It’s early January 2022, and as I sit here writing this, I’m struck by the number of stories of government organizations of all sizes who have turned to advanced analytics to empower their mission outcomes. Agencies as diverse in mission — from the U.S. Census to the State of Florida, the DOD and LA County — all made significant advancements in the way they gathered data, analyzed it, and automated their progress towards insights.
While diverse in mission, they’re unified in their approach to leveraging democratized analytics. In turn, they’ve realized significant progress on their analytics journey by leveraging an analytics platform that enables their data teams to upskill their analytic throughput without requiring access to specialized tools or hold advanced degrees in data science or mathematics.
The resulting efficiency realized by these teams is certainly time saved by their data workers, but the real value is realized in how service delivery is improved for the people, community, and employees these organizations serve. For example:
The U.S. Census is better able to deliver more complete, more timely insight on key U.S. Economic Indicators
The State of Florida was able to quickly reestablish the delivery of unemployment services
LA County was able to automate data and analytics to improve insights into public health
While these stories are clear indications of the value of a democratized approach to analytics, there is still lot of work to do within many government agencies. Consider the following:
44% of data workers time is wasted on searching, preparing, and analyzing data.
87% of Pub Sec leaders cite democratization of data and technology as critical to organizational success.
97% of organizations are failing in their analytics initiatives because of people and process issues.
As a result of these and other factors, it is estimated that across all industries, annually 500,000 days of a manager’s time (at a typical Fortune 500 organization), is wasted on ineffective decision-making.
The International Institute of Analytics (IIA) maps out the analytics maturity of an organization across a five-stage model that starts with stage one: the analytics beginner with no data-driven capabilities and culminates with an organization being a true, analytics and data-driven enterprise. According to the IAA, it turns out that most enterprises are below the midpoint and only average a 2.2.
This means that most organizations have some folks that can execute some level of “pretty analytics” but even at many “analytics-focused” enterprises, most knowledge workers do not have capabilities beyond basic spreadsheet manipulations. This is the situation that is repeated time and time again at many public sector organizations where analysts supporting key business functions, financial operations, human capital, public health, logistics, and many other missions are stuck working with 30-year-old desktop technology, working with pivot tables and relying on cut-and-paste capabilities.
Yes, there are pockets of more advanced analytics in many organizations, but many times that is the domain of highly trained specialists who have coding skills and access to specialized tools and expensive licenses. The other challenge is that there simply aren’t enough of these resources to go around. The good news is that pretty much all public sector organizations recognize that they want to move up the curve and are investing in analytics and data science. Certainly, most organizations have projects within their plans to build data lakes, implement updated data governance, and embrace automation efforts all in the name of digital transformation. In fact, 99% of organizations claim to be making these or similar investments. But what we see and what has been reported is that while organizations are having some success, many feel like it is going too slow … and their data science and dedicated analytics teams feel like they are tied-up with low-level, manual data janitorial services.
The Analytics Divide
What is seen in most organizations is that they have a small number of highly analytic individuals, Data Scientists, advanced developers in IT, and even data engineers who can perform analytics — and then there’s everyone else with business-related questions that need answers every day.
For example, take critical functions in any branch of the DOD. There are domain experts (data workers) that work with data every day in finance, budgeting, contracting, human resources, logistics, force readiness, and mission planning who need to leverage that data to create insight that will inform and guide decision making. Yet to get the answers they need, these data workers must rely on data scientists, data engineers, or others with higher-level analytic skills and capabilities to process the data and analysis to get the answers.
Even when this “throwing it over the wall” approach works perfectly and the question that was asked is answered many times the answers themselves lead to additional questions about the data, the analysis and why the results don’t support the assumptions. So even after all the investment in in data science and analytics, in most enterprises there is a significant and growing “analytics divide” that creates its own set of issues.
To address this, some organizations see the fix as training data workers to become more like data scientists. Teach them Python, pay for the additional access to the specialized tools and platforms, and the problem is solved. However real-world experience has informed us that this does not work.
This type of upskilling is complicated and the return on the investment is limited. It’s like training an aircraft maintenance person to become a pilot simply because they work with planes. This overlooks the critical factors of acumen, aptitude, and the years of training it takes to become proficient. For many data workers, the investment in their analytics technology and training has been limited to a 30-year-old spreadsheet, and if they were to become a data scientist you would have to invest hours in study and training before they could ever become effective.
Democratized Analytics: A Balanced Approach for the US Navy and others
What is more effective is a balanced approach where more self-service analytics capabilities are deployed through the enablement of data workers with right tools that support varying levels of technical acumen and can be shared between the various levels of data talent. This is what we define as “democratizing analytics.”
In the end, organizations have a choice of how much effort, resources, and dollars they are going to invest in democratizing analytics vs. tools that only specialized resources will leverage. While your data scientists will likely work on some of the most complex and important challenges, and hopefully empower the insights your organization needs, due to the limited number of data scientists you have the level of insights won’t be near enough to enable your organization to become fully data driven.
Meanwhile, the more numerous data workers across organizations who work on line-of-business projects can indeed generate more numerous insights that creates impact across lines of business with a democratized approach towards analytics.
As organizations like the US Navy face a growing complexity of threats from more advanced adversaries, climate change, and more infused technology within ships and systems, there is an understanding on the importance of data. To have a more robust readiness position means depending on a greater use of data analytics to anticipate maintenance and modernization requirements to keep the fleet ready to respond. For organizations like the US Navy and others, the ability to better leverage data will be built on the capability to upskill the domain experts in the use of analytics. These domain experts — better than anyone — know the questions to ask and the data that is needed to find the answers. In other words, with a democratized approach to analytics, functional domain experts will know where the gold is buried, will be able to self-serve their immediate business-related data needs, and create stronger relationships with data science resources to make it easier to implement solutions.
Want to learn more?
Then read Automating Analytics: A Human-Centered Approach to Transformative Business Outcomes.