A story about DO IT AGAIN
Dear Taro Backers!
I am Hao, founder of TARO, after 5 days trip to Shenzhen, China, I am back to our office now. There is a super good news to announce: after countless times of meetings and negotiations, we have successfully finished the selection process of all our suppliers, which means that TARO finally steps on the takeoff runway.
Here I want to say sorry, sorry for keep you waiting, but I would more wish to say thank you, thank you for standing with us from zero to today, and thank you for saving us many many times, even at the moment you don't realize. I could finally begin to tell the long story now:
Inventing something is hard, and inventing a new hardware for a group AI software geeks would be very close to a mission impossible.
On Oct. 2017 when we 3D printed our first prototype of TARO, we thought that we are just one step away from the final product, so we launched our KS campaign, and thanks to all your support, we successfully completed our campaign, and went through to the actual product test phase, at the beginning, we even believe that we could deliver the product before our promise date, however, the real world quickly welcomed us by a big slap.
In our KS page, we promised the tracking distance would be 20 meters, which we believe would be the perfect distance for most of the games and sports, however, I have to confess now, we did not do enough tests at that moment, so when we actually do that, we found out that, TARO fails, not a little, but totally fails on 20 meters tracking.
At the beginning, we didn't take this as a serious problem, because we believed that by simply increasing the power of our tracking tag, we would just fix it, however, when we really did that, we sadly found out that if we increased the power to the value which is good enough for our old tracking algorithm, our tracking tag would be hot enough to burn the egg after 5 minutes work. Then we tried every small change we could think about, and all of them failed, which left us with only one choice: redo all the designs and algorithms, in a word, DO IT AGAIN, remember this word. This is the most horrible word our whole team could imagine, and the even worse news is that we have to hear it again and again in the coming 10 months.
At that point, our team had a discussion, should we follow the old design and deliver something that only works for around 8 meters and don't postpone the schedule? After a long discussion, we all agreed that 20 meters would be very important to our backers since we had already received their support, it would be too shame for us to admit that we could only deliver them something that is not perfect. So, we decided to DO IT AGAIN.
This is the decision that dragged us to a 3-months' battle to the algorithm, I don't want to get to the details, but I clearly remember the night of Dec. 2017, just before we set off to the CES. At that night, as usual, we worked overnight, and after trying out all the possible ways we could think of at that moment, we found out that, we are totally exhausted. We could not even figure out a new way to meet our target. Our chief programmer, recked helplessly on the chair and asked: Hao, shall we just give up and refund the money? We had done our best, and it is not possible.
I don't know how to answer, and I opened our KS page, and scroll down all the comments, then I answered: There are too many backers there, the job of writing all the canceling emails is too tough to do. Maybe we could have some more tries.
So, we stayed, all because of you.
After this talk, we spend another 3 weeks on all the possible solutions, we failed many many times, and luckily, we finally found out a way to solve the problem. And, after achieving a series of very positive testing results for this new algorithm. We feel released and even believe that if we push the supply chain hard enough we could still meet the delivery date.
And it turns out to be the beginning of another series of nightmares.
Our new algorithm is much more complex than the old one, so we need to choose a new chip to carry all the calculation job. We roughly estimated our needs and chose a nice chip from our online supplier, the price is very expensive but still in our budget, so we happily begin the process of embedded development, to be honest, as a group of AI software engineers, embedded development is something that we learnt in school but not used for many years, so our team simply re-read all the books and cases and to secure the process, we even hired an out-source team to help us on the details in case that we are not familiar with all details.
After 3 weeks of peaceful coding job, we successfully coded our new algorithm on the chip, and the PCB design also arrived on time, so after one week of waiting on the SMT job, we are ready to do the test. This is where the nightmare begins: the new algorithm to too heavy, it requires too much calculation and thus cause very high power consumption, our battery would easily run out of juice in half an hour, OK, here comes to our promise again, we promised to you, we would last for 2 hours. And the even sad problem is that the chip is very very hot after continues working for 20 minutes, it even requires a radiating system. Those 2 big changes made our whole set of industrial design to go to the trash bin.
