09 May 2018

Dear Data Analytics companies. Seems like you've met your match

I came across ThoughtSpot on LinkedIn. While I felt it was silly of them to request users to sign up to see their video on how they add value, it was easy enough to find it on YouTube. If you are a Data Analytics company that's into Business Intelligence, do watch this. It'll give you a run for your money!


I've helped build a high volume cloud computing program for a company that needed to crunch Gigabytes of data at high speed. So I'm quite surprised that ThoughtSpot uses a relational database for data. Relational databases end up requiring quite a rigid schema, as compared to graph databases.
While the video does not explain how they do the data cleaning and segregation...which requires a huge amount of manual labour and no state-of-the-art AI can automatically do as of today...the video does introduce us to a very thoughtful search interface that allows the user to just type what they are searching for, and custom dashboards get created dynamically. That's where the money is!

As of today, data analytics companies spend months at a time, just to get the data organized and dashboards up and running. That incurs a very high opportunity cost for the client. Add to that an obvious high purchase cost for the analytics software, due to the highly skilled knowledge-workers that build it. Then comes the maintenance costs over the years for every little change and customization request.

The value-add that ThoughtSpot provides here, is just like what Google had provided to the world, when Yahoo Search was at it's primitive worst of requiring users to click through categories of data to find a web-page they were looking for.
It's not just clever, it's a huge money-saver and time-saver in the long-run for companies wanting to analyze their business data...which is what any company that is serious about business should be doing, pronto.

The only part I'm not yet convinced about is the amount of time that's lost in data cleaning and segregation. That's not something that can scale so easily. As for the UI and search capability ideas... brilliant!!!

No comments: