Beginning your journey towards a data-driven mindset comes with challenges but remember these five key points and you’ll avoid a lot of hiccups.
The idea of being a “data-driven” organization has been around for several years. The pursuit of making better decisions armed with facts is a good one.
As part of achieving this goal, a lot of companies have invested in data scientists, data lakes, new open-source storage capabilities, and a whole array of new and changing tools. These are not necessarily bad things or wrong decisions. However, this focus on being data driven sometimes has its own set of challenges.
In this blog, we walk through five key points to consider when pursuing your initiatives to be (more) data driven.
1. Don’t get lost when gathering data.
Initially, becoming a data-driven organization tends to be about identifying and centralizing data. Companies pursue creating data lakes, data ponds, data warehouses or a myriad of other data repositories. They pull all the data found together into one source. This takes a lot of time and energy. While centralizing the data to enable insight isn’t necessarily bad, where do you start? How much data is enough before the business side can see the value?
Organizations often get so involved in the architecture and technology efforts that they lose sight of the intended value. The timeline gets increasingly longer, and the business gets frustrated. Rather than trying to get all data into a centralized source, consider working with the business to identify core needs and related metrics. Identify the key questions you need to ask and the data that would enable the answers. Then you can focus on sets of related data and iterative insights to help drive business answers over time. You’ll find as you uncover these answers the business needs may change.
Seek to understand business priorities and metrics, along with the impact they can have. A business-first mindset will energize your team as solutions drive insight and productivity.
2. Technology shouldn’t drive the business. The business should drive technology.
Over the past several years, I have seen several companies relinquish their pursuit of being data driven to their IT department. They ask IT to produce the best path forward. From a pure technology perspective, many options, tools and potential skills are required. Machine learning and artificial intelligence can drive a lot of potentially interesting insights that excite most technologists.
However, basic business intelligence and reporting can also provide the right insight for many businesses. There is value to almost all solutions, but the value depends on the organization’s need and maturity. So, where do you start? It is important to understand what the business needs to improve its decision making and customer experience. Ask yourself, what skills, data and capabilities are already in place?
Understanding the basics is often critical to making the right decisions with more complex insights. Let the business side define the targeted goals, domains and expected potential impacts, so IT can apply the right technology. Don’t forget to invest in training for both technologists and data consumers.
3. Data is good. Data with context is even better.
I have worked with a lot of data people. Whether described as business intelligence experts or data scientists, they have great insight and understanding of their data. Many also have the ability to produce interesting and (often) complex reports and dashboards. The thing that often differentiates good data people from great ones is the ability to provide context.
The numbers are great, but how do they fit within the expectations of good and bad? The most valuable reports and dashboards explore and answer “Why.” If your reports only help provide awareness around what happened but cannot explore causes or potential actions, are they driving the intended value?
Great data people are those who understand and anticipate the actions of their users. Their report and dashboard design is intuitive and aligns with the business’s context and need to explore and understand the key factors impacting what they should do next.
Seek to understand your business priorities and metrics, along with the impact they can have. A business-first mindset will energize developers and consumers as solutions drive insight and productivity.
4. More is not necessarily better.
In the past 20 years, our ability to access and process data has grown astronomically. We stopped thinking a terabyte of data was enormous a long time ago. This access has helped in various ways. But does your business user actually need all of that data? Do all of those metrics and visuals need to be on the dashboard?
I have talked to many business users and executives about their reports and dashboards, and one thing keeps coming up. The business does not use, need or understand all of the reports. While the visuals may be beautiful and involve complex metrics, are they contributing to better decisions?
Many professionals I talk to describe a lot of the reports as cumbersome and difficult to interpret. While they agree the visuals are very impressive (and may even want to initially show them off), they also admit to downloading the base data to make the decisions they need to make. When questioned why they indicate the data scientists and business intelligence engineers make things too complex. While the metrics may be correct and interesting, they are unimportant to business decisions or strategies. Instead, they take up space and often cause confusion from a consumer perspective.
When business stakeholders ask the data scientists and business intelligence professionals about the complexity of the reports, they indicate that there is a lot to say. Their goal is to make sure the consumers have everything they need to be data driven. They don’t have a shared understanding of what’s important and what they could remove or pull for different consumers. The business and data side need to understand and align on the scope to drive the full value.
Not all data has the same value or impact on your business. There is no need to put all the metrics on reports or dashboards. Work to define priorities and scope. Start simple, and then add more if needed.
5. Small victories lead to bigger success.
Many organizations focus on final outcomes and wait for their data strategist to organize “all” of the data in some sort of repository (think data lake or data warehouse) to acknowledge progress. This means longer delays for the business and IT to realize the value of their efforts. Instead of focusing on specific business areas and targeted outcomes, you can start with related data sets to drive some initial value.
Even something specific like organizing master data and increasing related data quality can improve outcomes. The key is to think about what you need for incremental improvements. When adding data, think about adding integration and relationships. Each added piece and each integration will drive additional benefits if aligned to business needs.
As you recognize value, your entire team becomes more energized, driving better decisions and better data. And data-driven becomes a value-driven mantra across the business.
Conclusion
When improving your business’s data-driven goals, remember to drive with the business and do not let your teams get lost in the data or technology. Small improvements can lead to bigger value, and iterations can show big results in team engagement and customer satisfaction.
Struggling to sell machine learning to your organization? In this on-demand webinar, our data and analytics experts explain how to simplify your approach to machine learning so you can earn the buy-in and funding you need to create the advanced capabilities you desire.
