Five Myths and Five Ways to Create an Analytics Culture
This entry is the fifth and final in a series of blog posts discussing issues that the government faces with implementing big data strategies. The first blog post in this series described why data should be defined as "big" based on complexity of the data, not volume alone. The second blog post and third blog post explained four challenges that a big data strategy presents for public sector organizations. The fourth blog post provided examples of areas and organizations from the public sector that are developing and applying big data strategies.
What are the challenges of changing an organizations culture to embrace analytics and adopt a big data strategy? A culture that shares data should be part of that big data strategy. Recall the first blog acknowledged that not every organization will need or necessarily benefit from big data analytic tools. However, it is prudent for leaders to examine business or mission requirements to determine where they could apply analytics and if they have a need for big data analytic tools. I uncovered several cultural myths that leaders and analysts have faced, which I will address first. Then I will discuss some best practices that leaders can apply to change their organizations analytic and big data culture.
Cultural Myths that Challenge the Adoption of a Big Data Strategy
Myth 1: Data is Knowledge and Knowledge is Power
One of the most significant challenges to an organization's cultural approach to big data is that people see data as knowledge, and according to the age old adage that knowledge is power, so people perceive that restricting the use of data is one way to wield power. This view has led to the creation of data silos which inhibit people from sharing data with external organizations and often within organizations.
Myth 2: No One Understands the Data Like Me
Often analysts believe that they are the only ones that understand their data. If the data is dirty and/or it is not well defined, multiple analyses on the same set of data could result in different conclusions. Different analysts may see a problem through different lens, resulting in different assumptions and methodologies for cleansing the data. Presenting a decision maker with different recommendations for the same problem based on the same raw data results in a lack of trust and confidence from the decision maker and extra work for the analysts.
Myth 3: Big Data Technologies and Sharing Data are a Risk to Job Security
Various reports state that analysts spend anywhere from 60-80% of their time preparing data (e.g. organizing, cleansing, filtering, documenting). Given that high proportion of time spent in preparing the data, a number of analysts are concerned that open or shared data will reduce the requirement to prepare data; thus, reducing the number of jobs for analysts. Furthermore, big data technologies reduce the need to organize and clean data, also raising concern about reducing the number of jobs for analysts.
Myth 4: Change is Too Hard and Resource Intensive
Change is hard and it will require resources. Inherently, there is a strong desire to maintain status quo. However, technology, the environment, the competition, and the problem all continue to change. Change is the only path to process improvement and staying ahead. Changing will take time and money, but the benefits of better data and automation may be well worth the investment.
Myth 5: The State of my Data Warehouse is Embarrassing
When you have visitors over to your house, there is always at least one room where you stuff all of the junk, and that room is not on the tour. Similarly, no organization's data warehouse is perfectly organized and governed; there is always room for improvement. Likely, the worse the state of the data, the greater the need for a big data strategy. Being embarrassed and hiding the problem will not help solve the problem. Consider one of those TV shows about hoarders. The only way to fix the problem is to expose the problem and ask for help.
Best Practices in Changing an Organization's Cultural Approach to Analytics and Big Data
Practice 1: Identify an Appropriate Business Requirement
The first step to any problem is defining and understanding the problem. As I have emphasized many times in this blog series, you have to identify an appropriate business or mission requirement that will benefit from advanced analytic tools and/or a big data strategy. TechAmerica Foundation's report on Demystifying Big Data (http://www.techamerica.org/Docs/fileManager.cfm?f=techamerica-bigdatarep...) laid out four great recommendations for getting started.
- Understand the "Art of the Possible" -- Explore the case studies contained in the report linked above, in the fourth blog entry in this series, and otherwise in the public domain to find inspiration and practical examples.
- Identify 2-4 key business or mission requirements that Big Data can address for your agency, and define and develop underpinning use cases that would create value for both the agency and the public.
- Take inventory of your "data assets." Explore the data available both within the agency enterprise and across the government ecosystem within the context of the business requirements and the use cases.
- Assess your current capabilities and architecture against what is required to support your goals, and select the deployment entry point that best fits your Big Data challenge - volume, variety or velocity.
Practice 2: Leadership Establishes Clear Expectations
Leadership has the responsibility to establish clear expectations, which will lead to the change of organizational culture. This responsibility may require some difficult decisions pertaining to personnel actions and resource allocation. Leaders should emphasize that data is not power and get everyone rowing in the same direction for the good of the organization instead of personal gain or notoriety. Subject matter experts say that a carrot and stick approach works well in encouraging change. Rewarding those who share data, practice data governance, use metadata to document, and use the tools that translate into transparency in the data is effective in changing the culture and encourages others to adopt those practices. When rewards do not work, leaders need to make the unpopular decisions consistent with established expectations. Leaders can address the job security risk by repurposing analysts. Reducing the resources required to prepare the data for analysis allows leaders to reapportion analysts to spend more time analyzing the data and looking for insights or new ways of using the data.
Practice 3: Start Small and Build Iteratively
There is an old saying that advises to eat an elephant one bite at a time. Experts say that is the same way to embark on a big data strategy. Change is not too hard and resource intensive with the right approach. Start with a relatively small, yet appropriate business case. As that application proves successful, grow and build iteratively, which is the basis of a DevOps type approach discussed in the third entry of this blog series. Although the marginal cost in dollars might be higher, starting small should require a smaller comprehensive resource investment.
Practice 4: Build a Champion and Celebrate Small Victories
By starting small and growing iteratively, leaders and analysts can build champions along the way who will act as advocates for the new big data strategy, spreading the good word. As this group of champions grows, the naysayers will lose grip on the organization's culture, facilitating the path to change. In her interview on Conversations on Big Data (http://ourpublicservice.org/issues/modernize-management/conversations-on...), Lisa Danzig, Associate Director for Personnel and Performance, Office of Management and Budget, gives great advice on this topic including the recommendation to celebrate small victories. These celebrations become contagious, motivate the team, and encourage others to join the team. Everyone wants to be on a winning team.
Practice 5: Acknowledge that Failure is an Acceptable Way to Learn
An organization with zero tolerance for failure is an organization that is resistant to change. Fresh ideas and new technologies pose too great of a risk in this type of organization. Inherently, people are going to make mistakes. It is what people and organizations do in response to those mistakes which determine overall success or failure. An Army doesn't have to win every battle to win the war. Good leaders want their people to test and try new things; it is the only way to move forward. It is the leader's responsibility to ensure there are risk mitigation measures to reduce the impact if the experiment fails.
***The ideas and opinions presented in this paper are those of the author and do not represent an official statement by IBM, the U.S. Department of Defense, U.S. Army, or other government entity.***