We decided to take the plunge into Hardware around this time last year, after our early Beta users strongly suggested that we move on from an "App only" model. GPS, IMUs, and sensors in general find an obvious fit in Sports, as they allow for capture of more useful data and incredibly powerful UX.
This post isn't so much about our journey, as it is a list of learnings, lessons & mistakes we've come back from. So if you are thinking of starting a hardware / IoT Startup (esp. global-facing), this will be exceptionally useful for you.
1 - Spend 80% of your first 2 months defining the problem & market:
Unlike apps and software, once you decide your path, there's probably no turning back. You can't push an overnight update and wish away a mistake you make at this stage.
E.g. if you want to build a Child-safety product, should it be a watch, something you can clip onto children's clothing, or a pocket--monitor? Should it be enabled with GPS (but GPS doesn't work without open sky)? Would you sell this to schools, parents, etc? How HUGE is the market? Remember, hardware-mistakes can take several months to undo.
2 - Form is SUPER important:
If you're an engineer, you'll do well to remember this. Function & features are one thing, but the form makes or breaks your hardware product. This is especially true for wearables, and to a lesser extent with other IoT devices.
This is where you want to survey your target demographic and maybe ask them to use a "looks like, feels like" dummy (design the electronics on a parallel track). Remember the uniqueness of your product, and the size, weight, temperature-range, water-proof-ness expectations people might have from it. Build for these expectations.
3 - Make sure you've budgeted enough time to build a GREAT hardware team
I suppose you know that finding the perfect people for embedded systems and firmware roles is far more difficult (as of 2018) than hiring for software roles. It's considerably harder if you don't understand much of hardware to begin with; if you do, then it's a bonus.
But here's one thing that's a STRICT NO: don't hire consultants or outsource your development. You'll probably kill your startup even before you begin. Firstly, anyone who is good will burn a hole through your pocket. Secondly, they'll probably never understand your user the way you do. And last, how many iterations have you negotiated for? What if you need to go through 5 prototypes before finalizing on the design...
4 - Design for Manufacture. And for scale.
You need to think through how you want to produce these (100 units, 1000 units, 10k units, 1 lakh units etc). Best to choose a method that is fully-scaleable to meet demands of the Market you defined in Step 1.
Pro-Tip (I missed out on this earlier): if you can use an existing solution that's in the market - probably not even well known (HINT: Visit Shenzhen), then use it. You may find manufacturers willing to tweak & white-label existing solutions. You focus your collective energies on selling these in India. This is considerably easier (and far more popular) option.
However, if you - like us - choose to build disruptive IP, and a unique solution (because you think it is necessary, not because of vanity), then think through your procurement, logistics, payments, compliance etc. Earlier you do this, the better.
5 - Prototype electronics & mechanical designs. 3D Printing is a powerful gift!
Not enough if you have a 3D model / CAD file / sketch etc. Build it. Build it early. And build every iteration. You'll find new challenges in each one, that you won't realize while designing (e.g. How will this button press work? Component's Specifications sheet and actual dimensions don't match perfectly. Placement, design & routing problems etc).
Remember that 3D printing has multiple types. Choose the cheapest, fastest versions first. Then go for more expensive, representative solutions. Remember to always build what you can manufacture (it is easy to 3D Print complex designs which will cost you a fortune to mold).
Part 2 of this post COMING SOON. In case you have any questions / suggestions / comments up to this point, I'd love to hear from you. Cheers.
No comments:
Post a Comment