[apothegm]

Historically a ridiculous percentage of software projects either failed entirely or suffered massive overruns.

For larger, older companies, big-A Agile consultancies were able to sell training and services by touting Agile as a way to achieve savings. They may have even succeeded to some degree at achieving a higher likelihood of eventual delivery and lower overruns.

Smaller companies and early stage startups wanted and needed more lightweight processes. Those that used little-a agile were more likely to succeed and grow instead of going out of business. So there’s a survivor bias at play.

Then everyone else cargo cults and copies whoever looks like they know what they’re doing.

[nudpiedo]

I guess faster reaction times might have been decisive for many early startups.

Nowadays is another story because everyone does and cargo cults are in place.

[apothegm]

Sure. But going back to waterfall isn't going to improve reaction times for anyone. So still only the agile shops survive.