Abhishek Kumar - 10 months ago

Data Science is a buzz word everyone in technology industry want to hear and get associated with. This has created multi-million industries in itself by means of training institutes, start-up and special teams within organizations. In this article i will try to touch upon the aspect of its evolution and economical aspects of this growing service industry vertical.

The influx has been so much that it is becoming a norm to create or restructure a department within firm as Data Science/Analytics. This has lot of cost and talent issues associated with it. Scarcity of talent and sudden rush (without leadership in Data Science) has created lot of job opportunities without structure and well-defined roles. The same did happened in past during dot com bubble and recently in case of mobile app development.  I must say after talking to so many data science professionals and leadership from different corners of the world i find we are doing a injustice by not looking at the fundamental law of economics. And the i started using a word to explain my point of view as:

Return on Data investments (RoD) is a measure which can help organization set and monitor goals around the investment they are doing on Data related functions.

The data needs of an organization can be broadly divided into two sets:

  1. Mandatory Requirements: Organisations need to capture and store data for normal operations, reporting, budgeting etc. And this is nothing new!! organization has been storing and doing basic planning for long and that too is Data Science.
  2. Analytical Requirements:  With phenomenal advance in hardware and network technologies it has become possible to collect, store and distribute data beyond mandatory requirements. And this is the data which feeds all the Data Science/Analytics industry.

Another thing to understand is the timing of all this. From mid 90s data collection started in large scale and with early 2000s advancement in networking technology made this data available at click of a button. Now when market opened, and monopolies started breaking, the natural next area of improving profitability shifted from top line (revenue) to reducing operational cost. And voila!! Data Science got undivided attention. Now it’s time to get some return on data we were collecting for all those years by improving our operations and reduce cost of business.

Now, let’s tie this back to RoD, now the firms are suddenly spending a lot of money in the data resources and talent pool. We need to track that and get a comparable measure to keep our economics sound. Core business fundamental cannot be changed by a new application of data, example, a company selling pizza will still be selling pizza and selling pizza is what result in money not a fancy statistical model saying who will like what (that to with some probability not certainty). And hence i say, we need to look at what all these new investments adding to your revenue, how much to spend and where to stop. 

By ignoring the economical aspects we are creating a lot of new challenges for ourself. Sometime we are even defying the basic rule of business that profit is when your cost is lower than selling price. And somehow we tend to ignore the huge cost we are putting into data science still failing to measure it on cost vis-a-vis revenue. Some immediate issue that i see here are:

There is a huge industry created for training new talent into data science and analytics. The flux is so much that every other person defining and using the term Data Science to their convenience.

The rush in industry to have a data science and analytic team has lot of cost effects. At the same time, our traditional industry is not clear how to leverage this new talent and data infrastructure.  No one want to be left out!! but left out from what? no clarity!

Our academia is confused and in most cases the thought process is driven by "Paralysis by Analysis". Data Science is not a degree course!! Its not like mechanical engineering which you can teach at college and call yourself a mechanical engineering. Data Science is science around data, and data is everywhere. What academia is doing is re-packaging Mathematics courses that they have been running for long time as Data Science.

The new talent getting into this field by following the "BuZZ" is going to struggle in long term career. Being a multi-disciplinary and dynamic field its far more than learning a tool or doing some visualization. The short-term course and doing a quick fix to learn some statistics will not help.

On top of all this we data scientist are challenging the age old structure of doing business by new ways, and believe me its not different from automating a large portion of business and hence a resistance from business to prove RoD. You remember what havoc computers created when introduced in industry in 80s?

All said we are heading towards very intelligent systems and making human life easy so that human can invest time in better sustainable thing for future.

So, what are next steps?

Probyto publish market reports and conduct surveys to understanding changing landscape of data science. We can work to understand your culture and provide our suggestions for becoming better data driven organisation.

Contact Us