Data-driven organizations don't happen by accident.
The age of big data has opened possibilities for a data-driven organization like no other time in history. Organizations that have paved the way and tools that make powerful analysis easy make the move to a data-driven culture much simpler than it was just a decade ago. However, it still takes a focused effort to design the right organization if you hope to use your future analytic capability to drive a competitive advantage. Let's look at the organizational design elements of a data-driven company that's poised to outpace the competition.
SEE: 10 books to get you started on big data (TechRepublic)
It all starts with the corporate strategy, which is why it's important for the CEO to own the transformation--he or she is the only one who's accountable for your company's vision and purpose. Regardless of the products and services your company offers to its customers, the CEO must make it very clear that the future of the company will be powered by data.
It's also important for the CEO to establish why the company needs to shift to a data-driven culture. One reason is universal to everyone in today's global competitive environment (i.e., without some degree of analytic capability, there's no way to be competitive), but there must be other compelling reasons why it's imperative for your organization to be more analytic. The CEO must be crystal clear about what these reasons are before structuring the organization for success.
The ideal structure for a data-driven organization has a Chief Analytics Officer (CAO) or equivalent reporting to the CEO--this sends a strong message that big data analytics plays a strategic role in the organization. Remember that the CEO must retain ownership of big data analytics at the company, so the CAO is not a position for the CEO to delegate all accountability; instead, the CAO should be the chief advisor to the CEO and an integral member of the strategic team.
SEE: Job description: Chief data officer (Tech Pro Research)
Under the CAO should be a robust organization of data science. An appropriate top-level split for a data science organization is qualitative and quantitative functions. Qualitative data science is where you house your capability for exploratory data analysis, while quantitative data science is where you build strength in testing hypotheses. This is the most effective way to empower the business units.
Good business processes are what brings great analytic capabilities to the business units. In addition to analytics, there will be several functions (Finance, HR, etc.) that sit under the CEO, and there will also be business units. Since the business units reside in a different area than the analytics function, you must build cross-functional (i.e., lateral) processes to bring the power of analytics where they can provide the most benefit.
My advice is to eschew the notion of a matrix and build good global processes with process owners. Matrix organizations are confusing and violate unity of command. Instead, develop good global processes that start with the business units and flow through qualitative and quantitative analysis only to return back to the business units with great analytic value.
To make sure your global processes are effective, you must measure and reward performance--this starts at the organizational level with a scorecard. Attached to every corporate strategy should be a scorecard that clearly measures success; this is where the value of having analytic talent comes in handy. Whether you currently have a data science function or not, you can employ your current analytic talent to help build a robust scorecard.
Measurement and rewards must extend beyond the top level--every level of the organization should be measured and rewarded for moving the organization to a data-driven culture. Take this all the way down to the individual level, where workers are acknowledged for good analytic behaviors, like analyzing a problem before acting.
SEE: Digital transformation as a data-centric service (ZDNet)
Never forget that any organization is a collection of people that must be selected, guided, and nurtured. The last piece of the organizational design is arguably the most critical: making sure the right people are doing the right things.
The critical first step is selection. Once you have an idea of what your new analytic organization will look like, you must make sure it's staffed with the right people. For a new data science organization, you'll need data scientists who are multi-disciplined and flexible. For the rest of the organization, you'll need people who at least have an affinity for data.
SEE: Hiring kit: Data architect (Tech Pro Research)
Continuous learning and development also plays a critical role in the data-driven organization. Given the pace at which new technology and analytic techniques are discovered, there must be an enablement function in your organization that's constantly keeping the organization at the top of its game.
If your goal is to have a data-driven organization, you cannot take its design lightly. An important first step is understanding exactly where you want to go and why. Once that's established, structure the organization with a prominent data science function and build global processes that bring powerful analytics to the business units. Build scorecards and reward systems that extend from the top of the organization all the way down to the individual contributors and make sure you select and nurture the people that power your organization.
This classic five-step outline for designing an organization is your roadmap for building a data-driven organization that can outperform your competition. Take some time today to get your strategy straight. Then design your way to a more competitive future.
- How to build a data-driven culture with logic (TechRepublic)
- How to build a data-driven culture with credibility (TechRepublic)
- How to build a data-driven culture with emotion (TechRepublic)
- Chief data officer: Insight into a crucial role for the exabyte age (TechRepublic)
- How to build a successful data scientist career (free PDF) (TechRepublic)