What The C-Suite Needs to Know About Data Science – From a CEO And Data Scientist

The C-Suite is deeply invested in Machine Learning to drive revenue and maintain the business’s ability to compete effectively. That means involvement in the process at the right time and asking the right questions.

Vin Vashishta | Originally Published: October 10th, 2015

Book time with me for Career Coaching or sign up for my Business Strategy for Data Scientists class.

You need to be involved. While 75% of companies are planning to invest in big data over the next two years, the number that will see the kinds of returns they are expecting is much lower. There are several factors that lead to success with analytics, but without you none of those really matter.

Many executives get involved with analytics projects once they have overrun their budgets or failed to meet ROI goals. I have been called in to fix several derailed projects and we all say the same thing. “I wish we’d have gotten involved sooner.”

You do not need to run it and you do not have time to run it. The right roles for the C-Suite are strategy, oversight, and sponsorship. The wrong role is active, day to day management. I see a number senior executives, most often CMO’s, CIO’s and CTO’s, get dragged into running data science initiatives. That is a full-time job in and of itself, not one that a senior executive can focus on with everything else they are tasked with doing.

What the C-Suite needs to know revolves around building a smart data strategy, ensuring the goals are met and evangelizing a data driven culture. Let us start with motivation: why?

The Biggest Question – Why?

Three key groups’ expectations, customers, employees and investors, drive data science adoption in business. Customers, both B2C and B2B, expect personalized, consistent experiences and responsive service across a number of channels. Those expectations are met with advanced analytics and data driven automation.

Employee expectations are rising, especially in highly sought-after skill sets. Talent analytics shows the business how to attract, engage and retain their talent. This keeps productivity on the rise while keeping the cost of talent stable.

Investors have come to expect a well-defined data strategy from executive leadership. Their opinions on company valuation, competitive standing and future performance are impacted by the business’s ability to develop and execute that strategy. Data science is not an optional capability from the investor’s standpoint.

Building A Realistic Data Strategy

Data Science Realism 101:

  • You need a plan for responsibly collecting and securing data
  • The business needs a plan for building the capabilities and infrastructure to turn that data into insights to drive positive business outcomes
  • The plan needs to have measurable ROI that stakeholders agree on up front
  • You need a campaign to teach the talent what to expect and why the migration to data driven will improve the business
  • Everyone needs to be accountable for executing on the data strategy

  • Only collect data that will not damage the company’s image and credibility when it is revealed that data is being collected. Only collect what the business can secure because the cost of a breach is massive. Data governance is the term used to describe elements of data strategy like:

  • Data Quality – Making sure the data collected is accurate and useful in data science initiatives
  • Data Security – Keeping data safe from both internal and external loss
  • Data Compliance – Handling the complex legal and ethical implications of data science
  • Data Management – Where to keep it, how to transmit it and how to process it

  • Data is like any other capital resource. Businesses exist to turn capital resources into revenue streams and data is no exception. The capabilities to handle that transformation fall into two categories: infrastructure and talent. Infrastructure is where the data is physically stored and processed as well as the software tools needed to store, process, and consume the insights generated from the transformation.

    Talent are the people who add value to data. Build this around a leadership team, experienced in understanding business needs and how to meet those needs with data science solutions. The talent war in data science is real. While there are roughly 150,000 to 200,000 people in the world claiming to be data scientists, the top tier of talent is less than 20,000 people worldwide. One of these “unicorns” is worth the cost and effort to hire because they can produce systems on their own with high impacts to business performance. Give your HR and talent acquisition teams the flexibility and tools they need to attract and retain top talent as well as outsourcing where it makes sense.

    Overseeing Data Science Without Being Dragged into Managing It

    The difference between being pulled into the day to day operations and keeping a responsible level of oversight comes down to translating data science activities into KPIs the C-Suite can track. The jargon surrounding data science is thick. An easy way to cut through the buzzwords is with a clear focus on ROI and business goals. Data science organizations should be revenue generators and not cost centers. After the initial investment to ramp up, data science organizations should be like any other, a contributor to the bottom line.

    Many data scientists will argue that analytics organizations are different and need to be run or measured differently. In my experience, this is an argument to skirt oversight at best and an outright lie at worst. The rationales against oversight come down to a lack of trust between the data science organization and leadership. Executive involvement from day 1 is a key to remove this barrier.

    Building trust allows senior executive to delegate the daily operations of what will be one of the business’s most important organizations while maintaining a responsible level of oversight. The layer of translation between business goals and data science methods is also critical. Stakeholders need to be onboard with what is expected of the data science organization as well as how success will be defined and reported.

    Evangelizing Data Driven Business

    Accountability needs to extend to other organizations as well, the consumers of analytics. Data science teams cannot operate in a silo. To connect the team with business outcomes they must be integrated into each organization they support. That requires a level of accountability for turning the insights data science provides into outcomes the business expects. The data science team cannot do that alone and other business units will not act on insights they do not trust.

    Again, we come back to establishing trust which is a key function of the C-Suite in a data driven business. Accountability must be a two-way street for a data driven culture to take hold. The data science team is accountable for accurate, actionable insights. The consuming organization is accountable for making their business needs clear and acting on the insights they are handed. That does not work in companies where the teams do not build trust by working together closely.

    That working relationship starts with executive sponsors. Clear expectations and clear goals lead to success with data science as it does with any other organization or initiative. Making data driven a priority your leadership team is vocal about gets the organization on board with the changes needed to make that a reality.