Choosing & integrating business software

Nobody sets out to build a mess. It accumulates. A booking tool here because the old way was painful, a separate invoicing app because the accountant asked, a spreadsheet nobody will admit is load-bearing, a CRM someone signed up for and quietly abandoned. Each decision was reasonable on the day. The result is a business where the same customer exists in five places, none of which agree, and where somebody's afternoon is spent retyping between them.

The problem was never the individual tools. It's that nobody was asked to choose how they'd talk to each other.

Ask what it costs to leave, not what it costs to join

Pricing pages are designed to make joining easy. The number that matters is the cost of getting out. Can you export your data — all of it, including history and attachments — in a format that another system could actually read? Is there an API, and is it documented, or is it a page that says "contact sales"? If the honest answer is that leaving would mean re-keying three years of records by hand, you aren't buying software. You're signing a lease you can't break.

Decide which system tells the truth

This is the decision almost everyone skips, and it causes most of the pain that follows. For each important kind of information — customers, jobs, invoices, inventory — one system has to be the authority, and the others read from it. When two systems both think they own the customer record, they will drift apart, and then you're not running a business, you're refereeing one. Pick the source of truth first. Everything else gets easier.

Buy for the shape of your business, not the size you hope to be

Enterprise software is built for enterprise problems — approval chains, role hierarchies, compliance regimes — and it carries the cost and the complexity to match. A ten-person company running a platform designed for a thousand doesn't look ambitious. It looks stuck: expensive, slow to change, and requiring a specialist to do anything at all. Buy for the business you have, and buy things that can be replaced when you outgrow them.

The best software is the software people use

A tool your team quietly works around is worse than no tool, because now the real process is invisible. Before you commit, have the people who'll actually use it do their real job in it — not a demo, the real thing, with a real awkward customer. If they hate it in a trial, they will hate it in production, and they will go back to the spreadsheet the moment you stop watching.

Know where your customers' data lives

Canadian businesses have obligations to the people whose information they hold, and "the vendor handles it" is not a defence. Know what data each tool stores, which country it sits in, who at the vendor can see it, and what happens to it if you leave. This isn't a legal formality — it's the kind of thing that only ever becomes urgent at the worst possible moment.

A buying process that holds up

1. 01Write down the job to be done and the process that exists today — including the spreadsheet everyone pretends isn't there.
2. 02Decide which system will be the source of truth for each kind of record.
3. 03Shortlist on data portability and a real, documented API before you look at features.
4. 04Trial with the people who'll use it, doing real work, not a demo script.
5. 05Confirm exactly which fields sync where — and what happens when the sync breaks.
6. 06Check where the data lives and what your exit looks like.
7. 07Only then compare price.

Connected beats clever

A modest set of tools that pass information to each other reliably will beat a best-in-class collection that doesn't, every single time. The goal isn't a beautiful stack. It's that a customer's details are entered once, and everything that needs to know about them finds out on its own.