I maintain an open source push notification service called ntfy [1]. It's 100% open source, but I have recently started offering ntfy Pro, a paid SaaS offering.

I am telling you it's super hard to get people to pay for it. It may be because the free tier is too generous, or because I don't aggressively push or market it, but most likely it's because people can just selfhost it. I have 25k active daily users for the free tier, and probably hundreds or thousands of selfhosted installations, yet as of today, I have only 20 paying subscribers. It's quite brutal.

The worst part is that ever since I started offering paid subscriptions, donations have reduced and new donations have pretty much stopped, because people think I'm making bank with the paid stuff.

I know this doesn't answer your question, but it gives you the perspective you may be actually seeking. It's definitely hard to pull off.

[1] https://ntfy.sh/

I notice your lowest plan is 2500 daily notifications. So you’re stating you have 25k users that use less than 2500 daily notifications?

I’d suggest your free tier is too generous if you’re struggling to convert free to paid.

I feel like the value of open source saas is that you’re not locked into the vendors roadmap or pricing in the same way as a closed source saas.

If you need feature xx in the future, and they don’t provide it. You add that feature, and move to self hosting (or maybe they’d even accept the PR and it could become part of the hosted service, tho less likely). And if your business / service thrives, and you scale to the highest tiers of the pricing, you have the option to move to a self hosted version. Obviously self hosting has additional dev ops costs, but once you’re at a certain scale, and you have those dev ops resources anyway, it could make sense to take more control over the services you rely on.

Those options give me the confidence to deeply integrate and plan for the long term. But it’s not about free, it’s about that flexibility in the future.

An overly generous free tier seems more in line with a closed source saas lock-in strategy.

If donations make a sensitive income flow, why switch to paid SaaS offering ever? Keep your OpenSource as a non-profit and develop SaaS as another brand, separately. Offer a premium proprietary product as a separate product under the same brand.

Let's talk if I can help you: https://calendly.com/ignatev/coffee-online

> It may be because the free tier is too generous

Yea I also think 250 messages a day is a lot for the free tier. Thanks for sharing!

Btw I saw ntfy on HN a couple of months ago and I’m already one of the starrers.

It used to be 17k per day, and I've reduced it slowly, but I don't want to screw anyone over. Hehe.

Wow, 17k requests per day?! Is that like an enterprise plan or something?

Just emailed you with my “real” identity, maybe we can continue from there? :)

give a 3 months heads up and email every month then a week before and people should be fine to arrange payment

99% of my users do not sign up. And even the ones with user accounts do not have to provide email addresses. I am quite the privacy junkie, though I realize that's counter productive for a product haha.

That’s awful, though not surprising. Thanks for sharing.

I use anonaddy [0] because it's open source and self-hostable [1]. I don't have to worry about the service going under or jumping the shark, since I can always just self-host it on my own hardware and import my config should that happen. Of course I'd much prefer to pay someone else to run it, especially in the case of mail servers where self-hosting is notoriously tedious.

[0] https://anonaddy.com/ [1] https://github.com/anonaddy/anonaddy

Depends very heavily on the product, but short answer 'yes'. I already do pay for Open Source SaaS products.

For very critical infrastructure or for situations where I care a lot about data security, maybe not. It's still not a totally easy sell, I don't seek out SaaS solutions to my problems necessarily. But there are situations where I would (and again, actively do) pay for convenience and for someone else to handle deployment and security for me.

It's actually a much, much harder sell to get me to pay for a SaaS solution that can't be self-hosted and that isn't Open Source. Just as one example, I pay for Wallabag because I don't want to deal with deploying it. I would not pay for Pocket.

The Open Source part of a hosted software solution is about giving me a sense of security that it's safe to use the product in the first place. The way I see it, the hosting is a completely separate service from that. Once you've convinced me that it's safe to use your product and that if something goes wrong I can just move to self-hosting, then at that point you can make the sales pitch to me, "and also wouldn't it be nice if you didn't have to self-host?"

> would you still pay around $5 to get the cloud version of the app?

I'll also say that for personal usage around $5 a month falls basically squarely into my "this is kind of an impulse purchase" bucket. If you're charging $20-40 a month for a personal service that I'm using just for myself independently, then like... I don't know, do I really need it? Is it really worth paying for something I can do myself?

But $5 a month is very squarely in the range where that thinking completely flips and I start saying, "I don't know, do I really need to spend a day setting up a VPS? Over $5? It's not worth it, I want to just shoot the dev some money and have one less thing to worry about on my monthly todo list."

