Data Scientist and CIO
Data science has been characterized as one of the hottest disciplines in the 21st century. It is almost impossible to pick up a newspaper or magazine without reading an article about big data and/or analytics. This is certainly true of technology publications but, articles can also be found in most business magazines and newspapers as well. The data scientist and especially the chief data scientist is the person who is supposed to pull the right data together (both internal and external), extract amazing new insights and provide their company with competitive advantage that allows them to leapfrog their competition and become the next Progressive, Capital One or Harrah’s. These are some very lofty expectations!
While the future for data science is certainly bright, there are warning signs ahead.
Having been in a variety of information technology (IT) roles for the past 30 years, the current hype around data science is amazingly similar to that seen for IT back in the 1980’s and 1990’s. When American Airlines rolled out the Sabre reservation system and Merrill Lynch created the Cash Management Account, when the internet bubble began. Information technology was hot and companies scrambled to find people that would help them find competitive advantage.
Can companies learn from their experience with IT and not make the same mistakes as they build their data science organizations?
Over the succeeding 20+ years the IT profession saw tremendous growth, successes as well as failures, and experienced the ups and downs of inflated expectations and disillusionment. While some companies gained competitive advantage many came to question their IT investment and its growing share of the expense load. The inability of IT leaders to explain the IT ”black box”, to ensure that the IT investment delivered value and that their organizations were as tightly run as any in the company led many companies to reduce their investment and even outsource their entire IT organization. Ultimately, IT was increasingly seen as not delivering competitive advantage and that opinion was perhaps best captured in Nicholas Carr’s book Does IT really Matter? (N. Carr). While there are many potential reasons for this, we are only going to look at a few areas.
Transparency is necessary to demystify the value in data. With transparency comes trust and confidence. Data science, just like IT, is technical in nature and can be seen as a “black box” to those outside the field. Data scientists need to be able to distill technical work and ideas into easily understood business terms that communicate what the effort will add to an organization’s profitability, products/services, and overall efficiency.
Companies need to make sure they are not seduced by just building the “best model”. It was easy to get caught up in the building of technology for technology’s sake and it is just as easy to build an analytical model with the latest and greatest statistical advances… yet not deliver something that aligns with the business goals and objectives. Just like IT, data science must deliver solutions that add value. When large investments are being made companies need to have a clear understanding of what the investment is, what it will do and how it will be measured. Stakeholders want to see a positive RoI and data scientists need to demonstrate their success in measurable ways.
Data science and more
Multiple items need to come together to create a great solution. Data science (analytics), data, business understanding, real world constraints and non-traditional disciplines like psychology all need to be taken into account to deliver the best solutions. It is key to build a data science capability that combines a deep understanding of their business with what quantitative methods and data analysis provides. Successful data scientists, just like CIOs, will bring together multiple disciplines and experiences as well as their technical discipline (i.e. data science and IT).
Building a good model, one that will deliver value is necessary but not sufficient on its own. Models are of limited value unless they can be implemented and utilized. As with IT projects, putting the right structure in place for data science initiatives is a must and the old adage about being on time and on budget applies to data science as well. From gathering requirements to seamlessly executing the plan within the limits of a clearly defined scope – project controls and communication are a critical component for any project to be a value add.
As with IT 25+ years ago, the future outlook for data science is very bright. There is no question that there are tremendous opportunities to deliver value to companies using data science. As the hype of inflated expectations cools, those companies who are able to capitalize on the experiences of IT will achieve success by:
- building a data science capability that combines a deep understanding of their business
- understanding what quantitative methods and data can provide
- focusing on solutions that add value
- making sure that the project and change management processes are in place to ensure successful implementation
Successfully harnessing all of this does not ensure companies will become the next Progressive, Google, or Amazon… but they will be better off than their peers who aren’t.
About The Author
QuaEra Insights, LLC
Co-Founder & COO