An alternate that matches the datasheet can still fail in your machine. The approval process exists to find out which differences matter.
Component alternates arrive for two reasons: the original is unavailable, or the original has become expensive. Either way, the question is the same — does the substitute do the same job in this application, not in general.
A useful proposal compares the parameters that matter for the design, not every line of both datasheets. Electrical ratings and mechanical footprint are the easy part. What catches programs out are the second-order differences: a slightly different inrush characteristic, a connector that mates the same but retains differently, an LED with the same wavelength and a different viewing angle, a relay with the same contacts and a different coil suppression need.
We propose alternates with the comparison attached and let the customer decide, because the customer knows the application. Directed-buy components stay directed unless you release them. And a proposed alternate never enters the bill of materials until the release exists in writing — a verbal "that sounds fine" is not a release.
The reward for doing this properly shows up in year three, when the original goes end-of-life. A program with a set of already-qualified alternates absorbs the news in a planning meeting. A program without one absorbs it in a crisis, usually alongside a last-time-buy decision made under time pressure.
— GANI Manufacturing engineering team