What I don't understand is things like:

    - What's required to justify this cost?
    - How many features and updates does the app require?
    - What could the ongoing server costs be?
    - How many people maintain the software?

I've built some IoT projects and handling events from the hardware was remarkably inexpensive. Piping tiny telemetric packets, even at a high frequency, was no big deal. It wouldn't justify charging customers $20/month. Maybe $2.50?

Plus, these things are only piping out data when they're in use, right? So... Only 1/3 of the day, if that.

Then the feature set, who knows. Is it just a readout with some fixed controls for the firmware in the eight sleep?

How is that justifying $20? Every single month?

I know software (especially when hardware is involved) can be more complicated and demanding than it appears on the surface, so these are genuine questions. I'm very open to having bad assumptions here. It just doesn't map to my experiences properly. Especially since the customers pay a premium for the hardware upfront.

I guess if customers are willing to pay, it's fair game.

Indeed, it’s about what consumers are willing to pay, not what it costs to produce. It’s called value-based pricing.

Do you have the same reasoning with cigarettes? $10 every day is a small price to pay to avoid having to stop smoking.

Cigarettes is a consumable resources, as for any resources like that it has fifferen justification as you cant produce it.

The subscription for bed is not, it locks artificially features to pay monthly. Even more, it collecs data to improve the product (which sounds good) - but you need to pay for this. They have an ability to run model locally - they choose to not.

I like Topaz approach: you have an ability over some time (subscription period) to have up to date model that will help you recognise snoring etc, then if you choose not to pay - you stick with this model, but it still works.

Subscription in addition is something that limits an ability to sell it in the future.

I don't follow, sorry

in a way, yes. 20$/month to marginally improve sleep efficiency can be worth it, especially when you have high energy expenditure and need to be able to keep up.

on the other hand, paying 20$/month for the right to use the bed, that your purchased at 2000$ cost is a ripoff.

sleeping isn't costly, has never been, yet a company is trying to enforce it and i can see how it doesn't go well with most people.

How much can this help?

Depends person to person. For me it's the difference between waking up 6-8 times throughout the night, and sleeping for a sound 8 hours without interruption. For my wife, not much difference, other than we are able to sleep together, where as before our wildly different temperature tolerances meant separate rooms. I've seen a few people in this thread state it negatively impacted their sleep.