Rolling Out Job Management Software to Your Crew

Why rollouts fail, and it's rarely the software

The failure pattern is almost always identical. The system goes live before it's set up, so the first person to open it finds an empty shell and no rates in it. The crew are told about it in a two-minute conversation on site. Nobody takes the old way away, so the dockets keep coming, and now you're running two systems and getting the benefit of neither.

Six weeks later the data is half in and half out, which is worse than either. You can't trust a report, the invoices still need reconstructing on a Sunday, and the honest conclusion looks like the software was a dud. It wasn't. Nothing was ever fully entered into it.

Build it before anyone else sees it

Give yourself a week or two on your own with it before a single other person logs in. The crew should meet a system that already knows your clients, your prices and your terms — not a blank one they're expected to fill in while standing in a roof cavity.

1. Get your client list in. Export contacts from wherever they live now — your phone, your accounting software, a spreadsheet — and import them. Do this first, because everything else attaches to a client.
2. Build the rate library. This is the asset that makes the whole thing worth having. Take your last twenty or thirty jobs, pull out the line items that keep recurring, price them at current supplier cost, and save them. Nothing else you set up returns as much time.
3. Write your terms once. Quote validity, exclusions, deposit and payment terms, your variation process, licence and ABN details. Put them into the templates so every quote and invoice carries them without you thinking about it.
4. Connect your accounting software. Link Xero or MYOB and push one real invoice through end to end before you rely on it. Check the account codes and GST land where your bookkeeper expects them.
5. Decide who sees what. Set up users and permissions before the crew arrive, not after. Most owners want the field crew seeing job detail and their own hours, and not seeing costs, margins or what anyone else earns.
6. Run three real jobs through it yourself. Quote, schedule, complete and invoice three genuine jobs before you involve anyone. You'll find the five things that are set up wrong, and you'll find them without an audience.

Switch it on one stage at a time

Turning everything on at once means everybody's job changes on the same Monday. Roll it out in the order below and each group only has one new thing to learn at a time.

Stage one is deliberately invisible to the crew. It's the stage that pays for the software, it's entirely within your control, and it gives you a fortnight of using the thing daily before you ask anyone else to change what they do.

Stage three is the one that decides whether this works, because it's the only stage where somebody other than you has to change a habit under time pressure. Keep the ask small: for most field crews there are three or four actions that matter, and everything else can wait until those are automatic.

Train three audiences, three different ways

You didn't learn your trade in a day and nobody learns a new system in a toolbox meeting. The mistake is training everyone together, because the three groups need almost nothing in common.

• You. Learn all of it, properly, and first. You're the one who'll be rung at six in the morning by someone who can't find the job, and if you don't know the answer the whole thing stalls that day.
• Your office person. They need the workflow end to end — scheduling, ordering, invoicing, the accounting sync, and what to do when something doesn't match. They'll use it more hours a day than anyone. If they aren't sold on it, it will not survive.
• The field crew. They need the shortest version that exists: find my job, read the notes, log my hours, add photos and materials. Four things. Don't show them a report they'll never open.

Train on a real job, not a demo. Sit with each person, open a job that's actually happening this week, and have them do it themselves while you watch — not watch you do it. The difference in what sticks is enormous.

Pick one person on the crew who's naturally good with a phone and make them the go-to. Half the questions get answered on site without reaching you, and the crew ask a workmate things they'd never ring the boss about.

The pushback you'll get, and what's behind it

Almost none of the resistance is about the software. It's worth knowing what each objection usually means, because the fix is different in every case.

That last row is the common one. Nobody argues, everyone nods, and the jobs are still being written on the back of an invoice. It only shows up if you're looking at what's actually in the system each week.

Name the date the old way stops

Running the old way alongside the new one for a few weeks is sensible. Running it indefinitely is how rollouts die, because as long as the docket book still works, the docket book will be used every time someone's in a hurry.

• Pick a cutover date before you start and tell everyone what it is.
• After that date, no job gets invoiced unless it's in the system — including yours.
• Let jobs already underway finish the old way rather than migrating them mid-build.
• Say the rule plainly: if it's not on the job, it didn't happen and it doesn't get paid — for hours, materials and variations alike.
• Take the docket book away on that date. Leaving it in the ute is leaving the old system running.

Expect the fortnight either side of cutover to be the worst of it. That's normal, it's short, and it's the reason to do this in a quiet stretch of the year rather than the week before Christmas.

How to tell whether it's actually working

Don't judge it on how people say they're finding it. Check five things at thirty, sixty and ninety days — all of them are visible in the system itself.

• What share of jobs have hours logged against them. The single best measure of field adoption.
• How many quotes were built from the rate library rather than typed from scratch.
• Days from job complete to invoice sent. This is the number that pays for the subscription, and it should fall fast.
• How many invoices were raised outside the system. Should be zero after cutover.
• How many jobs have photos or notes on them. Cheap to check, and it tells you whether the crew have opened the app on site at all.
• Whether you can answer what a job made without going through the bank statement.

If a number is bad at thirty days, it's a training or setup problem and it's fixable. If it's still bad at ninety, something structural is wrong — usually the crew are being asked to do more on a phone than the job allows, or a step still requires typing something twice. Find that step and remove it.

Frequently asked questions