By Greg Allan, tradesman and creator of Data First Coaching
Why do variations go unsigned and unpaid on so many jobs?
Variations go unsigned because the builder’s crew treats them as part of getting the job done, not as a separate contract event. A site supervisor spots extra work on a renovation or extension, tells the subcontractor to crack on, and the paperwork gets chased after the fact, by which point the client has moved on mentally and the builder is left arguing over a job that is already built.
On a fixed price contract, every bit of extra scope, extra footings because of rock, a change to the kitchen layout after frame stage, an extra power point the electrician adds while they are already there, is meant to become a signed variation before the work happens. In practice, most of it happens because stopping the job to chase a signature feels slower than just getting on with it. By the time the final progress claim or the handover invoice goes out, the client is looking at a number they were not expecting, and the builder is explaining extra cost after the work is already in the wall.
The result is the same conversation on site after site: the crew did the work in good faith, the client never agreed to pay for it in writing, and somebody absorbs the difference. On a renovation with a tight margin, that difference can be the whole profit on the job.
What does an unsigned variation actually cost a building business?
The real cost of an unsigned variation is not just one unpaid invoice, it is the gross profit margin on that job quietly disappearing, job after job, without ever showing up as a single clear number the owner notices. Most builders see it as the client being tricky rather than seeing the pattern across the whole business.
A builder running fixed price contracts, cost plus jobs, or a mix of both, relies on the margin built into the estimate to cover overheads and leave a profit. Every variation that gets done but never gets paid for eats straight into that margin, because the labour and material went out the door regardless of who agreed to pay for it. Across a year of renovations, extensions and new homes, those small unpaid variations can be the difference between a business that looks busy and one that is actually profitable at handover.
Because each one feels small on its own, it rarely gets flagged as a problem. It gets written off as part of doing business, when it is actually a leak that repeats on every job where the signed variation step gets skipped.
What number in your own job data shows where variations are breaking down?
The number that tells the real story is not how many variations get raised, it is how many get signed before the work starts versus after, and how that gap moves against your gross profit margin on each job. That pattern, tracked job by job rather than remembered at the end of the year, shows exactly where in the process variations are leaking out of the business.
Memory tells a builder the business is pretty good at variations, because the big ones, the ones that caused an argument at the final progress claim, stick in the mind. The small ones, the extra power point, the slightly bigger slab, the extra day a subcontractor spent because the certifier wanted something fixed before sign off, do not get remembered at all, and that is exactly where most of the leak sits.
Looking at the business job by job, rather than from memory, shows which stage the gap opens at: at quoting, at frame stage, at lock-up, or in the run-up to handover. That is a data first way of looking at the business, because the numbers do not forget the small ones the way a site supervisor or an estimator does.
How can AI agents get variations signed before the work starts?
An AI agent is software that does a job the way you would, using your own prices, your own wording and your own rules, and in a building business it can sit in the exact gap where variations leak, stopping extra work from proceeding until the client has signed off on the price. That means the signature happens before the subcontractor picks up a tool, not after the renovation is finished.
Picture a builder running a kitchen and bathroom renovation where the site supervisor finds rot behind the old shower that was not in the original scope. Right now, the supervisor rings the office, someone drafts a variation, and the job either stalls waiting for signatures or, more often, the crew just gets on with it while the paperwork catches up, if it ever does. With an agent built around how that builder actually prices and words variations, the extra scope becomes a signed variation sent to the client the same day it is found, in the builder’s own language, with the builder’s own margin already built in.
Because the agent does this consistently, every time, it takes the chasing off the site supervisor and the office, and puts the decision back where it belongs, with the client, before the work proceeds. The crew spends their time building, not following up signatures, and the customer experience improves because the client is told the price before they are told the bill.
Should builders build their own variation system or buy an app?
Some off-the-shelf job management tools handle variations well enough for some builders, and the data on how a particular business actually quotes, contracts and claims shows which ones are worth keeping. For others, a system built around their own contract wording, their own prime cost items and their own progress claim schedule fits the job better than software built for every trade at once.
Many builders already run some kind of job management or estimating software, and for straightforward fixed price work with a simple variation process, that can be enough. The gap tends to show up on more complex jobs, cost plus contracts, multi-stage renovations, or new homes with a lot of prime cost allowances, where the standard template does not match how that builder actually structures a contract or writes up a variation on site.
Rather than assuming either option is right, the honest answer is to look at where the current system is losing variations, at quoting, at frame stage, at handover, and decide from there whether an off-the-shelf tool fits or whether a system built on the business’s own data and its own way of doing things closes the gap better. That decision belongs with the builder, not with whichever software they happened to start with.
How does Data First Coaching approach this?
Data First Coaching, built by Greg Allan, starts from your own job data rather than a general view of how variations should work. It looks at your gross profit margin job by job, through the Trade Business Intelligence, to find exactly which stage of the job, quoting, frame stage, lock-up or handover, is where variations are going unsigned and unpaid. That shows whether the real problem is the paperwork, the timing, or how the crew is trained to handle extra scope on site. The how, including where an AI agent is built into that process, is worked through inside the Academy, with other members, because the Trade Business Intelligence is powered by every member and works best with builders who are open minded, innovative and willing to share what they learn.
Looking for a business coach who starts from your numbers? Builder Business Coach or every answer for builders.