And it is after we start to do TARO, that we realized that a whole set of industrial design is not only the structure of the product, but also the container case, the outbox, and even the process of assembly, which spent us over 2 months to complete, and because of the chip problem, guess what, we need to DO IT AGAIN!!!!!
Part of our industrial design job is outsourced, and since we want to meet the delivery date, we pushed them really hard, actually they warned us many times about the uncertainty of our project and suggested us many times to begin the industrial design only after we confirmed our IC part, however, we have a delivery date to meet, so we ignored all their suggestions and pushed them really hard on all the related jobs. So, I guess you could imagine, what is their feedback when we tell them this three magic words: DO IT AGAIN, to be honest, I very appreciate that they let us walking out the meeting room instead of lying down.
OK, I don't want to recall the memory of that chaos period, "nightmare" is the most vivid word I could choose.
After 2 weeks of refining, the new set of design is almost ready, new outlook, new battery, new box, everything seems to be perfect except one: the chip. We first get this chip from an online supplier, because they are the quickest source to deliver. And when we began the process of production, we turned to the official supplier of the chip, and the price we got there was almost triple the price we have online. We were super angry and believe that the official supplier was trying to blackmail us, so we tried hard to find another supplier and even found a friend in the chip company, but all the price we got was almost the same. So we naturally doubt why we could have such a cheap price online? After a long meeting, we understood that what we got at the beginning was a second-hand chip, which had no quality insurance and no stable supply, and we came back and verified, Yes, it was a second-hand chip, and we were very lucky that it didn't break in our tests.
So..... we were facing this big problem again, we need to change the chip because it is too expensive to afford and in order to change the chip, we have to change our algorithm, and we have to change our IC design, and again we had to face our industrial design outsourcer and let them help us to redo the design. It is a big mess! After the exhausting work for more than a year, we are back to the starting point now, it is not one thing or two things we need to redo, we need to DO THE WHOLE THING AGAIN.
I don't remember very clearly how I explained this to everyone, but I do remember most of the conversation closed with something like this: We had to do this because the backers are waiting, they have waited for months, we have to deliver it.
We feel extreme graceful for everyone in our team and everyone in our out-source team, after hearing our promise to the backers, they all understand and fully supported us. We have done projects many times, before TARO, every time when we faced such a severe fail, we have to quit.
But we stayed this time, all because of your support.
Because this time we are not doing some cool geek projects, we are carrying the trust of all our backers, you chose to support us just via our promotion video and just through our KS page, you didn't actually know us, but you chose to believe in us, so we could not quit this time. It is not only our project now, it is the project we own with our backers.
After that, we chose another affordable chip from the official supplier and greatly simplified our algorithm, it was very lucky that after this simplify process, our performance had not dropped but even increased, however, even after simplifying, it would still take over 240% of the chip's calculation capability if we follow the standard way of embedded development coding. So our embedded coder has to work sentence by sentence to dig the possible room for saving calculation capability, and lucky we did it.
So here comes to today, that we finalized all our design and stepping on the pre-production phase. We know that what you care about is the delivery date. But to make this date very accurate, we need to collect the feedback from all our suppliers, which might take 2 weeks. And we decide that we should make this process more transparent, so we will do the bi-weekly update on the production schedule and let you sync the process of the production. Our estimation of the delivery would be the end of September, but, we would let you know.
We fully understand some of our backers could not wait for longer, so we promise to do the full refund, and we have done all the 12 refunds in the past months. So please let us know if you want the refund now.
Back to Oct. 2017, when we answer the comment with "Thank you for your support" it was no more than a polite word for us. But after those whole struggle months, we understand that your support is not something in the air, but it the big reason to keep us stay if not the only one. And we want to deliver a special thanks for Kickstarter to build such a tolerant and supportive community.
We would release some important updates very soon since we have almost nailed our supply chain.
Finally, and sincerely, Thank you for your support!
2018-8-2 Hao Qian @Home