What we learned shipping admin software to tradespeople
Short answer: nearly everything we got wrong came from designing for the moment when someone sits down to do admin. That moment does not reliably exist. What exists is twenty seconds in a van, and building for that changes almost every decision you would otherwise make.
Who is writing this
Green & Home Ltd is a UK trading business that builds software for its own operations and then sells it — Linnworks tools, quote-chasing tools, and a trade counter system. We wrote about how that happened. GraftG came out of the same instinct applied to a different problem: trade admin, over WhatsApp. It is in beta now, and these are the notes so far rather than a victory lap.
What did not survive
That people would install an app. They will. They just will not open it on the day it matters. An install is a decision, a login is another, and both sit in front of any value being delivered.
That structure up front was helpful. We assumed people would want to create the customer, then the job, then log against it. Tidy, and wrong. On site you have a fact — forty-one miles, £480 of parts — and no appetite for deciding where it belongs. The software should accept the fact and work out the filing afterwards. Demanding structure at the point of capture simply means nothing gets captured.
That mandatory fields produced better data. They produce abandoned entries, which is the worst possible outcome: the time was spent and nothing was recorded. A partial record you can finish later beats a perfect record nobody made.
That people would read anything. Onboarding, help text, tips. If a product needs explaining to someone who has been on site since seven, the product is the problem.
What did survive
Being where the conversation already is. The customer thread already holds the address, the photos and what was agreed. Working in the same place as that conversation removes a whole category of friction we had not even counted as friction.
Speed of fix over breadth of feature. With a small beta group, fixing the thing someone mentioned this morning by this afternoon does more for the product than any feature we could have planned.
Showing the working. Every figure the software produces is visible and correctable. We thought of this as basic decency; it turned out to be the thing that makes people willing to trust an early product with money at all.
Saying no to adjacent users. Agencies, landlords and other software companies asked. We said no, because the product is shaped around someone on site all day, and widening the user base would have widened the design until it fitted nobody. Being clear about who it is not for has been more useful than any targeting.
The uncomfortable bits
Free text is genuinely ambiguous, and sometimes the software has to ask a follow-up question that a well-designed form would never have needed. Message-based admin is excellent at capture and poorer at browsing — there is no dense overview screen, and for anyone coordinating several engineers that is a real limitation rather than a quibble.
And we have rough edges. Our own offer page says so: it is early, you will hit them, tell us on WhatsApp and we will fix them fast, usually the same day. We would rather lead with that than have someone discover it in week two.
What we would tell anyone building for the trades
Design for the worst twenty seconds of the day, not the best. Make the cost of capture lower than the cost of remembering, because your real competitor is someone’s memory and their memory is free. Take feedback from people with genuine volume — five or more jobs a week — because low-volume users will tell you everything is fine when it is not. And measure capture rather than engagement: the only question that matters is whether the thing that happened got recorded.
Most of this generalises past the trades, which is roughly the argument in why the first new-tool project usually fails: tools do not fix processes, and a tool that does not fit the moment of use never becomes a process at all.
Where it is now
GraftG does quotes, invoices, job tracking, mileage, receipts and a diary over WhatsApp, with no app to install. It is in beta, there are 21 Founding Tradies places — twelve months free on the Pro plan with no card, then half price for life — and what we ask back is two honest lines after a month about what changed.
If you are a plumber, gas engineer, electrician, builder or decorator with real admin, that feedback is the thing we are short of and the thing the product is built from. Normal pricing, when the free period ends, is at graftg.co.uk/pricing.
Questions, or want a tool we don't have yet? Email hello@grafto.co.uk — a real person replies.