itemised nowhere, so it never gets argued with.

It’s worth making it concrete with a composite.

What AI can realistically fix in 2026

Quoting is one of the places where the tooling has matured past the demo stage into something a small firm can actually lean on, and the fix splits cleanly along the two losses. (For where AI agents do and don’t fit more broadly, that’s its own piece.)

Fast, consistent quotes from the firm’s own data. An AI set up around a firm’s historical quotes, current rate card, supplier prices and pricing rules can turn an enquiry and a set of site notes into an accurate, properly structured quote in minutes rather than evenings. It applies the firm’s real margins consistently, every time, regardless of who’s asking or how busy the week is. The senior person moves from building the quote from scratch to checking and sending one — which means the quote goes out same-day, while the buyer is still deciding, at a price that protects the firm rather than one reached for under time pressure. It can also flag when a quote’s margin has dropped below the firm’s own threshold, so nothing goes out underpriced by accident.

Variation capture that runs through the job. The second agent’s job is to make sure nothing is given away once work starts. As changes come up — logged from a site note, a voice memo, a photo, an email from the client — it records each one against the job, prices it from the same rate card, and assembles the variations into something that actually reaches the invoice. The “while you’re here” extras stop being free, because there’s finally a system catching them at the moment they happen rather than a memory failing to reconstruct them weeks later.

In both halves the pricing judgement stays with a person: what to charge a long-standing client, when to sharpen a number to win strategic work, how to handle a tricky variation. What the AI strips out is the slowness and the leakage — the late quote and the uncaptured extra — which are the parts costing the money, not the judgement.

Two things to get right

Two things decide whether this actually works.

First, the quote is only as good as the data behind it. The value comes entirely from setting the system up around the firm’s real rates, real suppliers and real rules; a generic estimating bot that doesn’t know the firm’s costs will produce confident nonsense. As with most things AI, the value is in the setup, not the model — the work is in the configuration, not the software.

Second, fast must still mean considered. The aim isn’t to fire out an instant number that wins the job and loses the margin all over again — it’s to produce a quote that’s both quick and right. Speed only helps when the price underneath it protects the firm.

The bottom line

The losses that hurt a small firm most are the ones that don’t show up in the accounts as a single line. A quote that was never returned and an extra that was never billed both vanish into “that’s just how the numbers came out this year.” They’re not. They’re the predictable result of pricing winnable work by hand, in a hurry, with no record of what was promised.

The first move costs nothing. For one month, log every quote — when the enquiry came in, when the quote went out, whether it won, and whether the final invoice matched it. The gap between the enquiries and the quotes that went out fast is the work lost to speed. The gap between the quotes and the final invoices is the margin given away. For most small property-economy firms, that gap is the easiest money in the business to recover, and in 2026 it’s finally recoverable without putting a full-time estimator on the payroll.

Putting a first number on it costs nothing: the free AI Value Calculator gives a quick estimate of what hand-built quoting is costing in lost jobs and thin margins. If you’d rather have someone trace exactly where it leaks — how long quotes take to leave the desk, how far the final invoice tends to drift from what was quoted — that’s the work an AI Strategy & Operations Audit does. And the build itself — a quoting tool and a variation tracker wired to your own rates, suppliers and rules rather than a generic estimator — is a Bespoke AI & Automation Build.