I’m sorry, but most tech companies are clueless when it comes to management. Cost analysis as it relates to the actual activities of an engineering labor force are brushed to the side as overhead, which makes no sense since these are direct costs.
What’s the cost of a poorly specified product? What’s the cost of a meeting? What’s the cost of involving engineers in the hiring process? What’s the cost of multiple programming languages? What’s the cost of having separate teams for front-end, back-end, operations? What’s the cost of adding another engineer to a team?
I’ve never seen any effort put into any of the above. This is underpants gnomes territory:
1.) acquire new engineers
2.) ?
3.) profit
Hence massive overhiring with unaccounted for costs that lead shortly to layoffs and hiring freezes.
The more I learn about managerial accounting (thanks dad!) the more dysfunction I see at most organizations.
Apparently Oracle does attempt to uncovers these costs, but I can’t find the case study.
All of this has an incredibly large impact on financial reporting, hence rosy projections based on EBITDA followed by the sobering realizations when those projections are not met. Most tech firms actually base internal management on EBITDA! This is borderline insane internally and borderline fraud externally.
If we don't undo the winner-takes-all nature of tech (by forcing open protocols and mandatory interoperability) then tech will continue being a dystopian situation of take-all-never-innovate while relying on the difficulties of the network effect to prevent millions of brilliant ideas to never see the light of day.
TikTok, an Instagram competitor, had over 1,000 people working on it in 2020 (grown 130x since).
Those features you’ve listed were added in the 11 odd years since. I think that the original set of employees could have implemented every aspect of the application since the date of sale!
Could 13 people clone Instagram's featureset at the time of the sale? Yeah, it could probably be done by 3 people in a weekend. But you'd need thousands to emulate their current featureset.
But then Instagram was mostly just a simple photo editor app without the social graph complexity. Sure you can still build and maintain such an app with thirteen people.
I really like the new Canvas feature they have. I realize it's more updated version of posts that they already had but seems to work much better than posts IMHO.