But again, obviously that's somewhat software-specific and depends a little bit on whether or not it is genuinely something that I really want to use. For business use-cases where it's something more important, I'd probably be willing to pay more.

> The Open Source part of a hosted software solution is about giving me a sense of security that it's safe to use the product in the first place.

Do you mean as in auditable? Would you settle for source available for this use case? Sounds like you also value the ability to use it to self-host for free if something goes wrong, right?

My company makes a security and automation product that customers self-host (we do the deployment for them if they need, which typically they do except where you need to be cleared, then we train on a demo system and they do it themselves), but we don't make it open source because we think it's hard to capture the value in that case. As in: you can self-host valuable IP for free for commercial use under OSSL, and we get volume-discounted 5/seat/mo with SaaS if you really want to? We're scared people will just not pay.

I guess the jury is still out for us, but we used to have a OSS product with degraded quality, but we found it inefficient because customers would try to upgrade that rather than bothering with licensing, get stuck, then get fed up. So patient was a limited resource. So we remove the artificial limitations, but also make it non-open-source, but free for noncommercial use (governments, non-profits, etc, because they are either too big to close, or didn't have budget, but their adoption is good signal). So we won't do OSS in future for that product.

We make it source available, so it can be audited, and free to use non-commercially, but not free to self-host commercially (possible to, but you need to license it). We haven't gone the SaaS route because: 1) customers often want to use their own infra anyway, 2) customers have different customization needs, 3) we always thought about SaaS like 1000s of individual consumer accounts and we weren't sure that was the right fit, and 4) embarrassingly it's pretty complex to build the infra for SaaS especially without being sure there are customers for that.

Tho with the recession, we're considering adding SaaS. Our competitors often have SaaS offerings, and the pricing certainly can work at "developer prices" (5 a month or less, depending).

> Do you mean as in auditable?

Auditable is part of it I suppose; but no, that's not what I meant. I meant security in the sense of "do I feel safe that I'm not going to get burned by it?" Specifically, if something goes wrong or I need to self host or it needs to get forked, can the problem be fixed? Security in the sense of "am I tying myself to a service where I have no control?" It's less about fear of getting hacked and more about fear of a service turning into <insert-basically-any-number-of-proprietary-saas-services-from-the-past-few-years-here>.

Source available would be a harder sell for me but depending on the terms not an impossible sell. I do use some source available products, but not a ton of them. My main question when I'm looking at a SaaS service is "what are the situations where I'll suddenly lose access to this?"

Free for non-commercial use only works if it's a product where I absolutely know for certain that I'm not ever going to use it commercially (and I don't just mean selling hosting as a service; if I'm building any product at all that depends on it, I need to know really clearly in the license that I'm allowed to do so). Otherwise, I don't have that sense of security. But there are products I use where that's fine. If it is a purely personal thing where I'm not going to use it as a business, I wouldn't necessarily drop a product over a source available license I guess.

I pay for and I use Aseprite commercially and Aseprite is source-available. But Aseprite is not a SaaS service, and its license explicitly allows me to modify and self-compile it for internal use even commercially. The risks are not very high for me, now that I've bought the thing I'm basically allowed to do whatever with it. If I have to commercially license the product on a continual basis to use it commercially then yeah the auditability is nice, but otherwise I don't really see what the difference is with a proprietary product. If the company gets bought out Google and your product gets canceled or turned into a mess or merged into another Google product, I'm still going to be in the same situation where I won't be able to do anything about it and where I won't be able to self-host.

That's more of what I was trying to get at when I talked about lack of security.

---

I guess the TLDR is that if a SaaS product is actually OSS, my thought is:

- as an individual I'm willing to potentially consider paying for hosting if it's cheap enough, and as company I would pretty much always pay for hosting unless there's a strong reason for me to care about the data storage or customizations.

If it's source available, my thought is:

- as an individual I'm willing to consider paying for hosting depending on my usage, but as a company I'm probably not touching it.

I would only consider using a source-available SaaS product if I'm very confident that my usage is never going to be subject to the restrictions in the license. Note the phrase "using" there; there are some SaaS products that I avoid even though it would be free to use them. For example, I'm never going to choose to build a product I control on top of MongoDB if I can help it; even though I wouldn't have to pay for hosting to do so. Getting back to my original comment, there are two parts to selling SaaS:

1. Convincing me to use the product in the first place

2. Convincing me to pay for hosting

The first step is the hard part, the second step is the easier part. If I feel comfortable using it, it's not that hard to convince me to pay for hosting. For the most part, I do not want to self-host most of my software; I want that to be someone else's problem. But I'm never going to get to the "pay for hosting" step if I don't feel comfortable using it.

