I especially agree that businesses should focus less on feature lists and more on the actual results a product delivers. A cheaper or newer tool is not automatically better if it creates more manual work, poor integrations, or unreliable workflows.
The idea of testing alternatives against real business processes is especially important. Demos can make almost any product look impressive, but a small pilot often reveals how well it actually fits into the existing software stack. The best alternative is ultimately the one that solves the problem better with less cost, complexity, and unnecessary work.