A handwritten fiasco

Emerging from the ashes of my crash landing in China, in late 2003, desperately trying to build an industrial automation consulting activity like the one I lost in Italy, I got an idea and I fell in love with it.

Soon enough I realised that there was no space for me to sell automation services, especially software, in a country where the customers I was visiting were asking me: “yes we like automation, how many kg weights and what colour is your automation?”.

You cannot sell something like automation consultancy unless you seal an heavy stone in a box and put a label “AUTOMATION” on it.

I was in desperate need for a product to give a measurable weight to my expertise and make it saleable.

I was searching a broad spectrum of possible products that could become my automation Trojan-horse when by chance I stumbled upon something that was nothing better than a geek gadget. I found a pen paired with a clip. The clip could hold a piece of paper and anything the pen was writing on the paper held by the clip was automatically transferred to a PC to which the clip was connected with an USB cable.

The great idea

The guy introducing the product was showing me how he could sketch something and add an handwritten note and then the software was making a picture and a text file of the image and the handwritten text in it. In that moment the idea flashed to my mind.

With the eyes of my mind I saw a worker in a factory filling a production slip with the pen and all the information in the slip were being instantly fed to the IT system in the factory assuring 100% traceability with 0% changes in the manual procedures. For an expert in shop floor IT and data collection that was the best possible breakthrough.

I could also see receptionists at fair and events handing over forms and customer filling them with my system assuring 100% of data collection with 0% of manual data entry work required.

And so on, I was seeing applications in shopping malls, restaurants, banks, public offices and wherever there were forms to be filled manually

The advantage of immediate availability of data entered by handwriting, and the fact that it also had time stamp and secure identification of the pen used and its location was obvious and self selling.

I also figured out that, although the technology was all but 100% accurate, its limit was that it was thought to be for general purpose application and therefore not good enough for any of the purposes.

But I had an ace up my sleeve: the pen only had to recognise a limited selection of words within predefined grids or even just ticks on check boxes and it was definitely possible to specialise the technology to reach nearly 100% accuracy on those limited tasks. Moreover my software could very easily detect exceptions when the recognition accuracy dropped below a certain level or there was some inconsistency in the data collected and then submit the raw data collected from the pen for human revision, integrating thus limited human detection to make up for the fraction missed by the automation while still learning and improving.

The beginning of fall

A great idea, right? I was selling my vision upon it and I got a round of financing from friends around the world. I set a target of roughly 50,000 US$ for the first stage of development to get to a working proof of concept. I managed to get less than half of it.

That was the first signal that I had to revise something, but I was in love with my idea and I kept running according to my plan hoping that early success would have fostered more investment.

Then the time came when I started seeing a broader picture. All was very nice, but how was my system going to be able to know what kind of form was being filled and what was already on it? In production one form travels with the product being made and receives information at different stages. The pen was not able to detect what form it was.

I started thinking of complex ways to encode some sort of ID that could be entered again at each stage, but then all point of 0% change in procedures was lost. I imagined adding bar codes on the forms, but again that would have impacted on the existing procedures and added complexity.

I could use a different technology that used special paper capable of solving all my problems, but I rejected the idea . It would have been more expensive and not as elegant as my original solution.

So I spent out all the resources I had from the investors and my own trying hard to perfect a flawed idea until I had to accept failure.

A lifeboat

When I was still enthusiastic about my idea I eventually managed to convince the production director of an Italian furniture maker with operations in Shenzhen to apply my technology within a traceability system for their factory that grew in 3 years from 1 container per month production to 50 containers per day with more than 3,000 workers. They were facing all the obvious problems that such a booming growth created on a shop floor still managed by a bunch of foreign specialists managing local labour.

Within 9 months, with a team of 6 engineers, we built a pilot system and we managed to have reliable data collection and means to replicate the solution for the entire factory.

When I got the first batch of data I went to the boss and showed him. He looked at me and said: “ your system does not work, because if your data are right I am gonna close down in less than one year. Go back to the drawing board!”

Instead I went to the factory and spent 10 days collecting data manually. Then I was able to validate the data from my system and went again to the boss like: “Sorry boss, these are the data from the system and these the one verified with the actual operation. They are the same”.

What happened next is that I got paid all dues for the pilot project, the project was turned down and one year later the factory closed.

The lesson I learned

• Do not fall in love with your ideas.
• If an idea seems perfect and you don’t get the investment you expect then either others see the idea is not perfect or you are not good enough to show it. Anyway you are wrong and should reconsider.
• Be pragmatic, accept compromise. With a great idea you are providing a great solution to a problem. Even if your idea turns out to be not the right solution the problem is still there and you are in a better position to find a new fresh solution.
• Recognise the value of what you have so far and find a way to sell it. At least it will pay your tuition fee during the learning curve.