At its best, SaaS should make me feel like I'm getting rid of a dependency, that there's a part of my infrastructure I don't need to worry about because someone else is handling it. At its worst, SaaS introduces a new dependency because it means that I can lose access at any time and that any infrastructure I build around having access can be rendered useless at any time. At its best, SaaS makes my infrastructure more robust by delegating responsibility, at its worst it makes my infrastructure more fragile because it introduces another point of potential failure if the service goes away or becomes unusable.

And the license is very often the biggest thing that determines which of those scenarios is the case for a given service.

> That's more of what I was trying to get at when I talked about lack of security.

Ah, That makes sense! I don’t think we’d really considered that deeply. Thank you for your comprehensive response.

- If I'm using it for fun, learning, or testing it out: I'll probably self host it, I already have a server and enjoy that.

- If any of my friends use it who don't do self hosting normally (techy or otherwise): they'll probably use the cloud offering.

- If I'm using it for business use: cloud offering unless its mission critical or core infrastructure, then it depends.

- If it's one of my few "don't live without it" personal apps: cloud offering or I'll buy a license/do a donation.

I don't self-host, with a few exceptions. Generally speaking, my value proposition is app feature development, not managing the operational complexity of self-hosted software.

For personal projects, I either try to leverage a free tier (if available) or just not use that software, because very few of my projects make any money, and most of what I do is just for fun/learning, anyway.

For projects that make me money, I would absolutely pay $5 or whatever for the basic tier, and as traffic/revenue goes up, I'd happily pay for more. That hasn't happened yet for me, but I would if I made any money ;)

I do this for a number of services, obviously pricing needs to be reasonable. My reasoning:

- I don't want to deal with upgrading, managing infra, backups, outages, security etc etc

- For products I rely on it's good to pay, as that (usually) means getting support if I need it

- I want the developers to continue development and maintain the product, happy to pay for that

That said I only pay for apps that are great and I need and use. If you have a crappy app that I sort of need at some point, I might try installing from source.

If done right, with right transparency, disclosures, yes SaaS it is.

When building Wide Angle Analytics, we spent a lot of time on getting the legal, compliance, and transparency right. We were mocked by the "move fast, break stuff” crowd.

When we were looking for vendors and SaaS partners, we stumbled upon some very dodgy “businesses”.

No address? Nope. No director name? Nope. Dodgy jurisdiction? Nope. Parroting privacy and data, or poor or outright invalid security statements? Nope.

Do your due diligence.

I made https://github.com/nadermx/backgroundremover which can be installed and used fully with no restrictions, MIT licensed. Yet people still feel the need to pay for API usage or my GUI at https://backgroundremoverai.com

Yes. Self hosting for me can come because of a few different reasons: - I want to ensure my data is safe - I want to ensure uptime or performance - I want to ensure continuity if the SaaS company disappears

Pricing matters of course. I get really ticked off at the overprice of many SaaS.

Me for personal use -- yes, something like 5 bucks a month tops. Me for company use -- definitely yes, your pricing will dictate who will be your client tho, 5 per user per month might be too much or too little depending..

> You find a task manager that does pretty much exactly what you need BUT its also open source and you can self host it, would you still pay around $5 to get the cloud version of the app?

If it's for personal use: Probably. The cheapest droplet on Digital Ocean is $5 anyway (I know there are cheaper alternatives, but that's what I use), and I don't have a general purpose VPS that I use to self host stuff already.

If it's for business: Absolutely. The set up time and maintenance time cost dwarfs whatever pittance of savings you might be able to achieve with self hosting.

This also depends on the SaaS pricing and how complex the self hosting set up is.

I've been paying for Plausible Analytics (OSS & self-hostable) from day 1 of OnlineOrNot.

I don't want to save a few bucks a month at the expense of hours of messing around with servers and config, I want to build my product.

There are only an extremely tiny number of SaaS products I have any interest in at all, because the megacorps already have everything, often for free/youretheproductware.

If there was an app I wanted, and I had the money, I would absolutely pay for it rather than self host. Most likely self hosting isn't a thing I'd even consider, or suggest to anyone, it just seems totally insane to DIY something that may well wind up costing more even if your time value is minimum wage like me.

