Quick Take
Most software migrations fail for boring reasons. The data moved, the techs adapted, and then in week three somebody noticed that 400 membership renewals had been generating invoices with last year’s price. HVAC migrations add a second failure mode that pest control shops never face: your compliance records and your equipment history live inside the same system you are leaving.
Three datasets make an HVAC migration harder than a general field service migration. Equipment records filed by serial number, which almost no platform imports automatically. Refrigerant logs, which EPA Section 608 requires you to keep for three years regardless of what software you run. And membership billing, where a single missed retry on a declined card silently drops an agreement worth $300 to $600 a year.
The platforms that hold those records are also the platforms that make leaving hardest. ServiceTitan reviewers describe two-year commitments they believed were annual, and the Better Business Bureau record includes a $39,375 early-termination charge. FieldEdge is quote-gated, so the exit terms are whatever the quote says.
This guide covers the sequence: what to pull out of the old system, when in the calendar year to do it, how to run both platforms at once, and the contract timing that turns a six-week project into a fourteen-month bill for two subscriptions.
Before You Start: Deciding Whether to Switch
A migration for a 10-tech HVAC shop eats 60 to 120 hours of owner or office manager time across 6 to 16 weeks. At a $55/hour loaded cost that is $3,300 to $6,600 in labor before the new subscription starts. HVAC runs higher than pest control here because of the equipment register: somebody has to rebuild the serial-number database by hand if the old platform will not export it.
Reasons worth that cost:
- Your shop started quoting from a flat-rate book and the current platform has none. This is the most common trigger in this trade, and it is a real one.
- Summer dispatch has outgrown the scheduling board. If a dispatcher is routing 40 calls with a whiteboard and a phone, the platform is the bottleneck.
- You installed R-454B equipment this year and your refrigerant records are a paper binder in a truck. That binder will not survive an EPA audit.
- Membership count passed the point where the platform handles renewals cleanly. Shops on Jobber and Service Fusion describe trouble past a few hundred agreements.
- The contract is up for renewal and the new terms are worse than the market.
Reasons that are not:
- A demo looked good. Every demo looks good in February with 12 calls on the board.
- One feature is missing and you can live without it. Ask the vendor for a roadmap date in writing. If they give you one, wait for it.
- One technician is complaining. That is a training problem more often than a software problem, and migration does not fix training.
- The price went up. Price increases annoy everyone. Take the increase, multiply it by twelve, and compare it to the migration labor cost above. Sometimes switching still wins. Do the arithmetic before you decide.
Spend one week writing down every specific problem with the current platform, then sort the list into things software can fix and things a process change can fix. HVAC shops routinely find that half the list is data entry discipline, not missing features.
Phase 1: Data Export
Do this before you notify the vendor. Some platforms tighten data access the moment a cancellation is mentioned, and at least one HVAC vendor’s users report being quoted a backup fee only after they asked to leave.
The Equipment Register
This is the export that decides whether the migration takes six weeks or six months.
Every unit your company has ever installed or serviced should exist as a row: model number, serial number, install date, warranty start, warranty expiry, refrigerant type, the customer it belongs to, and the address it sits at. If the old platform stores equipment only as free text in a job note, you are rebuilding the register from invoices, warranty registrations and manufacturer portals.
Budget three minutes per unit for manual reconstruction. A five-tech residential shop with 800 units in the field is 40 hours. A commercial contractor with 3,000 rooftop units and split systems is 150 hours. Neither number is a reason to stay, but both should be on the project plan before anyone promises a go-live date.
Ask the old vendor for the export in one specific form: one row per serial number, with the service history attached to the serial number rather than the address. If the answer is that history attaches to the address, you have found the reason your techs cannot see what was done to a condenser before it was replaced.
Refrigerant Records
EPA Section 608 requires any shop that services refrigerant to keep records for three years, and the A2L transition has made that record longer, not shorter. R-454B and R-32 carry a mild flammability classification, and every unit installed under the new rules adds lines to a log somebody has to maintain.
Export the log in full: date, customer, equipment serial number, refrigerant type, amount added or recovered, cylinder tracking, and the EPA 608 certification number of the technician who did the work. Then check that the export includes the technician certification records themselves, because your techs need current certifications on file and those often live in a different module than the usage log.
The gap to avoid is a period where the old system is off, the new system is live, and neither holds a complete record. Keep the old platform readable until the new one has accepted a full year of usage data. This is worth arguing about with a vendor who wants to shut off access at cutover.
Forms, Price Book and Membership Agreements
The flat-rate price book. Export your task list with prices, your Good/Better/Best tier structure, and any markup or margin rules. If you use an independent book like Profit Rhino or Coolfront, that data lives outside your field service platform and migrates on its own, which is one small mercy. If your book is inside ServiceTitan Pricebook Pro or FieldEdge’s bundled Coolfront library, confirm whether the tasks you customized are exportable or whether you are rebuilding your own pricing from scratch. Custom tasks are where shops get surprised.
