In this interview Eric Migicovsky recounts Pebble's product-stage journey from late 2009 to 2010. He explains how the team used a clever free‑preorder beta launch to qualify high‑intent leads without taking money, then describes the modest cash burn on conferences, web hosting, and contractors while building the software and hardware. Migicovsky reflects on the pros and cons of contractors versus employees, the strategy of deferring legal fees and handling accounting in QuickBooks until the team grew. He outlines filing a provisional patent to secure an early filing date and the low‑cost, hands‑off approach to hardware prototyping in China—using local designers, promotional‑item factories, and a simple test‑jig workflow. He shares concrete cost figures: $5‑6k setup, $60‑70 per watch cost of goods, $150 retail price, and a 35% deposit structure. The narrative concludes with early prototype methods (3D‑printing, CNC milling) and lessons learned about lead generation, budgeting, and manufacturing for a first‑time hardware startup.
Chapters
For us, the product stage was launching the product publicly to find out whether people were actually interested in what we were making. We did a bit of a clever thing. We didn't have the product ready to sell, but we also didn't want to take money because we had never made a product before, and we had no idea how long it would be until we were going to have a product to sell. So we launched— A beta. It looked like a web store, but it had Impulse, the product, and you can kind of view the features and see some renders and stuff like that. And then when it said preorder or buy, you click through and it would kind of have that e-commerce flow. But at the end of the line, there was no credit card section. So you just, instead of buying it, you just preordered it for free. But we found that the quality of leads that came from that e-commerce style checkout were really strong because people thought that they were actually going to have to put their credit card in and buy the product, but at the end of the day, they just placed the preorder for free. So we got the benefit of knowing a hard lead on someone who wanted to buy our product, but we didn't have to take the responsibility of having their money without exactly knowing when or what we were going to ship. So we did this soft launch at the end of 2009, and we weren't actually able to ship the product to our very first customers until September 2010. So it was a good thing we kind of had that framework in place. So between the concept stage where we had this prototype and the day that we actually started shipping to beta customers, we started spending more money. We had a bit of a very small amount of money to live on. We had to travel to conferences. We didn't have to travel to conferences. We did travel to conferences. I guess that's a good point. Probably in hindsight, the conferences were a bit of a waste of money. We needed web servers to run our hosting. We ended up leaning fairly heavily on contractors to help with some software bits that we weren't confident that we could do. This was a mixed bag. Contractors are excellent in terms of being able to supplement your skill set in the short term, but The real disconnect that we had with contractors was that at the end of the day, we're the only people that care about the company and the product. And no matter how much money you pay people, you can't make someone care as much as you do. It's a little bit different with employees who are invested in the company and usually go through a little bit of a longer hiring process or a more engaged hiring process than a contractor. But I definitely had some pretty rough experiences with people who were just doing short-term work for me. And yeah, there's no real surefire way to be able to detect whether a contractor will be good, but I think it's a skill, an underrated skill. And if you have someone on your team who's pretty good at detecting that, then it's quite a useful skill. Didn't end up spending any money on legal in the early stages. We found a lawyer who was able to help us defer the costs. One thing I did ask is that the lawyer would tell us how much he was deferring so it wouldn't be like a huge shock at the end. Accounting, cheap, you know, someone to help do the books. I ended up managing most of the QuickBooks myself in the early stages and then worked with an accountant to do taxes at the end. Depending on, I guess, the number of people, like right now we have a bookkeeper because we have 18 employees and it's a little bit more moving pieces, but I was able to manage it with QuickBooks up until about 4 or 5 people, and then it became a little bit too much. Marketing, I think we made one of those roll-up banners to take around to trade shows and stuff like that, and business cards and t-shirts, pretty much all we did there. The office was the garage behind the house that I was living at, so I rented it from my landlord. And we ended up filing a provisional patent back in 2009. Provisional patents are what you do if you think you're going to get a patent, but you don't want to spend $10K or $15K right up front. So, you file a very rough outline of what your patent will be. You don't necessarily have to engage a lawyer. We ended up engaging an engineer who worked at a law firm to help us, guide us through this process. Couple thousand bucks, you get— the benefit of a provisional patent is you get an early filing date. So, I think our patent has a filing date of like fall 2009, because we were able to execute this provisional patent fairly quickly. Then you have a year to follow up with an actual patent. And that's when hopefully you'll have an idea of whether the thing you're trying to patent is what your company is, and hopefully have a bit more money. I'm still undecided on the value of patents, but we felt like it couldn't be a bad thing to at least get an early filing date. On the hardware side of things, we actually had a design by this point of what we wanted the shipping product impulse to be. We worked with a combination of manufacturers locally in Waterloo and a couple factories in China that made promotional items. Now promotional items factories are like the worst bottom of the barrel factories that you can think of. They're the people that make like $2 MP3 players or something like that, or wristbands, or free junk, but they were priced right for us at that early stage, and our industrial designer actually had designed Christmas promotional gifts for another company, so he had this network in already, and we were able to get quotes really quickly from a metal vendor, a plastic vendor, and an electronic PCB vendor in China. It was very hands-off. We kind of just emailed someone documents, and then we'd get a quote, and then 8 months— 8 weeks later, we'd get a FedEx shipment in the mail with prototypes. We didn't actually fly to China in the early days. It worked out well enough for us. The quality was pretty low. There was a high failure rate on the electronics, but we didn't spend much money, and we didn't know anything about manufacturing, so we really didn't know how bad it was. So overall, it worked out okay. We made the circuit board in China. They would send it back. We actually FedExed them a little test jig so that they would take our circuit board, put it into our test jig, and if the light came on in the right order, it would be good, and they would ship it to us. And if it was bad, they would just keep fixing it until it showed those lights. That whole China outsourcing process cost us about— I think it was about $5,000 or $6,000 to set up, and then the cost of goods Which was, I think it was like $60 or $70 per watch at that time. And we were selling them for $150. So we had to pay a deposit on the order. And that was 35%. And then we'd pay the final amount when the goods were shipped and qualified by us to be in fully working order. This got us to the phase where we had— we made about— Well, these are early sort of sketches of this next generation of products. The early prototype where we had a 3D-printed case, and then the electronics inside, and then a render of what the product was going to look like. And then this is actually an early CNC-milled prototype. So CNC is computer-controlled milling.
