Survey Finds Barriers That Prevent Organizations From Realizing Data Science Return on Investment

Print Friendly, PDF & Email

Domino Data Lab, provider of the advanced data science platform, announced the findings of a survey of  more than 250 data science leaders and practitioners — Survey Finds Barriers That Prevent Organizations From Realizing Data Science Return on Investment. The survey revealed factors that enable data science teams to achieve return on investment (ROI), top capabilities contributing to their success, and priorities they must address this  year. Ultimately, model-driven organizations — those who’ve mastered the technical and management practices necessary to embed algorithmic-driven decision making at the core of the business — are reaping the most returns.

The survey findings highlight a divide between organizations that view data science as a technical practice as opposed  to an organizational capability woven throughout the business fabric,” said Nick Elprin, co-founder and CEO at  Domino. “Those who treat data science as a core business capability are better equipped to build and deploy models  quickly, manage and monitor them in production, and calculate models’ impact on the business.”

The survey, which ran from November 2017 through January 2018, surfaced four fundamental
findings:

  • Collaboration is a foundational factor driving success. Collaboration was cited as the primary factor contributing to success with data science; 72 percent of organizations considered “model-driven” (based on number of models in production, ability to control them and quantify their impact) named collaboration as the main attribute of success, as opposed to 63 percent of organizations not considered model-driven.
  • Key barriers prevent organizations from achieving success. Most organizations (90 percent of respondents) see data science driving innovation in their business, but only nine percent can quantify the business impact of all their data science projects. Furthermore, only 30 percent of companies can claim to have more than five models in production. The challenges preventing organizations from becoming model-driven can be categorized into four barriers: silos of knowledge, iteration friction, static infrastructure, and model liability.
  • Model-driven companies are increasing their investment at a staggering rate and they threaten to widen the  competitive gap. Three distinct groups of organizations appear in survey data: model-driven (14 percent),  aspiring (40 percent), and laggards (46 percent). Of the model-driven organizations, 91 percent plan to grow their investment in data science teams by two times or more in the next 12 months, while only 36 percent of laggards plan to double their teams.
  • A consistent and sensible framework around people, process, and technology is required to become model- driven. Model-driven organizations out-perform laggards in each of these areas:

People: Half of model-driven companies’ respondents report spending at least 50 percent of their time on enjoyable work, whereas only 22 percent of laggards can say the same;

Process: 80 percent of model-driven companies get a model from idea to production in less than three months (vs. 60 percent for laggards); and

Technology: 86 percent of model-driven companies have adopted a data science platform (vs. 30 percent of laggards) to enable collaboration, facilitate knowledge sharing, and accelerate research.

Download the Domino Data Lab survey HERE.

 

Sign up for the free insideBIGDATA newsletter.

 

Speak Your Mind

*

Comments

  1. Not surprising it is the human factors that are the biggest issues. I’ll keep banging this drum for those who want to hear it, get some basic UX skills – data scientists! Simple things like running mini workshops, interviewing folks and using paper/white board wireframes make you WAY more effective.