The Dichotomy Between Data Science and Business Analytics

dichotomy between data science and business analytics

As an analytics professional, I take a keen interest in trends of all sorts. Thanks to Google Trends, we don’t have to question the accuracy of this trend – in my business, any other trend can and will be questioned six ways from Sunday!

I know that IT investments are no popularity contests, but I think search term traffic is a good proxy for what interests businesses worldwide at a given point in time. In this blog, my objective is to try and highlight my understanding of what might be driving these trends. I’m stating what is obvious from this chart.

  1. The interest in “Business Intelligence” has significantly waned in the last decade
  2. The post-recession interest in “Data Science” has grown exponentially (and surpassed BI just a year ago!), while
  3. “Business Analytics” has kept a decade-long marathon-paced trend in popularity

dichotomy between data science analytics

I started my career in analytics exactly 11 years ago. I think it coincided nicely with the emergence of analytics as a legitimate business function. And I worked at Hewlett Packard that could afford investments in high-performance software and research.

Business intelligence was very popular, and everyone was crazy about ETLs and data warehousing and “putting 2 and 2 together”. It was an exciting time to learn new stuff.

Read more: Visualizing retail trends during Black Friday and Cyber Monday with Tableau

Business analytics – an emerging field where science meets art – had distinctly strong believers and non-believers. Senior leadership remained skeptical at the time while younger managers saw it as a tool to further reinforce and justify their decisions.

Because of the now-clichéd “quick wins”, the early believers saw quick success, which in turn nudged the senior leadership to understand this field better. I don’t recollect the term “data science” being as popular back then as Google says it was, but the flat trend-less period from 2006-2011 further evidences that.

The fact that we see three distinct trends in popularity here implies that they have very different drivers. And I don’t think the analytics community uses any of these terms interchangeably, as they shouldn’t be. Having done both for over a decade, I find myself compelled to spell out the dichotomy between data science and business analytics. It is naïve to think of one without the other. However, I think there are key nuances at play here.

Business Analytics is a Super-set of Data Science

I like to think of data science as a cog in the wheel that traverses the path laid by business analytics towards monetizing data assets. Business analytics leverages domain and industry knowledge to provide the right data – a natural input for data science. It navigates organizational challenges in adopting and using data science. It has to repeatedly defend data science against nay-sayers and convince the leadership of its value within an organization.

Data science is conditional on data availability; business analytics is not. At the end of a successful demand forecasting project I concluded recently for a major hi-tech company, the SVP of supply chain asked me what accuracy the models we built were expected to yield. And I promptly answered – 66%. Six months out – we were obviously prepared with our simulations and stuff. He then said something unusual but insightful – “What data do you need to get it to 75%?”

Read more: 10 Must-read Books on Data Science, Analytics, and Big Data

What rankles decision makers is that there is always more data “somewhere” that you did not consider in data science. Sponsors of analytics expect a rational business answer to such questions, and that is what keeps them hooked to the data science journey.

Business Analytics Rationalizes Spend on Data Science

Data science comes at a steep cost – not just the expensive data scientists but the hardware, software, time, effort, and mindshare it requires. And the output of all that investment is always just incremental to what the organization would have achieved without it! This makes the check-book owner within an organization sometimes question his or her motives before going to bed.

I think business analytics helps businesses see the long-term value in data science. It is ostensibly simpler to establish the value of data science in making operational decisions while business analytics has to juggle a lot of balls while evidencing this value in making tactical and strategic decisions (a topic very close to my heart).

Business Analytics Interprets Data Science

You cannot explain multi-collinearity and variable importance to a business manager – not until we have data scientists leading business functions (we will get there, eventually!). A key role of business analytics is to break data science down to consumable chunks while a program is in motion.

Read more: 12 Situations Data Scientists Will Totally Relate to

Ninety-five percent of the marginal return on investment in analytics comes from a 5% chance that you will find something opportunistic and counter-intuitive (at least non-intuitive) to the business function’s knowledge from data science. If you don’t do a good job at defending that precious insight and value from every possible angle of attack (and you will face hell while doing this), it’s gone – relegated to a few PowerPoint pages, Excel sheets, and R models that will never see a double-click again.

So, where is the aforementioned trend headed? Here is what I think.

  • BI will become axiomatic, an absolute necessity for any organization and function that works for profit. That’s the real reason why BI’s popularity is trending down.
  • Business analytics will continue its upward trend – IoT, machine learning, and artificial intelligence will target the automation of operational decision-making. The biggest bang-for-buck business analytics in strategic and tactical decision-making is safe and secure in its marathon pace. And that is the sole reason why companies will invest in data science R&D.
  • Data science will see sustained popularity till the time machine learning and artificial intelligence take over operational decision-making. Beyond that point, I expect both data science and business analytics to amalgamate.

I’d love to hear what the community of analytics practitioners thinks on this subject. Please do leave a comment.

(This write-up first appeared on LinkedIn Pulse)


Think you’re aware of all the interesting application areas of artificial intelligence? Find out by checking out the video below!