How does one become a BI developer?
These days, you can study BI courses in University, or take various technical courses such as the ones we teach at Naya College. But back in the 90’s there was nothing. No-one knew what BI was.
My journey started in high school (1982-85), where I studied computers (the subject was called “Automatic Data Processing – Systems”). We learnt Pascal, Fortran, and mostly Cobol, and my final project was an ERP system for my father’s brush factory. Unlike most of the projects in the class, this one went into production within a few months, and kept working until I upgraded it to Microsoft Access 12 years later!
After my army duty and a B.A. in Mathematics, I discovered that there is no real work in Maths, unless you want to become a teacher (no way) or stay in academics (not really). So a friend of my dad took me in for a trial as a software programmer, and there I stayed for the next ten years. But that wasn’t BI yet…
I was working at Blades Technology Ltd., a company manufacturing blades for aircraft engines, and doing various programming jobs, mostly in Access. One day, a marketing guy from a software company came around to display his product – Panorama, a tool for creating data visualizations. No-one was impressed by the demo, but he agreed to leave us a copy of the software, so we could test it on our own data, and as the junior in the IT department it landed on me.
I was hooked. I connected to some of our marketing data, created charts, and showed how quickly we could now analyze the data. The company bought Panorama licenses, I was put in charge of the project, and soon we had a data warehouse and I was an expert in the field. I continued developing applications with Access until I left there, but my official position by then was BI Team Leader, and I’ve never looked back.
There are lots of additional BI stories from my time at BTL, but the most significant one was probably one of the first.
Our ERP was Priority, an Israeli system that is still around and has a large slice of the market. At some point, I was tasked with analyzing the changes in customer orders that our company was receiving. These were loaded automatically into Priority through some sort of communication interface (this was in 1996, nowadays it’s called an API), and the VP of Sales suspected that there were too many changes from day to day, and they were playing havoc with our production planning.
Priority wasn’t storing history, so I created the beginnings of a data warehouse, and started storing daily snapshots of the customer orders. Using Panorama to visualize the data, within a week we knew that the automatic data was deeply flawed, and orders were changing drastically from day to day – so that yesterday you could have an order for 300 blades of a certain part no. to be supplied by August, and today there would be just 100, but due in June.
More importantly, we could now prove it. Not long after that, the VP Sales travelled to visit our customer in France, laden with printed charts from Panorama, so he could show them the data. Incredibly, they were just as astounded by the data as we were. Apparently they had an automated planning system that sent the orders through the API without any control, and it was causing this whole mess. In time the problem was fixed, but more importantly – the data warehouse and BI had proved its worth.
And if I could give you just one takeaway from these stories: if you’re selling BI software, always enable your customers to use it and analyze their own data.
BI Stories 