Membership agreements. Every active agreement with the customer, the plan level, the annual or monthly amount, the next scheduled visit, the visit frequency, and the payment method token. Payment tokens almost never transfer between platforms, which means every member re-enters a card. Plan the customer communication for that, because 800 members receiving a “we need your card again” email in the same week generates 800 phone calls unless you write the email carefully.
Financials and communications. Two years of invoices, payments and outstanding balances for your accountant. Email and SMS logs with customer contact preferences. The communication history is the one operators skip and later regret.
Export Method and the Fee Question
Start with the platform’s own reporting and export tools. Then contact support with a data audit request, framed as an internal records review rather than a cancellation notice, and ask what a complete export includes and what it costs.
If a fee appears, check whether the standard CSV export through reporting is free. The charge is usually attached to a full database backup, which you may not need if you can reconstruct from CSVs. Offer to sign a release confirming receipt, which removes the vendor’s liability concern. Ask the incoming platform whether they will credit migration costs, since several now do this as a sales incentive. If none of that works, pay it. A few hundred dollars for three years of refrigerant records is not a close call.
Phase 2: Pick the Migration Window
No other trade has to schedule a software project around weather. HVAC does.
Summer call volume runs roughly 3.5 times spring volume. A platform that feels fine at 12 calls a day gets abandoned at 40, and techs abandon it quietly, which means the equipment history you bought the system to collect never gets entered. Winter has the same problem in cold markets, compressed into a shorter window with fewer techs on the road.
| Market | Worst Window | Best Window |
|---|---|---|
| Phoenix, Las Vegas, South Florida | May to September | January to February |
| Texas, Southeast | June to August | February to March, October |
| Midwest, Mid-Atlantic | July to August, December to February | April, September |
| Northeast, Upper Midwest | July to August, December to March | May, September |
Two rules make the calendar work. Finish cutover at least three weeks before the first heat wave or the first hard freeze, so the parallel run happens on a light board. And never schedule go-live for the week after a holiday weekend, because that is when deferred maintenance calls arrive all at once.
If the contract renewal does not line up with a shoulder season, you have a real conflict. Paying two subscriptions for an extra quarter is usually cheaper than migrating in July, and it is worth saying that out loud during negotiation.
Phase 3: Parallel Running
Two systems, live at the same time, for two to four weeks. The reason is narrower than general caution. A demo with 20 fake customers will never expose the bug where the mobile app drops the refrigerant entry on the fifteenth job of a hot day.
Weeks 1 and 2: new work only. New customers, new jobs and new quotes go into the new system. Existing membership visits continue in the old system. Your team learns the new platform on a volume they can actually handle, and if something breaks, the blast radius is new customers rather than your entire recurring base.
Weeks 3 and 4: move the agreements. Export active membership schedules. Import them with the next visit date set to the upcoming cycle. The old system finishes the current cycle, the new one picks up the next. Then watch the first full week of membership visits like a hawk, because a missed maintenance visit is a customer who calls a competitor while you are still troubleshooting.
One Source of Truth Per Data Type
The failure mode during parallel running is a customer who calls to reschedule. Does the office update both systems, or one? If the answer is unclear, staff will update neither, or update the wrong one.
Decide in advance. During the parallel window, the old system is the source of truth for existing membership schedules and historical service records. The new system is the source of truth for new customers, new jobs and all outgoing communication. Write that on a card and tape it to the dispatcher’s monitor. Ambiguity here is what produces a tech knocking on a door for a visit that was cancelled in the other system.
Phase 4: Staff Training
The software is only as good as the people running it, and HVAC training splits cleanly in two because the office and the truck use different halves of the product.
Office first. The dispatcher and office manager configure the platform, import data, build the price book and become the internal experts. Give them three full days before a technician opens the app. They should be able to schedule a call, dispatch it, price a repair, invoice it and run a membership renewal without help.
Technicians second, mobile only. A tech needs six things: see the route, navigate to the stop, price a repair from the book, log a refrigerant charge, capture a photo, and collect payment. Keep the session to 90 minutes. Anything about reporting belongs to the office.
The Session That Works
Skip the conference room projector. Instead, run it at the end of a workday with a device in every hand:
- Walk through one complete call on the new app, using a real customer and a real serial number.
- Have each tech price a condenser changeout with Good/Better/Best options and log a six-pound charge against a real unit.
- Have each tech repeat the flow three times while you watch.
- Let the fast ones teach the slow ones.
- Close with the two-week rule below.
The Two-Week Rule
For the first ten working days, the office manager asks each tech at the end of the day: what worked, what did not, and what took longer than the old system. Then fix it overnight. If pricing a capacitor takes six taps, reorder the price book. If the map is slow, check offline mode.
Feedback quality collapses after two weeks. Techs stop reporting problems and start working around them, and a workaround that survives a month becomes permanent. The first ten days are the only cheap window you get.
Phase 5: Cutover and the 30-Day Audit
Set a hard cutover date and tell the whole company two weeks out, again one week out, and again the day before. On cutover day the old system becomes read-only and every new job, invoice and communication lives in the new platform.
