Give a man an opinonated tool that support their job, and they will do their job well, give them a tool to configure and their job becomes configuring the tool.
If you put Jira and Linear through a proper RFP, Jira wins almost every time. It has more features, more integrations, more permission levels, more configuration options, more certifications and a bigger vendor behind it. Every row in the scoring matrix is a row Jira was built to tick. Ask most people who have actually used both, though, and they will tell you Linear is the better tool for running a business, and that is the part no RFP knows how to score.
What makes Linear better is mostly things it does not do. It is opinionated about how work should flow, it is fast, it has cycles that force a cadence, and it does not let every team configure its own workflow with fourteen statuses and a mandatory field nobody understands. People actually use it, and the data in it is therefore closer to the truth. None of that shows up as a feature. In a weighted scoring model, "less configuration" scores lower than "highly configurable", even though the configurability is exactly what turns Jira into a bureaucracy after two years.
This is what happens when procurement tries to turn a decision into an academic exercise. We set up requirements lists with hundreds of lines, give them weights, score vendors from one to five and average the results across a committee, and the number that comes out looks objective. It is not. It measures breadth, because breadth is what can be counted, and it ignores fit, speed, adoption and how the tool shapes the way people work, because those are hard to put in a cell. The process ends up rewarding the vendor that is best at answering RFPs, not the one that is best for the business.
The RFP also protects the people running it more than it helps the business. A decision backed by a 200-line scoring sheet is easy to defend if it goes wrong. A decision that says "we picked the one our teams actually work better in" feels exposed, even when it is the right call. So the process drifts towards the defensible choice instead of the good one.
If you want a better outcome, keep the hard requirements as a short pass or fail list: security, data residency, the integrations you cannot live without. Then stop scoring and start testing. Let two real teams run their actual work in each tool for a few weeks, call companies like yours that use them, and let the people who will live with the tool make the call. Some of the most important qualities of a tool cannot be quantified, and pretending they can is how you end up with Jira.