Yes. When I’m trying to make money from something I want someone to blame when the open source software inevitably craps out temporarily :)

I generally don’t care if something is open source or not. Though it can help my decision to choose one thing over another when weighing the risk of getting locked in. I would also pay someone else to pay developers to maintain it so that I can focus on paying for and maintaining my own software.

The advantages of the paid Saas to me would be:

1. Time. If it can be deployed to accessed much faster, or changed/updated faster than I can myself, it's worth the cost

2. Support. SAAS offerings usually have support that would reduce the need to debug everything myself

3. Convenience: The SAAS may be more easily accessible via the cloud portal, wouldn't require that I store and maintain things myself, and overall reduces the overhead in that regard.

Yes. I just don't have time to deal with self-hosting most things anymore. Minecraft for instance. Sure, I could host my own server, but it's a pain in the ass. I spent as much time dealing with the server as playing. I just wanna fire up the game and play for 30 mins or so at a time and have a space my friends and nephew can join. Realms does the job for $10/month. No question worth it.

Absolutely. A lot of the SaaS services I pay for are so that somebody else wakes up at 3am when it breaks so I don't have to.

I’m not sure if it is still an option today, but I have chosen paid SaaS Sentry over self hosted Sentry in the past.

Are you asking us for our perspective as hobbyists, or as b2b customers?

If the latter, yes, absolutely. I have seen a few major successes in my career thay involved a migration from a self-hosted solution to a managed offering. Trying to self-host is often "penny wise and pound foolish" in the long run.

Yes, I do that, examples for:

- https://joplinapp.org/ - https://bitwarden.com/

I self-host several services on a cheap VPS, but I'm paying a yearly subscription to wallabag.it.

I could very easily self-host it, but I chose to help fund its development though that subscription.

For work we always pick the cloud version. Chances are whoever is running that cloud version knows their shit better than we do and if we self host it it's on AWS anyway so probably not much cheaper if scale is any issue.

Absolutely, if we can get away with not hosting and paying a small fee we will almost always prefer that, otherwise we will build out the service ourself.

I don't self-host and will never. I don't have time for all that.

It's still the same fundamental SaaS purchase decision, build vs. buy.

The short answer is "yes" but it's nothing but easy.

SaaS or paying versions (whether the extra features are open source or not) are possible financing sources.

However SaaS requires specific investments which don't finance the open source software itself and you need to factor them in.

As you say, people can self host so you have a "competitor" for your offering. One way to handle this is to have the same free tier in your SaaS than in your self hosting and you can provide paying versions also on the self hosting. Note that you can do paying version that is open source. At XWiki (xwiki.org xwiki.com) we do this with our Apps Store and paying extensions, but all our extensions are Open Source but not deployed for free. This shows our commitment to Open Source. We show to the community that we need ways to pay, but we also show that we are not trying to close down the projection or refuse collaboration and competition with our community.

You could also not have a free tier or smaller tier on SaaS and concentrate on the easy of getting the full service running and ready to go. Now a significant part of open source users are here for the self hosting and keep control. Others are also in for getting things for free (either because they like free, or their boss likes free, or even because buying is complicated). But these free users are also your marketing. Never forget that when you try to get some revenue. It's easy to turn back on your community. In the end you need the free users to prove the project and if you make all paying the reach is reduced. From my pov, the goal is to create more Open Source, not more revenue. The revenue is there to make more Open Source.

So in the end, the paying users will always be a small percentage of you free users. It can be 1% percent. Could be more or less depending on the type of users or software. On SaaS with individuals same rules as free tier that's not open source can apply and be even less.

Now the key is also to have great software. More happy free users, more paying ones.. Same for donations.

But if you start asking for money, whichever method, it's good to be transparent about it. How much money is made, how.. What do you do for it. For example in one comment, somebody mentioned donations going down when launching a service. This can easily happen as users have not idea what's going behind the scenes, because they don't have the numbers. For cryptpad (cryptpad.org) we decided we should publish the numbers. It's not easy to keep it up to date as we work a lot to build the software, developer offers, find funding from projects and so on. But we try. See an example here: https://blog.cryptpad.org/2023/02/09/CryptPad-Funding-Status...

On the XWiki side, we would want to do it too, but lack time, and most of our funders are not companies, which are more interested on how competitive our offer is towards proprietary solutions, than how we fund it. This is also key in the end. Companies and Users have money. They just use it for what they really need and for that they will very often look for the best offer. As opensourcer you still have to convince that it's the best offer and most users won't really care about the fact that it's open source. We still have to convince them that Open Source is valuable in itself and that users should actually pay more for it..

Hope this helps.. Ludovic, XWiki and CryptPad