The ROI is the free software

Every good work of software starts by scratching a developer's personal itch.

You don't have to pay people to scratch their own itch. However, you do have to pay people to write/work on inferior software.

It depends on the software, just by being open source doesn't mean you give up all rights to earn money.

One example is the Aseprite editor https://github.com/aseprite/aseprite/ it is open source and you can build it from source and use it fine, but most average people can't be bothered to build from source, and the prebuilt binary downloads are paid, many people pay for this software.

Another example is apps that may run on the cloud, you may open source it so people can self-host it, but you'd still have a centrally hosted version that's paid.

It all depends on what type of software you are trying to sell.

Yes, most of the time.

You can sell your open-source software and/or services related it. You just need to provide the source and allow your customers the usual OSS freedoms.

You don't need to provide everything for free (as in free beer)

There are several ways to monetize open source

Open source is the best way to deliver software.

Closed source libraries are a pain to distribute. If you only provide binaries and not the source code you need to generate every single combination of platform and compiler, which specific versions, including all the external libraries it uses. It's simply crazy and a huge pain. One perfect example of this is CUDA from NVidia, it's a nightmare to use it because everything has to match exactly, including your OS version as the libraries needed usually come with the distribution.

With Open Source you can compile it anywhere. You also get to see what is does, and you can tweak it to fix bugs and improve it. None of that is available in closed sources.

You also can grow a community with open source, where other people fix the code and gives it back to you. Most software we have today is based on open source.

In terms of money, then it depends on your specific project. Open source has no limitations on making money. You just have to generate value with it so that people want to pay. A simple example would be selling the compiled version of your open source code, but there are multiple ways of doing it. You could have a "Community Version" open source and have extras paid, etc.

Once you have a monumental idea, a lot of people in the community will notice the potential. If you are just someone whom is so insecure about coding skills, software deployment skills, software development life cycle skills you will just end up like the supposed coder of a jumping fish game and then disappears without an update.

Making your software open source probably quite heavily depends on how much of an ideologue you are. That is, if you don’t have a solid business model behind doing so. Many big corpo projects like Facebook’s React are open source as a way to get free work from other developers, and developers are happy to do so because they like using React.

As others have said, offering your software in its entirety as self hosted if its a cloud option, and then just be the best and most authoritative source for hosting it if people don’t want to go through the self-hosted hassle.

Other reasons may include being a personal project that you want help with, similar to the previously mentioned React model.

Some may do so as a “giving back” model. Many coders know their success wouldnt have happened without open source. They enjoy coding, dont need the money and viola.

Another incentive is perhaps you have an employer or client who is paying you for work. If you have a strong negotiating position, you can make your contract stipulate that some or all code be made open source.

its not "worth it"

get a real job, contribute to your passion projects on the side.

don't expect much from open source

Adding to what k00b said, the open-source model is suitable for economically sustainable business if you use it essentially to make your product a standard and to make it easier to be tailored by the consumer. Then it can be so convenient that entire companies can develop around it. My golden example is Prusa, the 3D printer manufacturer, who made every single aspect of the product 100% open-source: both hardware and software. In doing so their designs became so ubiquitous that it became the de-facto standard, with all of the convenience that that brings with it in the technical world: abundance of both high quality documentation, experienced operators and higly knowledgeable communities. Calibration profiles are mercilessly battle tested so you can profit from infinite previous experience. And in any case, you're never locked to the vendor on anything: neither what you buy nor what you learn are vendor-specific. That did not but to allow a first niche to grow from tech-savvy people that in turn encouraged non-technical people to buy the machine directly from Prusa out of trust. The result has been a model to follow.

Now, it's not really an unexpected result. For example, the industry of flow-meter manufacturing is 100% standardized. It's not a law by any means, yet people do it anyways so to be fully compatible within the specific technical field. The standard is "open-source" in the sense that's freely available for everyone to be able to manufacture devices according to standard. Yet in despite of no intellectual property being possible there, the industry thrives. That's because, as k00b pointed out, profit is not in the standard itself (or "the code" if you will) but in the added value around it: manufacturing, calibration, assistance, etc.

If you think about your code as an ISO standard you can download from the web to use for a product, then you will always see if you're heading towards profit or not. If your project ends with the development of the standard (i.e. the code in itself), then you don't have a business model and will depend on charity.

It depends on the software. If the code is all that's needed to easily receive all the code's value, which is rarely the case, then it might not be worth it.

I am telling that you have to be a Linux Kernel programmer to finally feel the rewards that you are longing for. Open source is a hit and miss thing but if you finally hit it off you will be surprised. More than you will ever know.