Shenzhen Insights from Hardware Startups (French Tech Redux)
HAX (leading early stage investor in hardware) hosted recently an IoT Trends Event at KPMG in Hong Kong. We invited 3 startups to share their experience (see below and also the Hardware Investors Panel). For French Tech we asked two French startups from HAX to add their perspective.
Coline Juin, CEO of Moona
Q: The panelists said Shenzhen had a huge impact on their development.
Speed: Indeed, Shenzhen is a great ecosystem for prototyping and hardware development. In France you have to order parts (from China anyway) and it will take 2 weeks to get them. In Shenzhen you cross the street or get samples within a day. It speeds up the process a lot.
Cost: Testing is also easier thanks to cheaper spare parts, test jigs and 3D printing.
Software: Like some of the panelists, most of our software is done outside China (in France), while most of the hardware is now done in Shenzhen.
Clement Jambou, Founder & CEO of Unsupervised.ai
Speed: Shenzhen gave us the huge opportunity to reduce the time between product iterations. Usually in Europe/France a prototype iteration takes at least 3 weeks and is extremely expensive.
Lower costs and faster speed lead to frequent iterations: Outside China, longer time and higher costs and force startups to triple check everything before ordering any new iteration. Each of them becomes a big risk as you count the iterations that you can make before the end of your runway. In particular in Robotics where complexity adds on.
Manufacturing capabilities: It’s not only speed and money, manufacturing in Europe for small quantities also lacks flexibility: you need to exchange a few times with factories because they often cannot do what you are asking, in China they will ask their neighbor. Things are getting better in Europe with more and more prototyping labs and Fab Labs as well as the French Fab initiative. But these solutions are also problematic to scale to small batches.
At HAX we made great connections with the Shenzhen ecosystem. Since we came back to Europe, we continue to grow these relationships and have built a pipeline that works well both for iterations as well as small/medium batch (up to hundreds of robots). We’re happy to bring back knowledge from the Chinese ecosystem to improve best practices in Europe.
Hardware Startups Panel
- Lamont Tang, CEO, Oyalabs (Best of BabyTech Award at CES 2018)
- Jamie Salter, Lead Engineer, Carv (#1 connected sports product on Kickstarter)
- Florian Simmendinger, CEO & Founder, Soundbrenner (first wearable for musicians)
- Moderator: Karena Belin, Co-founder, WHub
*Slightly edited transcript below, video at the end.
Q: Karena — Can you tell us about your company?
Lamont — we use machine learning to enable children’s cognitive development. Our device listens to every word and conversation to track the development of vocabulary and help parents raise smart and healthy children. We’re backed by J&J and HAX.
Jamie — we’re a digital ski coach, also from HAX, using motion sensors and pressure sensors in your ski boots to see exactly what your feet are doing and give real-time coaching via earphones and data analysis on your phone. We’re not just a tracking device but trying to teach you how to do something better, starting with skiing, using feedback from sensors.
Florian — we are reinventing the metronome by making musicians feel the rhythm. We started in 2014 noticing that there was little innovation in the music space. We focused on rhythm as it’s something every musician cares about. We joined Brinc 3 years ago and want to be the best music technology company in the world.
Q: Karena — When did you decide to set up shop in HK / SZ?
Lamont — we started in HK with research in SF. HK has a lower cost of operating and you can prototype 5 to 10 times faster with SZ.
Jamie — We came to SZ 3 years ago, I have been living in SZ for about 2 years. There is no better place to be to develop hardware. I’ve looked into moving to HK because as a foreigner SZ is tougher socially. Hardware is in SZ, software, business and everything else is in Europe. Persuading team members to move to SZ is difficult because it doesn’t have the infrastructure — like joining sports teams, etc.: I commute every weekend to play hockey in HK — so if the Greater Bay Area (GBA) helps cross the border I’d love to live in HK but right now living in SZ is very fast.
For example we’re designing our version two. A week ago I messaged a supplier on wechat in the morning and about 3 hours later he said “check with reception”. Then I realized I ordered the wrong part. I messaged him again. About an hour later the new part was there. This ecosystem is the world’s capital of electronics so for me there is no replacement to SZ.
This ecosystem is the world’s capital of electronics. For me there is no replacement to Shenzhen. I can get parts delivered in an hour.
Q: Karena — Why is the software still in the UK?
Jamie — Mostly because you can’t ski in Shenzhen. All the hardware is designed in Shenzhen. If you design here, you can scale and validate here.
Florian — the core to our first product is the vibration motor. For our first prototype we used the motor of a Nokia phone. Then we looked for the best motor in all the electronic shops in Berlin… which only has two shops. We also ordered one from eBay.
