Three Reasons Why Enterprises Fail at Data Science

By July 18, 2017, 6:35 pm EDT
Share Tweet LinkedIn
atidiv sailesh shot copy2

During Chicago’s Growth Summit, we brought together a veritable cast of experts to discuss data science (aka data intelligence), one of the buzziest of business buzzwords this year. Panelists included:

  • Sailesh Mohapatra (moderator): MD, Atidiv –  a “full stack” data science solutions provider, helping clients across a number of industries understand the meaning and value of their data and turn it into actionable insights.
  • George Souri: CEO & CO-Founder, LQD Business Finance– a tech enabled commercial lender, using intelligent decision systems to make the  underwriting process faster and more accurate.
  • Sean Brown: CEO, YCharts – a SAAS company, taking data and providing the tools and analytics to help investors make smarter investment decisions.
  • Stu Frankel: CEO & Co-Founder, Narrative Science – a company spun out of research done at Northwestern. They’ve created an Advanced Natural Language development platform called Quill, which takes data, figures out what’s important, and then creates language to explain it’s findings- everything from verbal conversation to 10 page reports. Key clients include financial services firms and the intelligence community.
  • Dan Wagner: CEO & Founder, Civis Analytics – a data science consulting operation for internal teams, and SAAS platform with an elastically scalable machine learning back end.

Topics ranged from Machine Learning and AI to predictive learning, natural language processing and more. Sailesh of Atidiv notes that in “the last five years the sector has heated up so much that [many people] don’t even understand what data science means anymore.” George Souri points out that “data science” isn’t new – before personal computers, companies like IBM were using chalkboards and punchcards to quantify their business processes.

In a particularly illuminating moment, Dan Wagner pointed out the three biggest barriers to implementing data science at large enterprises:

  • 1. Personnel – companies need both technical personnel, who build technical assets and run predictive workflows, as well as translators, people who translate those insights into business processes that make a difference to the bottom line.
  • 2. Technology – many businesses are going through a “pit of despair,” making investments in technology without a clear strategy, including creating “data lakes” or hiring a couple of people internally who aren’t equipped with the right software for success.
  • 3. Culture – a lot of enterprises face resistance to data science from within. There are “a set of people with entrenched ideas of how to operate, then the nerds come in saying ‘this is how you should do things,’ and when the nerds fail at their first attempt, a process of re-entrenchment begins.”

Dan then suggests three ways to break that cycle:

  • 1. Hiring an aggressive combination of new people that have a different way of looking at data, and use different workflows
  • 2. Investing in new technology to accelerate the adoption of data science into a company
  • 3. Conducting a cultural transformation by leveling up staff who support data science, and building “translator” capabilities

Watch the full panel here to learn more, including why “99.9% of finance people don’t understand data”!





Share Tweet LinkedIn

Techweek Newsletter Signup