Some of the expenses mentioned in the article can be easily taken off with those free tools which are just as good, if not better. If you develop an in-house tool for comms and collaboration, it will be an overkill for even a big startup.
If I had any more people I would still be taking a hard look at existing, well-understood, open technologies before signing my soul to a private company in business solely due to shackling other companies to their proprietary software, and dealing with their lock in when they inevitably fail me and my needs at a random date in the future.
Then again, I'm also not one of those people who believe the world changed when slack was released, so maybe I'm a little biased.
If we had been OK just using whatever styles they gave us, it would have been fine. But we have a designer who designs how everything is supposed to look. And even the "style-less" react components have been a nightmare. Just a constant train of bugs to fix, and they still barely function.
It would have been way faster to start from scratch, in my opinion. This has all been over 6 months, so it's not like there was some initial payoff period either.
Secondly, frontend is hard. Having a senior react dev who can accurately weigh the trade offs of using external components like that would have been invaluable here. People underestimate how much different in skill level a junior react/frontend dev is from a senior engineer with experience in multiple frameworks. 6 months is long time in terms of dev hours - writing a table from scratch for your use case that works with design is usually a 1-2 sprint task, and that’s absolutely the hardest one there. But more reasonably, work should have been done to make a reusable style library to make this even easier.
+1 on not letting design dictate. Just gotta sit in the room and weigh pros and cons.