Your "guess" at the end... regarding commercial vs. hobby... like anything I do, "that depends". The impetus, as usual, is for my own hobby application. But I've had reasonable success in this world on the principle that "I'm nothing special". No, that's not a self-confidence issue that my mother has to worry about... its a business mantra that suggests if I have a need for something... so do others. Its virtually impossible to be unique in this world. The market may not be large (a relative term if there ever was one) but in my ventures so far, I've found them to be "large enough".
Take the two examples I touched on earlier:
1) An automotive cruise control. I bought a car. It happened to be a rather quirky car, but has its cult-like following. For whatever reason, it isn't offered with cruise control - either standard or as an option. I wanted cruise control. To my surprise, none of the existing aftermarket cruise products (costing $80 to $300) would work because of the particular way this car operates (100% drive by wire - getting more common). That meant the cruise control had to actually communicate with the car's computer to work rather than doing it the traditional mechanical way because the "gas pedal" was really a large potentiometer so there was no direct (or indirect) linkage between the pedal and the engine. What did I know about cruise controls? Nothing. What did I know about cars? Next to nothing. But I learned. My first tip came from a common product called a ScanGauge that plugs into any newer car's OBD port and reads all sorts of data. Right there I knew it was possible, at least, to get the necessary information from the car (speed, throttle position, etc). I learned that this car used CANBUS as its network. What did I know about CANBUS? Nothing. But I learned. So from that I determined I'd need some sort of "microchip"... and I put that in quotes for a good reason... at this point, I had no idea "Microchip" was a brand. That's how little I knew about ANY of this. But I learned. And it so happened, that the PIC was a good way to go. I studied datasheets (up until this point I never knew there was such a thing) and I got the general gist of it. I was going to do it all myself, as funds were very limited, and I searched for some sort of device to hook into my car to read the CANBUS data. I found a sub-$100 unit and contacted the small company that made it with a bunch of questions. Keying in on my novice-ness, he offered to help. For a very reasonable sum, he designed the hardware on this and I focused on the firmware and getting all the data from the car. What did I know about firmware? Nothing. But I learned. I learned C+ from scratch (had some experience in Perl and php programming) and spent a lot of time with my ICD2's help file
I spent DAYS watching the screen of my laptop with hex numbers flying by a million lines a minute as it took in the raw data from the car, which I had to manually decode (absolutely no manufacturer support or documentation). What did I know about hex? Nothing. I learned. I had no wiring diagram for the car... but that's not rocket science. Anything I didn't know, I learned. What kind of a market is there for a cruise control specific to ONE quirky car?? NOT BLOODY MUCH! But that's the definition of "niche", now isn't it. I made 1000 of those units and so far I've sold 800 in a little over 2 years. @ 1000 units, they cost me $70 each to make and, as I mentioned before, that INCLUDES the cost of buying the car, paying the hardware guy, all PCB production and assembly, AND injection moulds for two plastic parts (people told me those would be $40k... I ended up getting them for $3.3k), custom (nice!) packaging - boxes and plastic inserts - very professional, all components and printed manuals. Everything but my own time. I sell them for $500 each and have not discounted any of them except for a few promotional cases... do the math.
2) An EL inverter backlight circuit. I'm not going to go into nearly the detail ("thank god", I know). But similar thing... it started as a requirement for ME. I fly model airplanes and the radio I chose doesn't have a backlit LCD. Its a popular radio and "I'm nothing special" so if I want it, others want it. And I found tons of people that did it... using cheap crappy inverters already on the market along with EL panels that they had to trim to size. And those cheap inverters are NOISY and HUGE and relatively power hungry. This time I did all the hardware myself. I found an IC specifically for driving EL panels far more efficiently than the simple inverter setup, so I went to work... reading, and learning, and doing. This project isn't nearly as profitable, but that's the nature of things sometimes. Its a $60 radio, so people are willing to pay hundreds for an add-on
Still, I make them for about $15 a piece (500 at a time) and sell them for $40 and so far, so good.
My point is this: commercial possibilities are everywhere. Much of it starts with the confidence of knowing it CAN be done (you just have to find a way) and just because you don't know how to do something doesn't mean you can't LEARN!
Take the two examples I touched on earlier:
1) An automotive cruise control. I bought a car. It happened to be a rather quirky car, but has its cult-like following. For whatever reason, it isn't offered with cruise control - either standard or as an option. I wanted cruise control. To my surprise, none of the existing aftermarket cruise products (costing $80 to $300) would work because of the particular way this car operates (100% drive by wire - getting more common). That meant the cruise control had to actually communicate with the car's computer to work rather than doing it the traditional mechanical way because the "gas pedal" was really a large potentiometer so there was no direct (or indirect) linkage between the pedal and the engine. What did I know about cruise controls? Nothing. What did I know about cars? Next to nothing. But I learned. My first tip came from a common product called a ScanGauge that plugs into any newer car's OBD port and reads all sorts of data. Right there I knew it was possible, at least, to get the necessary information from the car (speed, throttle position, etc). I learned that this car used CANBUS as its network. What did I know about CANBUS? Nothing. But I learned. So from that I determined I'd need some sort of "microchip"... and I put that in quotes for a good reason... at this point, I had no idea "Microchip" was a brand. That's how little I knew about ANY of this. But I learned. And it so happened, that the PIC was a good way to go. I studied datasheets (up until this point I never knew there was such a thing) and I got the general gist of it. I was going to do it all myself, as funds were very limited, and I searched for some sort of device to hook into my car to read the CANBUS data. I found a sub-$100 unit and contacted the small company that made it with a bunch of questions. Keying in on my novice-ness, he offered to help. For a very reasonable sum, he designed the hardware on this and I focused on the firmware and getting all the data from the car. What did I know about firmware? Nothing. But I learned. I learned C+ from scratch (had some experience in Perl and php programming) and spent a lot of time with my ICD2's help file
2) An EL inverter backlight circuit. I'm not going to go into nearly the detail ("thank god", I know). But similar thing... it started as a requirement for ME. I fly model airplanes and the radio I chose doesn't have a backlit LCD. Its a popular radio and "I'm nothing special" so if I want it, others want it. And I found tons of people that did it... using cheap crappy inverters already on the market along with EL panels that they had to trim to size. And those cheap inverters are NOISY and HUGE and relatively power hungry. This time I did all the hardware myself. I found an IC specifically for driving EL panels far more efficiently than the simple inverter setup, so I went to work... reading, and learning, and doing. This project isn't nearly as profitable, but that's the nature of things sometimes. Its a $60 radio, so people are willing to pay hundreds for an add-on
My point is this: commercial possibilities are everywhere. Much of it starts with the confidence of knowing it CAN be done (you just have to find a way) and just because you don't know how to do something doesn't mean you can't LEARN!