I think there's lots of interesting example companies/products like this outside the startup/software space.

For those into photography, Sigma makes great 3rd party lenses at reasonable prices for other brands, which drives their revenue. But they are privately held and able to take more experimental risk making oddball cameras at low volumes. A lot of people laud their product design and give fairly positive reviews without actually buying them.

I've owned a few myself, but they are always a "camera for someone who already has 2 other cameras" type of product. Sometimes "no one has designed a product like this before" is for good reason, and predictive of poor sales.

So maybe to bring this back to software - consider if your product simplifies a customers life / replaces anything, or simply adds more complexity & risk to their stack.

I think this also has a lot to do with HOW your product is bought.

May products are discovered. They're cool, they're novel, let's try. Those commonly end up in this situation where users like the thing but never buy it.

But the closer your product category is to being infrastructure, the less this happens.

In the space I'm in (billing/metering) but also many others, you don't stumble upon a product, think "that's cool" and hot-swap an important part of your architecture. You convene a buying committee and compare vedors.

Of course this advantage (everyone you speak with has higher buying intent) is counteracted by the fact that these types of products aren't "sexy" in the sense that they'll go viral with a snazzy animation.

Another, reason products end in "purgatory" is that the customer is already using something like it. During market research that's the final question people forget to ask. You might ask "Do you have a need for this product?" - "yes". "Do you like the product?" - "yes!", "Would you pay money for it?" - also, "yes". If you stop here and go hire 10 developers and spend millions of dollars to build it you might be screwed, because you forgot to ask "But do you already use something like that?" and the answer might also be "yes".