Our Data-Driven World- What drives success with data science?
3 steps to successfully integrate data-driven decisions and solutions into any organization.
Idea in brief
There is a lot of talk these days about data science, machine learning, artificial intelligence, and the need to use data in decision-making in a variety of situations. We use data to make decisions related to our businesses, community efforts, government policies, shopping for the best deal, personal and family health and many other day-to-day experiences.
In most organizations, leaders believe in the need for data to make more informed decisions. The challenge is — how do we get the most out of the data to generate positive results? Success in data science is about more than just the data itself. It is about taking action — execution- based on the data, engaging multi-disciplinary teams in this process, and embedding a data-driven philosophy into the culture of the organization.
What we know about data and algorithms
Of course, we do need to start with the data itself. It is critical to make sure the data is in good shape; it’s clean, unbiased, and that we have enough of it to build solid algorithms with strong predictive capabilities.
We also need to accept- otherwise reject any denial- that algorithms perform better than humans, even experts. While this statement may surprise you, there is a lot of research confirming this.
Daniel Kahneman, world renowned psychologist, economist and recipient of the Nobel Prize in Economic Sciences, is an expert in judgment and decision making. In this interview, author Daniel Pink interviews Kahneman and discusses algorithms and human (professional) judgement and how unreliable human judgment really is. It is a bit shocking, but listen for two minutes and let it sink in.
Interview with Daniel Kahneman
Starting at 4:30 in the interview, Kahneman talks about biases and noise in the data and how
“In general, the rules beat the experts.”
He shares meta research that tells us that the algorithms and formulas simply perform better and can help us to reduce noise or variability in decision making. So, if the algorithms beat the experts, does that mean we don’t need humans or experts anymore?
Humans are still needed
While the reality is that there are and will be some employment impacts due to utilizing AI solutions, particularly for some of the more repetitive jobs, humans and experts are still needed. In the end, we have to realize that our focus needs to be not only on the data and algorithms, but also on people, process and culture.
People are still needed to run the processes, to identify the business questions and hypotheses, to plan and build the models, as well as to develop, implement and monitor actions taken as a result of the analyses. I am confident that
Data or machine learning alone will not drive the transformation of our organizations and communities. Data and ML models are “enablers” of decision-making and transformation, they are not the transformation itself.
Demet Tatar, PhD in Material Science and CEO/Founder of Crimson Data Analytics, builds multi-disciplinary teams of experts to create data science solutions for clients in different industries. She says
“Data science allows us to build very powerful tools to help us understand complex problems, make data-driven informed decisions and design creative solutions. We should be thinking in terms of data science being complementary to the human decision-making process. Experts from both academia and industry are still extremely valuable in the implementation of these solutions.”
I agree completely with Demet on this. Instead of shying away from the data and algorithms, we need to embrace them as critically important-- and treat them as complementary to our expert judgment and intuition. By fully leveraging the data, we will be able to make more targeted, better decisions and have greater bottom line success.
As Harvard Business School Professor Karim Lakhani confirmed in a recent Harvard Business Analytics Program (#HBAP) discussion on data science in healthcare:
“Leaders who use algorithms will outperform those who do not. Managers who use algorithms will outperform those who do not. Doctors who use algorithms will outperform those who do not.”
And you can plug in any role or field and keep going…
A simple equation to visualize data science success
A simple equation can help to visualize the relationship between data science, humans and culture within organizations: DS Success = D + Ex + P + C
(DS=Data Science; D= Data (cleansed, unbiased, strong predictive value); Ex=Execution; P=People; C=Culture)
To reap the benefits of what data science has to offer, there has to be focus on how the data is operationalized. We need to have people in place to ensure the insights are translated into actions within the organization’s strategies and day-to-day operations and, even more broadly, that the practice of data-driven decision-making is woven into the culture or DNA of the organization.
Idea in Action- 3 steps to successful data analytics integration
Operationalizing the data will not just happen organically, without a more purposeful and disciplined approach in the organization. I have learned, through my time leading Customer Experience teams and efforts, how to best utilize data strategies within the organization and drive data-driven solutions out to customers.
Here are 3 steps that can help to ensure you translate the data into actions and integrate data science into the culture of any organization:
- Build execution muscle and accountability Build a team (doesn’t have to be big!), or expand a team already in place, who will be responsible for leading the development of solutions and execution of actions. This is how I built the team when I was Head of Customer Experience (CX) with AT&T Mexico. In addition to our Voice of Customer and CX Design teams, we had a team of data scientists building and running models as well as an Execution team responsible for building cross-functional team efforts to develop tangible actions, identify owners and involved teams, and ensure the actions are implemented and monitored continuously.
- Engage employees Develop and implement a process to regularly and consistently share key data and insights to leadership and across the entire organization. Work with leaders across all functional areas, and especially the HR team, to make sure to engage key stakeholders and all employees in the process. Education about the data and what success looks like is an important part of integrating data-driven practices into the culture. It doesn’t happen overnight. It takes time, meticulous planning and a strong communication process to increase adoption of this discipline within the organization. Continuously work to engage employees in this process because without a cross-functional approach, the implemented actions and related results will not be sustainable.
- Communicate actions and successes Share success stories with external customers and key stakeholder groups (including your own employees). “We have learned x from the data and have translated this into better solutions for you.” Communications and closed feedback loops with customers, regardless of organization or industry, will help drive increased loyalty and bottom-line results.
Building execution muscle, engaging your employees and communicating actions and successes, within and outside of your organization, will help to build and sustain positive results coming from your data science efforts.
Remember the other portion of the equation — execution, people and culture — is equally important as the data itself. DS Success = D + Ex + P + C.
How are you integrating data and data-driven strategies into your organizations? Would love to hear your experiences. If you are just getting started, and/or want to discuss further, please send me a message or email me Michelle@leadwithcx.com.