IoT implementations are difficult. You have not just the hardware piece, but also the software (firmware and the web+mobile platform.). And they need to talk well with each other. And all that talk needs to make sense to the client. So the complexity increases by multifold.

For my first IoT-led first digital transformation project involved us declaring with after the successful implementation. After all, the hardware was working well, sending strings to the software and the analytics was analysing it well. So what was wrong in celebrating? Well, a lot of things.

We trained the client to use it. She was happy.

In a few weeks, she came back and started saying that our data was wrong. That was not possible. The IoT devices were capturing the data and sending it to us. We were capturing it and showing the analytics. Why would anything be wrong? But she persisted.

Soon we realised that we didn’t know what we didn’t know. Here, indeed the data was wrong by about 10-15%. Mainly due to the way backend algorithms were capturing the jumps (due to the loss of network) and the way it was correcting it.

How could we have avoided it?

Answer: By testing the platform after going live. It’s kind of calibration. Where you actually sit there and physically measure whatever you are measuring through IoT. For 100s of times and scenarios.

100s? sounds crazy? But that’s how great products are built. Believe me this is going to save your time in the long run.