It appears to be time for another lesson about the dangers of relying on one particular item of hardware. News broke yesterday that Pebble, crowdfunded manufacturer of low-cost e-ink smartwatches, has just sold its software assets to fitness-tracker company Fitbit and is set to go out of business.

Pebble has announced that it will try to shift Pebble watches away from relying on cloud-based software, but this may result in loss of existing functionality. (BoingBoing commenters seem to think the only actual functionality that relies on the cloud is the speech-to-text function for text message replies in later watches, which would mean earlier ones that don’t even have that function should be fine.) Meanwhile, people who backed the latest Pebble Kickstarter but haven’t received their watches yet will get refunds—but nobody else will, and warranties on all Pebble devices currently in the wild are null and void. So long and thanks for all the fish.

BoingBoing turns this into a screed against the evils of DRM, but I think it’s actually missing the point. True, DRM is obnoxious and annoying, but it has little to do with the Pebble. I don’t know if there is any DRM on it. Even if there is, developing applications for the watch isn’t something any Pebble owner could do if it was removed—whereas, using ebooks with different devices is something anyone could do if DRM was removed on them. But there are more useful parallels to draw to the ebook industry.

How many times have we seen this sort of service shutdown or acquisition-shutdown happen, even in the ebook-and-ereader industry alone? Sony’s hand-off of all its customers to Kobo comes immediately to mind, and so does Barnes & Noble’s abrupt shutdown of Nook UK services (followed by Sainsbury’s ereader service, to which Nook had dumped all its users, also selling out to Kobo). And then there was what happened with Google and Etherpad, or Safari Books Online and Ibis Reader. Aquisitions where the acquiring party improves the thing it just purchased—such as Pinterest’s purchase of Instapaper—are sadly few and far-between. I bought a Pebble last year and found it useful for its long battery life and low cost, but people like me are sadly out of luck now.

It’s not even something you can blame Fitbit for being jerks about. It turns out that Pebble was in the process of collapsing even before Fitbit stepped in to scoop up the tech assets it wanted on fire sale. Ars Technica notes that those assets sold for $40 million at most–slightly less than the total amount of funding Pebble has raised over its 4-year life. It’s also about 1/20 of the $740 million Citizen reportedly offered for it in 2015, and half of the $70 million Intel reportedly offered earlier this year. Pebble laid off 25% of its workforce in March, and Fitbit’s only hiring 40% of its remaining staff now. It’s like the collapse of PDA-maker Palm in fast-forward.

Perhaps the biggest lesson to take from this is the danger of basing too much of your business on crowdfunding campaigns. That’s how Pebble got the funding to make its watches, after all—getting excited early-adopters to kick in to launch early models, and then getting more of them to kick in for later models. Unfortunately, that sort of funding model tends to make it hard to judge just how much interest there really is in your product.

Do your Kickstarter backers represent 10% of your potential audience, in which case you can use the profits as seed money go into manufacturing to serve the rest? Or do they represent 90% of your audience, meaning that you’ve already satisfied most of the people who would be interested? From inside the Kickstarter, it can be hard to tell. You only see the ones who wanted in. Ones who didn’t are effectively invisible to you, whether they might want to buy your product later or not. On top of that, a watch isn’t something people are going to feel the need to replace that often, even if a new one comes out with better features, so many of the people who bought your first product probably won’t feel the need to upgrade to an even better one.

As one of the earliest smartwatches, the Pebble appealed most of all to tech geeks and early adopters—not unlike the early days of the ebook market. The problem is, there simply aren’t all that many tech geeks and early adopters, compared to the overall marketplace. That’s why the ebook market stagnated until Amazon figured out how to make an ebook reader that would appeal to ordinary people.

A friend with whom I was discussing the matter earlier compared it to the low-end tablet market as well. It used to be that the only low-end tablets you could buy were irredeemably cheap Chinese things that performed poorly and were probably infested with malware to boot. But here came Amazon again, launching its $50 Fire—and now Barnes & Noble has followed suit. Suddenly there’s a big low-end tablet business now.

Unfortunately, Pebble never got to the point of widespread adoption. Other manufacturers like Samsung and Apple charged into the smartwatch market and started making watches that were much fancier (even if their battery life often wasn’t sufficient to last for even one entire day). And while it might be enough for a tech geek to do little more than read text messages and use a limited selection of apps on a black-and-white screen, the majority of the audience was going to want to pay less for a Fitbit which didn’t do much except tell the time and track their footsteps, or pay more for a Samsung or Apple watch that was more like strapping a miniature smartphone to your wrist.

As another friend mentioned during that conversation, Pebble’s highly-successful Kickstarters were basically a bubble. They made it appear there was demand, but didn’t actually create much of it beyond those people who kicked in at the time. Effectively, just about the only people who were excited enough to buy a low-end semi-smart smartwatch already had. If Pebble’s management had been smart themselves, they would have taken that offer from Citizen, if it really existed. Surely they must have known by then that their watches weren’t selling in the kind of numbers they needed for long-term survival. But then, hindsight is usually 20/20, and it can be hard to know what to do when you’re actually in the moment.

Meanwhile, Fitbit is in the awkward position of seeing its own potential market evaporate, as more and more modern smartwatches are incorporating extensive fitness-tracking functions of their own. If you can pay either $100 for something that only tracks your fitness or $300 for something that tracks your fitness and does about a million other things, why would you invest so much money in something that can do so little?

Clearly, Fitbit wants to become a player in the smartwatch market itself, and is acqui-hiring much of the staff who built the first successful smartwatch. But the fact that they were able to do so this cheaply doesn’t seem to bode well for their chances against the companies that are already better-entrenched in the market than Pebble ever was.

In any event, the sponsors of other hardware crowdfunding projects could probably learn from this failure. For example, Michael Kozlowski’s 13.3″ e-ink tablet IndieGoGo project had raised $270,499 by the time it closed—over 330% of its goal. And if he can deliver the tablet those backers expected, good for him. But that doesn’t mean he should try to go into business manufacturing 13.3″ e-ink tablets and putting them into Best Buy, because it seems likely that most of the people who would be interested in such a thing have already put their money down.

Crowdfunding might be most useful for something like self-publishing books and ebooks, because a book has few marginal costs after it’s been edited, formatted, and assembled. Ebooks have zero marginal cost, but even the print version only requires spraying ink onto dead trees, then slapping a little glue on to stick the whole thing together. The biggest costs are in putting the book together in the first place, and those costs are covered by the original Kickstarter. (If the sponsor of the Kickstarter is wise enough to set their campaign up to succeed, at least.) Keeping the book in print after that is effectively free.

But those who kickstart expensive hardware projects might do well to remember that this hardware is going to continue costing money to manufacture after the Kickstarter is over—and if nobody wants to buy more of it, continuing to make more is just throwing their money away. It might be best to treat Kickstarter hardware campaigns as one-and-done—at least until you can be sure enough people do actually want your product for you to continue making it.