Datafication – why it is important?

In the projects we undertake in the area of big data and analytics one of the tasks we perform is to understand the key performance indicators (KPI). When we inspect how a KPI came into being then we find out that the data used to calculate KPI might a very important component missing in it. When we do a “5 Whys” analysis of the KPIs we uncover that the way data collected and about what the data is collected has some problems. We work with clients to come up with new KPIs or revise their KPIs with new data which they never imagined.

We work with textile clients in India. They have few KPIs related to an order they are working on. They were selling their apparels in all the channels they can and their online sales via popular portal like Amazon, Flipkart and sometimes customers would return the products saying the color was different than what they showed in the website. When we digged into this problem we found out that they are storing all colors used in their company in names that only people in their company can understand. This caused a problem for them when the outsourced their dyeing activity to 3rd party The color’s name and sample was given to the dyeing company and then they would see the sample in a different lighting and mix colors in wrong proportion and final output was not expected.

When the listed the colors available for their apparel in online site they take the picture in different lighting and it appears brighter than usual and then when users order it they find that it is different and return it. The same problem happened in the catalog business where the color on the printed paper appeared different from the color of the apparel. We suggested them to use digital color codes i.e. hex colors and then add hue and saturation to it. Now they are first digitizing the colors in their computer and only after that they share it with their network for dyeing, listing and sales. The challenge they had was that they did not know how to datafi their colors. In much complex projects we face datafication challenges for many things that are so tightly weaved together. 

This is an interesting case study from Japan. Professor Shigeomi Koshimizu in Advanced Institute of Industrial Technology in Tokyo studied how people sit and how their weight gets distributed when the site. He worked to come up with a system that will capture the data necessary to identify each person uniquely. His team was able to come up with lot of metrics and identify pressure points when people sit. They built a model that was fairly able to identify people uniquely using these data and now they are working on use this as anti-theft device in cars. This is a very interesting project that I came across in 2011 when I started my work on big data and analytics.

Hope this post puts you in the pedestal to think about the KPIs used in your organization and how the data used for it may need to be under a new microscope of big data and analytics to check for its relevance and accuracy. Please share your ideas on how datafication will help your company.