Leaving the spreadsheet without losing a week

Most software migrations in small service businesses fail the same way. The owner decides to move everything at once, over a weekend, in the busy season, and by Wednesday the crews are running on paper again because the schedule was half-entered and nobody trusted it.

The fix is boring: move in an order that matches how the work flows, and accept that some of your history is not worth carrying.

Move in this order

1. Clients and properties

Everything else hangs off these. Export whatever you have, whether that is a spreadsheet, the contacts in your phone or the customer list in your accounting software, and clean it before it lands anywhere new.

The cleaning matters more than the moving. Duplicates created now will haunt every report you run for the next three years. Sort by address, not by name; the same customer appears twice under "Bob Whitfield" and "Robert Whitfield" far more often than anyone expects.

A property is not a customer. A property manager with fourteen buildings is one client and fourteen sites, and getting this shape right on day one saves an enormous amount of retrofitting.

2. The price book

This is the highest-value thing you will move, and the one most people skip because it feels like data entry.

You are not transcribing prices. You are deciding, for the first time in writing, what each task costs and what it sells for. Half the shops that do this discover two or three tasks they have been selling below cost for years.

Start with your twenty most common tasks. They are probably 80% of your revenue. The long tail can be added as it comes up. The quickstart walks through the import.

3. Open work only

Do not migrate finished jobs. Migrate the work that is currently live: scheduled visits, open quotes, unpaid invoices. Everything else stays where it is, readable, for as long as you want it.

This is the step people get wrong. Carrying five years of closed jobs into a new system costs a week of effort, makes search worse, and gets referenced approximately never. If you need a historical job, you can still open the old spreadsheet.

4. Recurring agreements

Once clients, sites and the price book are in, set up your maintenance agreements and let them generate their visits. This usually reveals that two or three agreements have been under-served, which is uncomfortable and extremely worth knowing.

Run both systems for exactly one pay period

Not three months. One pay period.

Parallel running is a safety net that turns into a hammock. Pick a date, run both systems until the first payroll clears, compare the hours and the invoices, fix what disagrees, and then stop updating the old one completely, including the wall calendar.

Announce the stop date to the crews before it arrives. The single biggest predictor of a failed migration is a dispatcher who keeps a private spreadsheet "just in case", because everyone else eventually has to check both.

What to tell the crews

Techs do not care about your migration. They care about three things:

• Will I know where I am going tomorrow?
• Will my hours be right on Friday?
• How many taps does it take to close a job?

Answer those three, in that order, in one short session with the phone in their hand. Do not run a training course on features they will never open. The office needs the tour; the field needs the four screens they will actually use.

Week one is slower. Say so.

Productivity dips for about a week. Pretending otherwise makes people think the tool is broken when they are simply new at it. Tell everyone that week one is slower, that week three will be faster than the old way, and then make sure week three is actually faster by removing whatever is still being double-entered.

What you get for the disruption

The point of leaving the spreadsheet is not the software. It is that the same fact stops existing in four places. A window changes once, and the customer, the crew and the invoice all learn about it. A price changes once. A note about a dog at 14 Alder Street is on the visit card forever, instead of in the memory of a tech who left in March.

That is worth one slow week. Migration is included on every paid plan, and the free plan is a real account rather than a trial, so the first week costs nothing but the week.