During the Nasscom Emerge 50 awards [India has the world’s third-largest startup base–meet the product firms exhibiting at NASSCOM Product Conclave 2017] I got into an argument with a Japanese entrepreneur. He was a bit unhappy that we were not able to make it to the CEATEC expo [missed due to a visa issue] that he helped arrange for our startup and get media exposure and business expansion in Japan.
It turned into a discussion on what we should be building. The gentleman was of the view that we should not be building hardware and focus more on building software. [As stereotypes go, this is an “advice” I get a lot]. Indians, while we can do good software, don’t have much experience doing hardware. He was of the view that while our robot might be good enough for India, Japan has a higher hardware standard. Thus, he advised us to port our software to a Japanese robot company and use that to enter the market.
While I agree that our hardware doesn’t match Japan’s standards now, the goal for us is to reduce the delta between what we have and what Japan wants in a progressive way. However, we cannot take the software-only approach, for several reasons.
First, the issue is that even Japan has only one commercial robot in our category [Pepper robot from Softbank] and that means we don’t have very many hardware options to “port” our software into. Pepper is completely booked with past orders and Aldebaran has nothing to spare. And it is only sold in Japan, making it hard for us to import and build software for.
Second, Pepper doesn’t have many of the features we seek. It is quite short – forcing people to bend down for interaction. We rely on face recognition and thus the camera has to be at the height of an average person. Its wheels are quite small – limiting its movement and it doesn’t have provision for autonomous navigation [we need to retrofit our sensors and that would spoil its aesthetics].
Third and most importantly, I don’t agree on splitting the hardware and software – especially in something as integrated as ours. I have seen the folly of this at Microsoft where the quality of the OS on Dell/HP was substantially inferior compared to the homegrown hardware of Surface Pro. And in robotics, the variables are a lot – from having right cameras to right microphone array to right sensors for navigation. Splitting the hardware and software would be disastrous for product quality.
Oh and we want to do hardware. That is our passion and something we want to fight traditional stereotypes and build a successful startup around it.
The discussion then went into how Pepper is superior to us in hardware and appearance. There is no question that Pepper looks better in appearance.
But, Softbank Robotics is 100x our team size, 10x older than us and backed by a parent organization that has 100,000,000 times more money that we have in our bank. Not to mention the absence of government support or infrastructure we have to work in India. The fact that he and our prospective customers are even bringing our products in comparison is flattering for us.
Perfection is not built in a day. It requires time, money and customer feedback. Three of the best companies of our time that are known for their design started like this.
And it makes perfect sense. You don’t have fetuses that look like George Clooney or Aishwarya Rai at their prime.
And this takes us to the Lean Product approach.
If we wait for a perfect product, we will never enter the market in time. The goal of the lean methodology is to build a good enough product – called the Minimum Viable Product – with a unique positioning. Apple, Google and Amazon didn’t wait to build a perfect product. They came with a good product and rapidly innovated. In a few years, they went miles ahead of anyone in their category.
For us, the unique positioning is a commercially available human-sized robot at a viable price point that can navigate around, recognize faces, get into conversations and provide product recommendations. There is no one else that satisfies all that and that is the niche we are getting.
Once we prove the hypothesis that this is a large, viable market we can always get investors to bring in money and substantially improve the quality of injection moulding process.
You can only judge quality by looking at the past. You should have seen our product last November. The first version is substantially backward compared to what we are launching for 1 year later. Sneak preview of what is coming next for Mitra. And as we learn manufacturing processes and make revenues, we can go for better moulding process and refine the design.
Again, this is the same approach the Japanese took in the 1940s-60s. Their products were not known for their quality in the west, but their companies boldly challenged the stereotypes. Sony brought a simple transistor radio to the west that was not anywhere near the music quality of B&O or Marconi. But, good and cheap enough to create a whole market segment. Same for Toyota taking on the likes of General Motors, Benz and BMW. They didn’t start out with making amazing cars but ended up making amazingly durable, efficient cars that fit the economic needs. Same with Honda took on the bike market dominated by the likes of Harley Davidson.
Perfection is not daydreaming of an imaginary product, but a process of constant improvement making sure the product of yesterday looks worse compared to the product of tomorrow.