5 ms·

Ask HN: Hardware startup? What surprised you about forecasting your business?

• bjelkeman-again 5y ago
  The engineering discipline in getting iterations done in hardware must be much more rigorous than in software. Get people used to designing hardware, or it comes back and bites hard.
  • pedalpete 5y ago
    We've been very fortunate to work with a great hardware engineer, and we've taken a "go slow to go fast" mentality. Building larger than necessary boards where we can cut traces, use 0603s so we can replace caps and resistors if necessary, and then going smaller once we've had that first gen working.
• GianFabien 5y ago
  It is hard to respond to your question without knowing where in the development cycle you are. Have you got working prototype? Are you at the ready to manufacture stage? etc. In my experience, the firmware development cycle is far slower than for native software. The interactions between electronics and the code can be very difficult to debug. Getting the power source to work reliably. Manufacturing the case in volume can hard to ramp up. Injection moulding, assembly, retail packaging, etc can cost more to develop than the product. When you write that your unit economics are good, I wonder how far along the path of being ready to manufacture, package and distribute you have gotten. Any team experienced with manufacturing consumer electronics at scale would have prior experience with all the necessary steps. If you are looking at sub-contract design/manufacture, then you need to make a very substantial up-front investment and guarantee certain volumes, etc.
  • pedalpete 5y ago
    It's interesting, we've heard this sort of thing before. We've had very few problems with firmware development, though we've been able to leverage esp32 IDF so far. We have working prototype have costed out our BOM and are in the Alpha design phase. We've costed out the materials cost, packaging, tooling. Our industrial designer has helped us to estimate the costs as well and figures we're in the right ballpark. They actually suspect that as we scale, our costs will drop faster than what I am forecasting. I'm trying to figure out, is there some big mystery surprise I'm waiting to knock me down later? How can I be prepared.
    • GianFabien 5y ago
      Thanks for the additional information. Based on that extra info. My next step would be to validate the industrial designer's estimates with quotations from a couple of contract manufacturers. The areas that would concern me would be: problems identified during the first production run, distribution, marketing, sales and support. If your product needs to be paired with smartphones (BlueTooth? WiFi? - you mentioned ESP32) then you would want to make sure you can support several generations of both iPhones and Android phones. Some SmartWatches have basic EEG capabilities and corresponding apps. So I would see marketing and sales as the biggest concerns. I'm assuming you have sufficient runway to address those issues effectively and in a timely manner.
      • pedalpete 5y ago
        Smartphones have ECG, not EEG. EEG is brainwave. Yes, marketing is a "concern" but we're not going for a spray and pray, make 10k devices. We've got a few very specific use cases, and partner channels, etc. TOTALLY agree on the bluetooth thing. It's one of my biggest concerns as I keep seeing almost every hardware device with bluetooth users complaining about pairing issues. I am SUPER concerned about that, and we are spending lots of time trying to focus on making that SUPER solid. But we also haven't really been able to get bluetooth connectivity to fail yet! We keep trying more and more devices, but it always works. I do think BLE has resolved lots of issues over Bluetooth classic, but if you've got any tips on how the hell we make sure that is rock solid, please let me know.
• b20000 5y ago
  your question is too vague, rephrase or come up with a list of clear questions?