By Greg Allan, tradesman and creator of Data First Coaching
Why does the gross profit figure in my cleaning business never match what I expect?
Because a cleaning business is rarely one business. It is several, running under the same roof, and most software quietly blends them into a single number that nobody can trust. The gross profit figure on the P&L stops meaning anything once that happens.
A recurring commercial clean runs on rostered hours and a steady list of consumables. A bond clean or end of lease job runs on a fixed quote against a fixed scope, done once, often against the clock. Carpet and upholstery work runs on equipment hire or ownership and chemical cost per site. Insurance restoration work on water damage, fire damage and mould runs on specialist drying equipment, longer site time, and a cost structure that looks nothing like a weekly office contract. When labour hours or chemical and equipment costs from one of these gets coded against another, the margin on each job type becomes a guess dressed up as a report.
Why doesn't one piece of software fit everything a cleaning business runs?
It doesn't, and chasing the one system that will do it all can become its own job. Most cleaning business software is built for one kind of work, a roster tool for recurring contracts, a quoting tool for bond cleans, a job tracker for restoration, and getting them to talk to each other usually means an owner manually re-entering the same job somewhere else.
That manual reconciling eats the hours an owner should be spending on quality checks, on strata managers, on the insurance assessors who send the water and fire jobs through. Many cleaners accept it as the cost of doing business, because the alternative looks like starting from scratch. But every hour spent copying a job from one spreadsheet into another is an hour the business paid for twice.
What should cleaning business software actually keep separate?
It should keep gross profit margin visible by job type, checked at the moment you quote, not worked out months later when the bank balance and the P&L have already drifted apart. If a bond clean, a carpet job and a mould remediation all land in the same undifferentiated bucket, you cannot tell which one is actually paying the wages.
That means labour hours tagged to the job type they were worked on, chemicals and equipment costed against the site they were used at, and a contract's recurring margin tracked separately from a one-off restoration job's margin. Once you can see that at the point of quoting, you quote differently. You stop pricing a water damage job the same way you price an office clean, because you can finally see they never cost the same to run.
What happens when AI agents take over the office work in a cleaning business?
An AI agent is software that does a piece of the office work the way you would do it yourself, with your prices, your wording and your rules, instead of a person retyping the same job details into three different places. For a cleaning business, that means the roster entry, the chemical cost, and the quote all carry the right job type from the first time they're entered, not after a monthly tidy up.
AI agents can sit at the point where a job is booked and ask the question a busy cleaner never gets time to ask: is this a recurring commercial clean, a bond clean, a carpet job or an insurance restoration job, and is the cost being recorded against the right one. On sites the owner never visits, an agent can also hold the quality check and the schedule together, so a missed step on a restoration job or a late crew on a contract site shows up before the customer has to raise it. This is data first, the business runs on what actually happened on site, not on what the owner remembers happening.
Should I buy ready-made software or build systems around my own jobs?
Both can work, and the honest answer depends on what your own data shows is actually breaking down. Some off-the-shelf tools are genuinely fit for a cleaning business, particularly for simple recurring contract rosters, and there is no need to replace what already works.
Where it usually breaks is at the join between job types, where a bond clean, a carpet job and a restoration job all need to sit in the same set of books but behave completely differently. Picture a cleaning business owner running commercial contracts, bond cleans and insurance restoration work side by side, whose rostering software never opened up its own data for six months despite repeated requests. Rather than keep waiting, the owner started building the business its own system, shaped around how the jobs actually run, with the dashboards already working. The plan is to run it alongside the old software, move the crew across gradually, and switch the old one off once everyone is comfortable using the new one. That is a choice some owners can make once they can see exactly where the old system was costing them visibility, never a promise that building beats buying for everyone.
How does Data First Coaching approach this?
Data First Coaching, built by Greg Allan, starts from what the owner's own job and finance data actually shows, not from a description of how the cleaning business runs. For a cleaning business, that means looking at gross profit margin by job type first, recurring contracts against bond cleans against restoration work, to find where the real number stops matching the story the owner has been telling themselves. Greg has described this as working through the research into finding that perfect balance of systems, because a cleaning business built around several job types rarely settles on one answer straight away. The how, including what gets built and in what order, is worked through inside the Academy, with members, because the Trade Business Intelligence is powered by every member and the Academy is built for owners who are open minded, innovative and willing to share what they find.

