Most software gets chosen for this quarter's problem. You need a thing that does X, you find the tool that does X best in a demo, you buy it. It's a perfectly reasonable way to make a decision and a terrible way to live with one, because the tool you choose for a quarterly problem becomes a system you depend on for years. Choosing the five-year tool — the one you won't resent in 2031 — means weighting completely different things than the demo shows you.
The demo optimizes for the wrong horizon.
A demo is designed to win the quarterly decision. It shows the feature depth, the slick onboarding, the impressive edge cases. None of that predicts how you'll feel in year three, when the question isn't “does it do X” but “do I own my data, can I leave, has it gotten better or worse, and do I trust the company behind it.” The qualities that determine a five-year relationship are almost invisible in the thing designed to close a five-week one.
The demo answers “is this good now?” The only question that matters is “will I resent this in three years?” — and the demo is structurally unable to ask it.
What actually predicts the long run.
Three things, mostly, and none of them are features. Data ownership: can you get everything out, with structure intact, for free, today? A tool you can leave is one you can stay with on your own terms. The trajectory: is the company making the product better for customers, or better at extracting from them? Look at the last two years of changes — they predict the next five. Trust: does the vendor charge for the exit, use dark patterns, hide the roadmap? Small dishonesties now become large ones over five years.
The consolidation angle.
There's also a structural case for choosing fewer, more fundamental tools for the long run. A point tool solving a narrow problem is, over five years, more likely to be commoditized, acquired, shut down, or left behind than a platform that owns a core part of how you operate. The five-year tool tends to be the one that's load-bearing enough that the company building it has every reason to keep investing — and central enough that it compounds in value as your data accumulates in it.
Choosing software for the quarter is how you end up with a stack you resent — each tool a reasonable decision at the time, collectively a trap. Choose the five-year tool instead: own your data, confirm you can leave, watch the trajectory, trust the company. The features will change. Those things are what you'll actually be living with.