…Then we came to Shenzhen to Huaqiangbei and an old lady with a cart on wheels had 100 motors — every motor ever built into a consumer electronic device. My cofounder and I looked at each other and realized ‘this is where we need to be’. We thought HK made sense for hardware and biz dev. We could hire a marketing team able to talk to both U.S. and Chinese customers. They can fly to Los Angeles to talk to artists.
Q: Karena — How about speed?
Florian — speed is huge. On crowdfunding sites you can see who’s in China and who isn’t based on their delay.
Jamie — we just hired a production manager. We needed a bilingual person. We used Chinese job sites and LinkedIn. We had about 100 applicants, had calls with about 30 then 10 face-to-face. 9 out of 10 were from HK. Some were mainland Chinese based there. HK seems more suitable for recruiting. We see Chinese teams at HAX and it seems very easy for them to hire. But if you’re looking for bilingual engineers you can find them in HK. We hired someone who lives in HK, has his kids in school in HK and commutes to SZ a few times a week.
Lamont — for one engineer in Silicon Valley you can get 2 to 4 engineers in HK or SZ. That said one engineer in HK we told he had to travel to SZ said ‘then it’s not a job for me’, so some are still reluctant. Compared to SV, it’s much easier to recruit hardware engineers here so most of our hardware development is here, while business and natural language processing is in SV.
For one engineer in Silicon Valley you can get 2 to 4 engineers in HK or SZ. It is also much easier to recruit hardware engineers.
Q: Karena — What were some of the disappointments?
Lamont — we are a US-based company with a HK subsidiary, which can be a turn off for some investors in SV. In HK investors are more late-stage. In mainland we found quite a few suitors. We are still trying to figure out how to work the ecosystem.
Jamie — I was hoping we could go to a factory with a prototype and we would figure out a process and it would be done. But if you’re developing hardware on a budget you can’t go to tier 1 factories. So you go to tier 3–4–5–6 who accept to work with you on low volume because they’re not very good. So it’s a bit painful. You need to put in a lot of work to ensure you get a quality product.
You have to spend enough money to not go too far down the tiers. The people who try to leave China end up having communication problems and find them when prototypes come back once a week or once a month. It’s much better to be in the factory to review what’s going on and drive the changes.
We shipped 2,500 units last season, 5,000 next season with $600k total investment from design to shipping, which is basically nothing. The only way we could do that was by being in the factory.
You need to be close to factories and put in a lot of work to get a quality product.
Florian — our accelerator told us about the pitfalls. We had trouble finding app developers so we do it in Berlin.
Q: Karena — how about IP?
Lamont — we’re a software / A.I. company so we actually encourage if someone can make the hardware cheaper.
Jamie — it’s all in the software.
Florian — software is the way to go. It would be very risky to build a pure hardware company like GoPro today.
Jamie — finding [bilingual] data scientists and software engineers in China is tricky so our team is in London where there is a good community.
IP? It’s all in the software.
Q: Karena — your consumers are in the West? Can the GBA help access China?
Lamont — we do our market entry in the US, then APAC.
Jamie — there will be 300M Chinese skiers by 2030, doubling the ski market. Ski companies are looking at it closely and Beijing Winter Olympics are coming too. Distributors hinted that Chinese customers need a lot more customer support — 24h wechat access, for instance — which would require a lot more investment.
Florian — tigers moms get their kids to learn piano and 100M kids learn music. China will be our #1 or #2 market. We are looking into entering the market in a big way.
Q: Benjamin (HAX) — this is the real insider’s panel — people on the ground building products. I also noticed a common thread. We hear a lot of talks about A.I. taking over the world and robots killing us, but here we have three companies that are using A.I., machine learning, sensors to ‘close the loop’ and teach humans: teaching language, skiing, rhythm. Those are companies building tech helping us improve. Tech can make us better.
AUDIENCE Q&A
Q: If you were to start again, would you do it any other way?
Lamont — we’re very early stage. Ask us in a year or two.
Jamie — we made lots of mistakes, but mostly right. We could have given a more realistic shipping date to Kickstarter backers.
Florian — we did a pretty good job, like a toddler learning to walk.
Q: Anson (KPMG) — Give us a buzzword and your thoughts about GBA
Lamont — ‘Human capital’: the core need and mission of our company. And the promise of GBA is growing.
Jamie — ‘Closing the loop’, using sensors to give feedback. If GBA gives access to better lifestyle, all the better!
Florian — Music is the mission of our company. For hardware, I think the key is ‘vertical integration’: hardware, software, content. And you get a business like Peloton (a hardware unicorn).
Karena — ‘Passion’. And the GBA potential is not about the past but about the future.
