You Don’t Need a Data Scientist

Headlines tout messages like Data is the New Oil  [1]” and tell us that a “Data Scientist [is] the Sexiest Job of the 21st Century [2].” These messages are trumpeted from the rooftops and have convinced executives that not only do they need data, and lots of it, they need someone called a data scientist to figure out what it all means. These data scientists are lauded as wizards, able to magically coax insights, knowledge, and wisdom from data.  Recently I lauded the role of the analyst in analytics, but I worry that combined in the cacophonous call for data scientists, the message may be misinterpreted or tainted. Let me pull back the curtain on the data wizards of Oz and explain what I mean when I say you don’t need a data scientist.

Before I go any further, let me be clear: there is an absolute need and role for experts in data and business analysis to help you get the most out of your data. Without expert advice, assistance, and support, you will not get the benefits you would otherwise realize. But most companies should think twice about either investing too heavily on building internal specialization or relying solely on a cadre of data scientists to provide all analysis and insights. The path that led to this mindset was a long one, which is why it will be so difficult to displace.

Twenty years ago, the world of data analytics had very high barriers to entry. Computing power and storage were at a premium and to use either required detailed and in-depth knowledge of coding. The language R, released in 1993, provided one of the popular frameworks for statistical analysis and is still in use today. Scala and Python offer alternatives, but in each case, at least a fair amount of coding knowledge is required to be able to use it.

More recently, the world was introduced to Hadoop, Spark, Hive, and a litany of other novel concepts, platforms, and products. The power of these dwarf earlier tools, making analysis of terabytes and petabytes possible. Largely, however; to use these tools, one still had to have very high levels of technical proficiency. Hence, the need for data scientists.

Gradually, the C-suite has become inured to the fact that to gain any benefit from data, a team of data scientists must be employed, and to really gain benefit, brought in-house. I have attended conferences with sessions dedicated to “bringing your analytics in-house” and “why you don’t want to outsource,” but these only exacerbate what I suggest is the problem.

Insights can only be gained, according to this proposition, from the great and glorious data scientist. It suggests that these experts must be part of the staff, and that only these oracles of wisdom can dispense data driven insight to the business. This is wrong-headed, misguided, and just plain lazy.

Insights and analytics are no longer the domain of the few. With tools like IBM’s Watson Analytics, it is the time of “citizen analysts.” [3] Many tools no longer required extensive coding knowledge to perform analysis. Insights are delivered in an understandable and actionable format. Data exploration is possible for the many, and questions never before thought of can be asked and answered—almost by anyone.

Maybe I was wrong in the title of my post. It probably should be called “You don’t need a data scientist—you need every employee to be a data scientist.” Successful organizations and businesses will begin an era of democratized data and insights; a model where any can ask questions and get answers will become the norm. Frontline employees to C-level executives will have access to tools to find better ways of doing business through their data. There is still an important role for specialists and experts to perform complex computations, develop algorithms, and implement machine learning code within an organization. My suggestion is simply that all employees be empowered through tools to gain the insights to move forward. Innovation and insight can come from anywhere in a company, and all should have the tools to help them disrupt the sometimes arcane existing models.

Implementing such a starkly different philosophy is complicated. A partner already expert in tools like Watson Analytics and with traditional big data and analytics tools can help provide the guidance, training, and on-going support absolutely vital to your success. Half-hearted or ill conceived attempts will result in very costly failures and resentment among employees. Archetype SC has the knowledge, resources, and experience to provide the support you need for success in your complicated data journey, from early stage conception to on-going support. Archetype SC: we do complicated.

I would love to hear your thoughts about my ideas presented here; do you think I’m right? Wrong? Do you have examples of how a democratized data culture has worked for you or how it has hasn’t? I welcome civil discourse and will engage thoughtful responses. You can reach me at

[1] To illustrate how ubiquitous this adage has become, each word in the quote links to a different article with it as at least part of the headline. Sources range from highly credible to more marginal on purpose. Inclusion of these links in no way implies that I agree with the content; purposefully I have included some that does not meet my standard of journalism excellence.  To learn more about the origins of the phrase, I recommend reading:

[2] I’m pretty sure my wife was terrified when the Harvard Business Review first published this headline in 2012. She had no idea that her mild-mannered, nerd of a husband had magically been transformed into a 21st century sex symbol. Fortunately for my marriage, it is still every bit as uncool, and unsexy, to be a math nerd.

[3] IBM has begun using this as one of their marketing phrases; I first heard it during an IBM Watson event, but I posit it will become a widely used and accepted phrase, adopted by other players in the analytics space.

© Copyright 2019 Archetype SC, Inc.