A high tech toy for pets. I have a lot of experience in software at Unicorn telcos and also some other big names ...
But my god 'hardware is hard'.
Just some quick points for the uninitiated:
There are 'creative' aspects both to design, and to play, and then of course issues with pets. Industrial design to make the physicality work, some high tech secrets, a lot of prototyping on Arduino/Raspi. Hardware customization, PCBs, optimizing for manufacturing ... getting many pieces to fight just right together (I mean product/market mix, i.e. features, price - not just physical) with long dev cycles, waiting for parts, trying to predict the cost of various things at various quantities. Setting up relationships with overseas manufacturers - which is partly a product issue ... but mostly - working capital requirement.
If you're well funded up front, you can get 'very expensive' units built, put them in the hand of journalists, bloggers who will write about it. Even better - get your first production run in place and give them those to coordinate with first sales run. But who invests in hardware startups enough to support working capital? Not many - it's hard enough just operating on regular Angel ... but coming up with the money for the first 'x units' is a huge bet. If you can't get prototypes in the hands of buyers, bloggers (and good ones) it increases risk quite a lot.
So the biggest issue with factories is not so much 'getting it made' because ultimately if you're careful you'll definitely find someone competent to make it ... but it's how they get paid, which is related to the effect of the 'hotness' of your product. Factories are not interested in 'one off' runs, they want something that's hot, the hotter, the better the terms. Things like moulding etc. which drive up the initial unit costs - factories will eat the cost of this if they really want your business. And of course paying for the production ... if you have something white-hot or an operational history - they might give you credit or much better terms, which makes getting to market more accessible.
Finally there are the channel issues - buyers that want guarantees, the ability to return product ... retailers who don't put those terms of front but return a bunch of inventory anyhow. Just because. Every buyer wants a 'unique' aspect to differentiate their offer, but it's hardware, so that's hard, you have to get creative etc..
Online channels are quite fundamentally different than retail - you can sell at higher price points online to cash flush buyers, but in-store, things are extremely price sensitive. It hardly matters what you are selling: once you're past $30 with a retail physical thing ... it gets risky to the buyers who want to see big ad spends, lots of social media activity, yada yada.
This later part is why so many really cool products fail. They're just too expensive for physical retail channels: plastic needs to have 'velocity' or it's a big problem, and most 'cool things' are quite expensive. As soon as you have a decent camera, or wifi, or a high end processor ... it starts to get very hard to get prices down to a retail happy point.
And so many products tend to be 'hit driven' meaning it's hard to turn all of that into something more long lived. Consider how many toy companies come and go - they are so often 'betting the whole company' on every new line.
Unless you have some kind of well oiled machine with all of those things in place, it's hard to break into - you kind of need a hit, and serious backing/believers in order to create the initial momentum. There are a few SV companies who've done that, kudos to them.
Very fun but very hard, and very 'business' i.e. a lot less about 'fun factor' 'design', many more raw operational concerns than a software startup.
These are just some quick thoughts, by no means comprehensive.
