A man, we’ll call him Bob, walks down the street and approaches another man with a dog standing on the sidewalk. Bob asks, “Does your dog bite?” The man responds, “No, he doesn’t.” As Bob reaches his hand out to pet the dog, he’s promptly bitten. “I thought you said your dog doesn’t bite?!” shouts Bob. “He doesn’t,” the man replies. “But that’s not my dog.”
The obvious lesson is that while Bob might be an expert in dogs, able to identify breed, weight, age and lineage at a glance, his failure to ask the right question came back to bite him. Literally. The same can be true for an organization and its data. We may have a deep knowledge of our data architecture, lines of code, or master data files, but it’s the things we don’t know to ask of the data that can keep us “stuck” when it comes to digital transformation. Before getting to the right questions to ask, let’s examine why data is so important to digital transformation, and look at some common challenges organizations face in acquiring and consuming data.
In our 2019 Client Global Insights, more than 1,550 client executives we interviewed indicate that data and predictive analytics continue to be a top investment area for innovation. This makes sense, given that data insight helps us make better decisions, and most digital technologies rely on and create an astronomical amount of data, such as from the Internet of Things (IoT) or artificial intelligence. But, why is using data effectively often a struggle? Many times, the root can be traced back to analytics programs that get stuck in a cycle of build-analyze, when they really need to reimagine the whole process—flipping it on its head to become an analyze-build development model.
Chasing your tail: the challenges of data acquisition
Developing a transformative data strategy is a series of stepping stones, with data acquisition being the first. This step is a frequent stumbling point, and where my team often first engages with a client. Many organizations have built their data infrastructure over years, incorporating disparate systems that each look at data in a slightly different way. Reengineering (or unifying these data sources and making the data more usable) is no small feat, and chasing after efficiency using a limited set of structured data is a common scenario.
The problem occurs when organizations focus solely on reengineering as an end goal for their data strategy. The data in their legacy systems most likely is structured data versus unstructured or qualitative data. As a result, organizations tend to limit the kind of problems the data can solve.
Playing fetch: the constraints of data consumption
Structured data stuck behind inefficient systems has limited value. As a result, organizations can fall into the trap of repeating the same kinds of questions of that data. More specifically, business users develop a rhythm of asking the data team for a limited set of answers. The data team delivers the answers, often presenting the data in more useful ways with modern data visualization tools or accessing the data more efficiently through upgraded systems. They may also integrate data from other parts of the organization to create a more holistic view.
Don’t get me wrong: these are important steps in the data strategy journey that can help steer an organization with valuable insight. But, they are waypoints to true transformation. What’s needed today is more than reengineering one’s data strategy, it takes completely reimagining it.
Teaching your data (and processes) new tricks
The deluge of new data sources created by the digital economy (e.g., mobile device, IoT, social media, etc.) and structures (e.g. data lakes) allows organizations to reimagine completely how they operate. Greenfield organizations – those born in the digital age – may appear to have a leg up when it comes to transformation and digital disruption because they don’t have to re-imagine well-established processes. However, established organizations can still thrive with the right approach to transformation.
Currently, many established organizations rely on a build-analyze model where IT first builds data layers based on known business questions and business users then analyze the data once it is available to them. The build-analyze mode limits the focus to known questions. To accelerate transformation, organizations need the capability to analyze all internal and external data to uncover new questions and patterns, and then build advanced analytics and processes accordingly. This change in approach – the analyze-build model – is what we refer to as the modern data life cycle.
Of course, every organization is different in their transformation journey. Each will run into different sticking points along the way. Some may be data-related while others may have an entirely different set of challenges. In our new CGI viewpoint¸ we look at some of the root causes of these transformation challenges, and provide real-world solutions for how to overcome them.