How to Evaluate HVAC Software Without Getting Sold

A demo is a piece of theatre. It is built to show a product at its best, in a scenario the presenter chose, with data the presenter prepared. That is not dishonest — it is just not an evaluation.

Last reviewed: August 2026 Next review: August 2027

The five stages, in order

Stage 1: Write your requirements before you speak to anyone

This single step changes every conversation that follows. Without it, each vendor defines the criteria for you — and naturally defines them as the things they happen to do well.

Do not write a feature wish list. Write problems you currently have, ranked, with the cost of each. Something like:

• Invoices go out 3–5 weeks late — roughly $18,000 tied up at any moment
• Nobody tracks maintenance agreement renewals — we lost 40 last year
• Two technicians quote the same repair differently
• I cannot see margin by job type

Then split them into three groups: must fix, should fix, and nice to have. Most shops find they have three or four genuine must-fixes and a long tail of things they simply noticed during a demo.

Stage 2: Shortlist three, not seven

Two platforms in the tier that fits your headcount, and one from the tier above so you know what you are declining and why. More than three and every demo blurs into the others.

How the tiers break down by team size is covered in what HVAC software costs. The short version: buy for your headcount in 24 months, not today, and not for the shop you hope to be in five years.

Stage 3: Bring your own scenarios

This is where you take control of the demo. Send three scenarios in writing, before the call, and ask for them to be run live in the product rather than described.

Three that expose the most:

The emergency insertion. “It is 11am on a full Tuesday in July. An agreement holder calls with no cooling. Show me the whole sequence — who you would assign, what happens to the displaced job, and what the affected customer receives.”

This tests dispatch, re-optimisation, priority rules and customer communication in one go. A product with a real dispatch board handles it fluidly. A calendar with colours does not, and you will see the difference immediately. There is more on what to watch for in how a dispatch board works.

The cold estimate. “Here is a $12,000 replacement proposal sent three weeks ago with no response. Show me what the system has done about it, and what it will do at day 30 and day 60.”

This separates a genuine pipeline from a list of pending estimates.

The management question. “Show me gross margin by job type for last quarter, and close rate by technician.”

Ask them to build it live rather than show a prepared dashboard. If it takes fifteen minutes and three exports, that is your answer about reporting.

Stage 4: The trial is the actual evaluation

Everything before this is research. The trial is where you learn the thing that decides success: whether your field team will use it.

Four rules:

Pick your two most sceptical technicians, not your most enthusiastic. The enthusiast will make anything work. The sceptic will find what is actually wrong with it, which is the information you need.

Run real jobs for a full week. A single day tells you nothing — day one is always awkward. By day four you find out whether it is getting easier or not.

Test the worst conditions on purpose. A crawl space with no signal. A mechanical room in a basement. A rural address at the edge of your territory. Ask specifically whether the app caches work offline and queues updates, then verify it rather than accepting the answer.

Let the technicians report without you in the room. A technician will not tell the owner that the app is confusing while a salesperson is standing there.

Stage 5: Reference calls that produce something useful

Any vendor will supply happy references. The trick is what you ask them.

Useless question: “Are you happy with it?” They are a reference. They will say yes.

Questions that work:

• “What went wrong during implementation?” Everyone has something. A reference who claims nothing went wrong is not telling you much.
• “How long until your technicians stopped complaining?” The honest answer is usually four to eight weeks, and it tells you what to brace for.
• “What do you still do in a spreadsheet?” This reveals the platform’s real gaps faster than any feature comparison.
• “What would you do differently?” Usually the most valuable answer you will get all week.
• “What did it actually cost in the first year, all in?” Subscription, processing, implementation and internal hours.

Also try to find one reference the vendor did not give you — a contractor of similar size in a trade group or online community. Not to catch anyone out, but because a vendor-supplied reference and a randomly-found one usually emphasise different things.

Questions vendors would rather you did not ask

None of these are traps. They are just the questions that get skipped, and each one has cost somebody money.

Red flags

Pressure to sign before a deadline. A discount that expires Friday is a sales technique, not an opportunity. Any platform worth a multi-year commitment is worth a month of evaluation.

Reluctance to run your scenarios. If a vendor deflects your scenario back to their standard demo flow, that is information about the product.

No trial, no product access before signing. Some platforms require a demo before showing you anything. That is a legitimate model, but it means you are committing on the strength of a presentation — factor that into your timeline and your caution.

Vague answers on cost. “It depends” is reasonable for a first call. By the second, you should have written numbers.

Every answer is yes. A vendor who says their product does everything you ask, without a single limitation, is either not listening or not telling you. The most trustworthy answer in any demo is “no, we do not do that well.”

Frequently asked questions

This guide describes an evaluation process based on practices documented across industry sources. We do not test software and we are not affiliated with any platform. Vendor policies on trials, contracts and pricing vary and change — confirm current terms directly before committing.