Thing is in theory, SaaS / Cloud should be win-win due to specialization. However the pricing I have seen is anything but - they need to charge that to support the sky high valuations and funding rounds. Which is why most scramble to do vendor lock in because without a barrier to entry they will need to keep reducing their prices down the equilibrium level that others can charge and have normal (but not abnormally high) economic profits.
To give you an example, I was quoted 12,000$ per year for an enterprise b2b database solution to record questionnaire answers. Great software, meeting a need we have. But I’m building a basic postgresql database and a very basic web front end for free in its place. 80% of the functionality, 0% of the cost.
I admit the remaining last 20% is the most complex and hardest part; also marketing and all the business aspect is a whole another game. So I have no intentions trying to compete with them, I respect them but at the same time I can’t justify to my company to spend 12,000$ per year for something that an in-house solution can cover most of. Even my CEO said - isn’t that just a database?
This case doesn’t apply generally, but there should be enough examples of SaaS that can be replicated and we don’t need to pay an inordinate fee to use. I would place substack firmly in this list - it is a mailing list software with a text editor (I wouldn’t be surprised if they just reused TinyMCE for this).
Most of the examples you have are very advanced software and I agree with you that having them readily available for all companies is a net positive. Perhaps sales force could be replicated, the rest require serious programming chops and investment.
I basically agree with you but want to share a few more data points on uncommon cases to show the angle I’m coming from. My belief is substack is in this set of uncommon cases.
Your ROI calculation is wrong. Your homegrown solution costs you [your hourly wage] x [hours spent working on it] + [infrastructure costs]. If you have an engineer who spends 3 or more weeks a year on this (incl. initial development cost amortized over # of years), then you lost money.
Worst, you spent time building a commoditized solution which you could have spent improving your own product!
I see these "I could build this in a weekend" type responses regularly on HN, and in my opinion it could not be more wrong. Substack isn't a CRUD app with a few features, it is man-years of work on the technical side. More importantly, they've had amazing execution on the business and product side. Building a business is really hard, and because you could replicate substack-the-app does not mean you would can build substack-the-business.
> It works to some degree, but I’m looking forward to the day that there are enough programming specialists that many solutions are done in house
What a waste this would be! Starting a business today is tractable because we build on the shoulders of giants. Tools like JIRA, Github, Sentry, Salesforce, Hubspot, Zoom, Slack, Notion, Stripe, GSuite, ... and many more are so much better and cheaper than anything you'd build in house.