Software for trade businesses: what actually matters

Most trade software is sold on features. Almost none of them are the thing that decides whether it works for you. Here is what to judge instead.

Every trade software vendor sells you a feature list. Scheduling, invoicing, quotes, a mobile app, GPS, reporting. They all have roughly the same list, which tells you the list is not the thing that matters.

After building these systems and, more usefully, after sitting with operators who are unhappy with the one they have, the things that decide whether software works come down to about five questions. None of them appear on a pricing page.

1. Does the job flow all the way through, or does it stop?

Follow one real job from the enquiry landing to the money hitting the bank. Count how many times a human re-enters something that already exists somewhere else.

Every re-entry is a leak. Most trade software handles the middle bit brilliantly and then stops just before the part that costs you time. Very common example: the app captures everything on site beautifully, and then invoicing is a separate exercise where somebody reads the job and types an invoice.

The question is not "does it do invoicing". It is "does the invoice build itself from what the crew actually recorded".

2. What happens on the exceptions?

Software demos always show the happy path. Your week is not the happy path.

What happens when the site is inaccessible? When the customer is not home? When the job takes two days instead of one? When the material comes back contaminated and cannot be taken? When a job gets split across two crews?

Ask the vendor to show you those. If the answer is "you would just add a note", that note is going to become your problem, because notes do not flow into invoices.

3. Can your least technical person use it in the rain?

A field app is used one-handed, on a cracked screen, in bad light, sometimes with no signal, by someone who did not choose it and does not care about software.

Two things matter enormously and are rarely demoed: does it work offline and sync later, and how many taps is the most common action? If logging a completed job takes eleven taps, it will get done at the end of the day from memory, which means it will be wrong.

4. What does it cost when you grow?

Per-seat pricing means every new hire increases your software bill forever. That is fine at three people. At fifteen it becomes a genuine line item, and the incentive is quietly to share logins, which wrecks your data.

Do the sum at the size you want to be in three years, not the size you are today.

5. Who owns the data if you leave?

Ask specifically: can you export every job, customer, photo and invoice, in a usable format, without paying for the privilege? Get the answer before you put five years of history in.

This is the one people find out about at the worst possible moment.

When off-the-shelf is the right answer

We build custom software, so take this with the appropriate salt, but: if a product does 90% of what you need and you can live with its 10%, buy it. It is cheaper and it is available on Monday. ServiceM8, Tradify, simPRO and the rest are good products used happily by thousands of businesses.

Off-the-shelf stops making sense at a specific point, and it is quite easy to recognise. It is when the workarounds become the system. When there is a spreadsheet beside the software that everyone actually relies on. When you employ someone partly to move data between two tools. When the thing that makes your business good is the exact thing the software cannot do.

At that point you are already paying for custom software. You are just paying for it in wages and evenings instead of in a build.

The honest comparison

Compare the true annual cost, not the subscription. Subscription plus per-seat growth, plus the hours lost to the workarounds, plus the money leaked through under-billing, plus the spreadsheet nobody admits to.

Then compare that to a one-off build you own outright with no per-seat fee. Sometimes off-the-shelf still wins comfortably. Sometimes the sum is not close. Either way you will have made the decision on numbers instead of on a feature list.