#15 Kick-starting Organizational Transformation
Subscribe to get the latest
on 2020-08-21 00:00:00 +0000
with Darren W Pulsipher, Sarah Kalicin,
Creating successful data-driven results starts with a strong organizational foundation. Darren and his guest Sarah Kalicin, Lead Data Scientist Data Center Group Intel, discuss the key aspects to this fundamental change.
Keywords
#datamanagement #analytics #people #policy
Building a Data-Driven Organization
Success based on data starts with an organizational foundation. This means that management has a key role in driving a valuable outcome. Rather than simply recognizing the need for data, asking for insight, and expecting results, management must provide a path for success, beginning with a fundamental question: What is the business value that we want to get from the data?
Once management determines the business questions, resources must be available to support the process: putting the right people in place, training, data collection, preparation, insight creation, and operationalization. This takes sufficient resources and time; the organization must support it from a foundational and cultural level, with a full plan in place.
Obtaining Value from AI
In an October 2019 MIT Sloan Management Review article, “Winning with AI,” the authors showed that organizations defined with a “pioneer” persona got the most out of their AI initiatives. The reason is because they were highly focused on their business strategy and made sure that the data they were using influenced their entire business model. On a basic level, they were using the data to figure out how to maximize revenue while minimizing operating expenses. They were generating value from AI revenue rather than cost savings alone.
The authors also found that these organizations are more successful when C-level executives, rather than IT, drive the AI initiatives. The C-level executives are closer to the business model and the context of how the data is being used. This structure helps avoid the problem of the analytics becoming merely an academic exercise.
Identifying Questions to Create Business Value
What questions should organizations ask to create business value? A good place to start is asking those in the business unit what they worry about and where they lack insight. After brainstorming through those issues, identify the big impact, low complexity problems. Then, figure out what data you have, or can acquire, that can answer these questions. Getting the data you need is not easy and requires discipline. This is where management support and commitment through the process comes in.
Committing to a Strong Organizational Foundation
A strong organizational foundation is not a buy-in, but a commitment by the whole organization to a problem-solving process. Once you’ve defined the problems or the business value that you want, break it down in to workable steps such as finding the data, having the right people in place, and management sponsorship. A problem-solving approach where everyone agrees on the breakdown and process rather than just trying to figure out an answer is essential. There also has to be a commitment to the necessary resources and time.
Feedback and checking throughout the process is important. The team and management must understand that this is not a linear process, but a continuous improvement practice. It may turn out, for example, that the most convenient data is perhaps not the right data. You may have to find a different source or clean up existing data in a way that is usable.
Another part of the organizational foundation is having the right software and hardware infrastructure. Big data needs a sophisticated pipeline. Management needs to understand that they are going to be investing money in the technology to process the data in a useful way. They also need to invest in people and provide training using real analytics software so they can do more with their data.
All of this feeds into the culture of an organization that embraces digital insights and recognizes the value in it.
Defining Roles and Responsibilities
Although some IT roles have been around for a while, it is useful to define the roles and responsibilities for key executives in the analytics phase.
The Chief Analytics Officer (CAO) enables analytics and AI to work to create value for the organization. These are the analytic translators who work with the C-suite executives to figure out how they can leverage analytics and AI through delivery and execution.
The Chief Data Officer (CDO) is responsible for curating the organization’s data so the CAO and their data science team can utilize the data. The data strategy, in addition to curation, is security, maintenance, and quality.
The Chief Information Officer (CIO) secures, builds, and maintains the software and hardware infrastructure to support data, analytic, and AI work. The CIO and their team ensure the data can flow based on the requirements from the data engineers and data scientists.
All of these officers and their teams need to work together. The CAO and data scientists define how the data will be used, building the models and dashboards to provide the insights. The CDO and data engineers curate the data and make sure it is ready for the analytic work, while the CIO and infrastructure teams and solution architects look to the data engineers, analysts, and data scientists to determine what hardware and software can enable their work.
Organizational Alignment Models
With these new C-suite positions, there are several options for organizational alignment along a scale of completely decentralized to fully centralized.
In a 2018 McKinsey report, “Ten Red Flags Signaling Your Analytics Program Will Fail,” it showcases the pros and cons of organizational alignment models. One of the key ideas shows that the benefit of having complete decentralization is that you are putting the expertise right within the business. The data workers will be closely involved with and understand the data, creating high value. Depending on the organization, however, you might not be able to support having so many data professionals in each of the business units. In addition, if there are only a few data professionals, they might not be able to leverage other expertise within the company. In this case, something more centralized could be more beneficial.
Conclusion
Organizations are facing a lot of new changes to become data centric, not only in the culture, but in the organizational structure. It’s not enough to simply want the benefits new AI brings; it requires fundamental changes in the way we think about the organization itself.