Only 40 percent of Canadian software buyers say they're satisfied with what they bought, and 89 percent of the ones who hit real trouble during rollout ended up regretting the purchase entirely (Capterra, 2025). That's a rough failure rate for a decision most organizations make once every few years, if that. The same research also found what separates the satisfied 40 percent from everyone else, and it isn't the product they picked. It's how they picked it.
We sit inside Architecture & Vendor Selection, Stage 6 of our own methodology, often enough to see the same four mistakes on repeat. None of them are exotic. All of them are avoidable, and none require a bigger budget to fix.
Mistake one: a shortlist that never actually got short
Ninety-nine percent of satisfied Canadian buyers narrowed their choice to five vendors or fewer before making a decision. Among disappointed buyers, that number drops to 90 percent (Capterra, 2025). The gap looks small until you notice what it represents: the buyers who kept comparing six, eight, ten options were spreading the same amount of due-diligence time across more vendors, which means less time on each one. A shortlist of nine isn't rigor. It's indecision wearing a spreadsheet.
The fix is blunt: do the wide search first, then force yourself to cut to five before you book a single demo. If you can't cut the list, that usually means your requirements weren't specific enough to rule anyone out yet, so go back and tighten those instead of watching another sales call.
Mistake two: buying the demo instead of the vendor
When Canadian buyers who regretted a purchase were asked what went wrong on the vendor side, the top answers were unmanaged expectations (41 percent), a rocky handoff from sales to implementation (38 percent), and the vendor simply not living up to what they'd promised (36 percent) (Capterra, 2025). None of those show up in a demo. A demo is a rehearsed thirty minutes with a dataset built to make the product look good. What it can't show you is who answers the phone in month four, or whether the person who sold you the deal is still involved once the contract is signed.
BDC's advice here is plain and worth taking literally: talk to peers in your sector who've already bought the thing, sit through more than one demo, and actually test the tool with your own data before you sign (BDC, 2022). For a nine-person accounting firm in Saint John choosing a practice management platform, that might mean one phone call to a firm in Fredericton that switched eighteen months ago, and a straight question: would you make the same choice again, and what do you wish you'd asked?
Mistake three: treating the signed contract as the finish line
About half of satisfied buyers had planned their rollout stages before the ink was dry. Among disappointed buyers, only 27 percent had (Capterra, 2025). Vendor selection doesn't end at signature. It ends when the tool is actually being used the way it was supposed to be used, by the people who have to use it. If the rollout plan gets written after the contract instead of before it, you've made a purchasing decision without knowing whether the thing is workable for your team, and you find that out the expensive way.
A rollout plan doesn't need to be elaborate. It needs three things before you sign: who trains whom, in what order the team migrates over, and what the first 30 days look like if adoption is slower than the vendor promised it would be.
Mistake four: skipping the three-year math
BDC puts the average Canadian SMB's annual digital technology spend at roughly $118,000, with cost (42 percent) and integration with existing systems (27 percent) among the top barriers organizations report (BDC, 2022). Neither of those shows up in a year-one quote. Licensing compounds as seats grow. Integration work rarely stops at go-live, since every connected system you add later inherits whatever limitations the first vendor built in. Before signing anything, ask what the same contract costs in year three at your expected headcount, and what it actually takes (in hours, not just dollars) to leave if the vendor gets acquired or the product stops improving.
This is where a lot of well-intentioned buyers get quietly boxed in. A platform that looked cheap at the demo can be the most expensive thing in the budget by year three, and by then the exit cost is baked into the decision to stay, not the decision to switch.
Why this hits smaller, less mature organizations hardest
CFIB's research on digital adoption found that firms further along in digital maturity see returns on new technology much faster than firms just starting out: over 55 percent of the most digitally mature small businesses reported ROI within the first year, compared to 27 percent of the least mature (CFIB, 2025). That gap is uncomfortable, because it means the organizations with the least internal expertise to catch a bad vendor pick are also the ones with the least room to absorb one. A 200-person company with an IT department can survive a mediocre platform choice. A 12-person manufacturer in Truro choosing its first real ERP system generally can't afford to get it wrong and quietly eat the cost. That's exactly the situation where the five mistakes above matter most, not least.
What a disciplined process actually looks like
- Requirements before names. Write down what the tool has to do, in order of priority, before you look at a single vendor website. Otherwise the shortlist gets built around whoever you'd already heard of.
- Cut to five, then stop adding. A sixth option rarely changes the outcome. It just delays it.
- One reference you didn't get from the vendor. A name the salesperson gave you is a curated reference. A name you found yourself, through your own network, is the one worth the phone call.
- Price the exit before you price the entry. Ask what leaving costs in year three, in hours and dollars both, while you still have leverage to negotiate it.
- Write the rollout plan before the signature, not after. If nobody can describe week one without the contract already signed, the decision was made too early.
If you're still deciding whether to build something custom instead of buying at all, that's a step earlier in the process, and we've written a full build vs. buy framework for that decision specifically. And if the vendor in question is a consultant rather than a software platform, most of the same discipline applies, which is why we wrote a companion piece on choosing a consultant using nearly the same questions.
If you're in the middle of a vendor decision right now and want a second set of eyes before you sign, a free 30-minute call is enough time to walk through the shortlist and flag anything we'd push back on.