I see this in tech all the time. It is indeed very hard to measure the return on investment for tooling and infrastructure in tech! Any infra work, whether it's

I see this in tech all the time. It is indeed very hard to measure the return on investment for tooling and infrastructure in tech! Any infra work, whether it's splitting up your monolith into components, improved developer tooling, or fixing flaky builds has a vague and hand-wavy return on investment and has to compete with "Well customer FooBar will sign a $10M contract in Q2 if we build feature BashBaz instead," and now good luck as an engineer explaining how and when exactly your investment in developer tooling is going to make the company $10M in Q2. I agree that it's hard to quantify those kinds of things, but I might quibble over the degree of difficulty. At the very least, some techniques for dealing with these kinds of "hard to quantify" scenarios are known, published, and have been analyzed to varying degrees. For a company the size of SouthWest, with the amount of money that's at stake, I find it surprising that they wouldn't use such methods (or maybe they do and just still managed to screw the pooch, I don't know). What means am I talking about? Well, you could start with by the book Scenario Planning[1], and then have domain experts assign costs to different scenarios and use calibrated probability estimation[2] to assign probabilities to those scenarios. And then you could use Monte Carlo simulation[3] to get an idea of the likely distribution of outcomes. From there you can calculate a range of likely costs (or returns, depending on which way you frame the problem) for dealing with, or not dealing with, specific scenarios. This is a very high level and hand-wavy description, btw, of the methodology Douglas Hubbard[4] lays out in his book How to Measure Anything[5]. [1]: https://en.wikipedia.org/wiki/Scenario_planning https://en.wikipedia.org/wiki/Scenario_planning [2]: https://en.wikipedia.org/wiki/Calibrated_probability_assessment https://en.wikipedia.org/wiki/Calibrated_probability_assessm... [3]: https://en.wikipedia.org/wiki/Monte_Carlo_method https://en.wikipedia.org/wiki/Monte_Carlo_method [4]: https://en.wikipedia.org/wiki/Douglas_W._Hubbard https://en.wikipedia.org/wiki/Douglas_W._Hubbard [5]: https://www.amazon.com/How-Measure-Anything-Intangibles-Business/dp/1118539273 https://www.amazon.com/How-Measure-Anything-Intangibles-Busi...