The problem is in measurability of a change. If I improve a developer or analyst workflow that makes 100 people 1% more successful, that's likely both more beneficial and also harder to measure than a change that makes 5 people 10% more successful.
That's the bit where the GGP's "optimising for the right factors", or something adjacent to it comes in. Many things can't be objectively measured, and upon realising that, many techies give up. But there is subjective truth, which for sure is a lesser truth than the objective sort, but truth none the less. How did you come to believe that you've made 100 people 1% more successful? If you can convince yourself that you've done good, important work, you can probably convince other people too -- but you have to do it, they're not going to convince themselves. Collect the evidence, such as it is, and articulate your case.
