It made me more wary of SaaS businesses with teams and venture capital behind.

In my experience companies with VC money are much more likely to "flip" a business and sell to someone else causing disruption, or bad updates who make the product worse.

I still haven't had any of the subscriptions for products made by indie developers get significantly worse or change significantly.

They keep the lights on, add some features without breaking existing stuff.

If they're in for the lifestyle more than the money, I think they're less likely to sell and more likely to hold and maintain (assuming they're charging enough to make a living).

Yep, sustainable dependencies are a key criteria for us. I'm happy to see OSS with a big community or paid hosting. OTOH, grown wary to see VC-backed (esp. big rounds) or corporate side-projects (esp. at unicorns with high churn in teams + tech, or bigco's with no OSS community governance track records).

Yes, it's doable to replace things after a year or two. But, if a bunch of dependencies have risk, that turns into a rip-and-replace treadmill of negative customer value and burns out a team.

This is definitely one of the downsides- but also possible that Google for example gets bored of one of its products and shuts it down, or acquires a product just to shut it down.

Personally I think one of the upsides is that the semi-solo team might actually decide that the successful product is "done" and move on or go into maintenance mode. IMO products like Spotify and Slack could have reached that point a while ago- sure they could still work on uptime, data compression, etc. But there is only so much you can tack onto a music player or a chatroom. They keep shipping more redesigns and new features that don't make users any better off but need to keep people busy.

I mean at sometime, you may as well just shift the engineering team onto another/newer product. Spotify could have totally made something like clubhouse.....

> How can I invest time and money in using a product if I'm afraid the founder will bail after a few months?

Would you rather a risk a much bigger one-off payment? For software? Sure, it works now, but it might never get updated, and eventually stop working altogether due to OS updates or other factors.

The one-time payment model is a much greater risk to the consumer than monthly subscriptions.

I don't agree with this at all. If a software product doesn't get upgraded you have 1) time to look for an alternative according to your own priorities 2) ways of mitigating potential risks (if it's a security problem, like airgapping the solution) 3) the ability to make value-based calls on the risks of not having updates

With SaaS you have all of these same constraints (as there's no guarantee they update the service either), plus the risk that they just turn it off one day on a timeline you can't control, plus potential risks that they will leak your data if they don't maintain security (or want additional money), and with the added risk/cost that they may decide to no longer support something you need due to a forced "upgrade" (like API compatibility).

These are also valid points.

Anecdote: A company I know used to have an on-premise product. The customer paid for a yearly license*. Then they developed a cloud based offering, a SaaS. In order to force customers over to this solution they simply stopped updating and giving support to their old product.

* Sometimes we forget that the old software model of offline or customer-installed (client, or server, on premise) also has fees attached to it, in the form of a license you have to keep maintaining. Often yearly.

If the customer wanted to keep using the product, they could not stop paying the license fee just because the vendor stopped updating and supporting it.

> Sometimes we forget that the old software model of offline or customer-installed (client, or server, on premise) also has fees attached to it, in the form of a license you have to keep maintaining. Often yearly.

A lot of the enterprise on-premises software that I use (from Atlassian, Microsoft, Oracle, etc.) come with perpetual licences. The annual maintenance cost is for support and for updates. I can think of a few exceptions that only have term licences, like IBM PurifyPlus, but based on my personal experience, a majority of on-premises software falls under the former category.

Yes but basically all of that software is so mission critical that you simply cannot not pay for support and upgrades. Ergo you have an annual license fee to be able to actually operate it.

> For software? Sure, it works now, but it might never get updated

That's often good rather than bad, as many products are optimized away from usability over time.

> and eventually stop working altogether due to OS updates or other factors.

There are means to mitigate that (e.g. running software in VMs), that work fine as long as software isn't fighting it.

Arguably the biggest factor causing software rot, making one-time payment risky, is purposefully adding dependencies on vendor's servers (e.g. for "collaboration") where they don't belong[0]. This is also what is used to justify the continuous stream of updates - whereas in reality, it's just there to create a plausible reason for why a perfectly good product now requires a subscription. Software that's designed to be paid one-off from the start doesn't require these kinds of anti-features, and doesn't rot nearly as fast.

--

[0] - It's famously said that Dropbox was doomed from the start, because it's a feature, not a product. Unfortunately, that view is part of what's making SaaS so incredibly user-hostile. Syncing data is a feature completely orthogonal to applications, and even operating systems. There's no need for a SaaS vendor to make their own synced storage for app data, unless they're purposefully trying to entrap their users.

> That's often good rather than bad, as many products are optimized away from usability over time.

I feel Adobe products are a great example of this.

> The one-time payment model is a much greater risk to the consumer than monthly subscriptions.

Short term - yes. But long term? It seems there is no point in planning to use a software for years now. And I don't like to migrate my knowledge base to new tool, every 2 years. But is there a way other than mastering org-mode on weekends?

> But is there a way other than mastering org-mode on weekends?

1. a quick tutorial will be plenty to get you started: https://mickael.kerjean.me/2017/03/20/emacs-tutorial-series-...

2. if you don't want to bother learning emacs, no worries, use one of the web apps such as the one I built: https://demo.filestash.app/login?next=/view/org/emacs.org#ty...

See also ObsidianMD and Athens Research.

Long-term? Not much beyond org-mode + org-roam at the moment. I've seen one or two alternatives (open-source products), but I don't remember their names.

The rule still applies: whatever you chose, do not become dependent on a SaaS offering. The problem is in the name: service. A service is something other people do for you. They can stop at any time.

We are dependent on food and water, which requires ongoing support from a vendor.

The problem is in having a trustworthy vendor or a healthy marketplace of alternatives you can switch to.

That's true. We're all dependent on plenty of services. But the difference between those and SaaS is that they're all commodity services. This means you have a steady selection of service providers, all providing equivalent service, and there's no risk with any individual provider disappearing - you can just switch over to a different one, and get the same thing.

An important aspect of most SaaS businesses is that they're not commodities. It's a core part of most SaaS business models - they're purposefully sticky. When you become dependent on one, and it shuts down, or gets acquired, or decides to discontinue the features you care about, there usually is no drop-in replacement. There may not be a replacement at all, for the combination of features you need. And whatever alternative you find, migrating all your data is usually difficult, and often not possible.

> Would you rather a risk a much bigger one-off payment?

That's a false choice. The problem with the proliferation of "services" is that many are hard to cancel, or are betting that I'll forget to cancel.