Then run this audit thirty days later:
- Unit count. Compare equipment records by serial number in both systems. A gap of more than 2% means units were dropped during import, and dropped units are warranty claims you will pay for yourself.
- Refrigerant continuity. Pull the usage log for the migration month and the month after. Any gap in the sequence is a Section 608 problem, and it is much easier to patch at day 30 than at an audit.
- Membership revenue. Compare recurring billing for the last 30 days against the same period last year. A shortfall usually traces to agreements that imported without a payment token.
- Ten random customers. Verify equipment list, service history and last invoice match between systems.
- One question per tech. Ask how much they prefer the new app, on a scale of one to ten. Anyone who answers four or below gets thirty minutes of your time, because a tech who quietly reverts to the old system takes your equipment history with them.
Keep the old platform readable for 60 to 90 days after cutover. You will need it for a warranty argument, a refrigerant record from two winters ago, or a customer disputing a charge you can no longer look up. Then time your cancellation notice so the access window and the auto-renewal window do not overlap.
The Contract Trap
The most expensive migration failure in this trade is not technical. It is a contract that renews while you are still importing data.
You plan a ten-week project in March. In week four the old platform auto-renews for twelve months because the cancellation notice window was 60 days and nobody read it. You now pay for two platforms until next spring, and the savings that justified the switch are gone.
Read the old contract before the project starts. Find the cancellation clause. Write down the notice period and the renewal date. Set reminders at 90, 60 and 30 days. If the notice period is longer than the migration, submit the cancellation notice before you begin. You can rescind a cancellation. You cannot rescind an auto-renewal after the window shuts.
This is the argument for month-to-month platforms, and it is stronger in HVAC than in most trades. Jobber and Housecall Pro publish prices and sell monthly. A shop on either can run a migration without a contract calendar hanging over it. That flexibility is why both appear so often in our 1-5 tech recommendations, even though neither ships a native refrigerant log.
Migration Timeline by Platform
| Migration | Difficulty | Typical Timeline | Main Risk |
|---|---|---|---|
| Jobber → Housecall Pro | Easy | 3-5 weeks | Membership billing model differs; rebuild agreements |
| Housecall Pro → FieldEdge | Moderate | 6-10 weeks | Equipment register must be rebuilt at serial-number level |
| Jobber → ServiceTitan | Hard | 12-16 weeks | Price book build-out, implementation fee, 12 to 24 month term |
| Service Fusion → FieldEdge | Moderate | 8-12 weeks | Unlimited-user pricing advantage disappears |
| ServiceTitan → FieldEdge | Hard | 12-16 weeks | Export negotiation plus contract term to run out |
| Workiz → Housecall Pro | Moderate | 4-8 weeks | Decoupling the built-in phone system |
| Commusoft → ServiceTitan | Moderate | 8-12 weeks | Site and asset hierarchy has to be remapped |
| FieldEdge → ServiceTitan | Hard | 12-16 weeks | Coolfront price book portability, implementation cost |
Difficulty tracks two things: whether the destination imports equipment records automatically, and how much of the old contract you have left to serve. Everything else is configuration.
Bottom Line
An HVAC migration is a data project with a weather clause attached. Pull the equipment register and the refrigerant log out of the old system before you say the word cancellation. Pick a shoulder month and finish cutover three weeks ahead of the season. Run both platforms for a month with one source of truth per data type. Train the office for three days and the techs for ninety minutes. Audit at day 30, and keep the old system readable for a quarter after that.
The shops with the worst migration stories share one habit: they moved fast to escape a problem and moved in season to save a quarter of subscription overlap. Both decisions feel efficient at the time. The bill arrives in August, when the membership renewals did not fire, the refrigerant log has a hole in it, and the dispatcher is learning a new board on the busiest week of the year. Slow, out of season, and boring is the version that works.
How I Evaluated
This guide draws on:
- Verified Capterra reviews across ServiceTitan (4.3/5, 339 reviews), FieldEdge (4.2/5, 312), Housecall Pro (4.7/5, 2,741), Jobber (4.6/5, 1,458), Service Fusion (4.3/5, 308), Workiz (4.4/5, 218) and Commusoft (4.7/5, 211), read with specific attention to users describing switching to or away from a platform
- Reported implementation costs, contract disputes and termination charges, including the $39,375 early-termination figure in the ServiceTitan Better Business Bureau record and reviewers who describe signing two-year terms they understood to be annual
- Vendor documentation and published pricing pages, verified September 2026. Pricing for the quote-gated platforms is estimated from user reports and marked as such
- EPA Section 608 recordkeeping requirements, including the three-year retention rule, and the A2L transition from R-410A to R-454B and R-32
- Seasonal call volume patterns in residential HVAC, where summer volume runs roughly 3.5 times spring volume
- Platform-level findings from our pricing breakdown, membership plan coverage and HVAC versus general FSM comparison
Where a figure could not be verified across at least two sources, it is described rather than stated as a number.