The term “data driven” has become widely used in recent years with lots said and written about becoming a data-driven organization. I think it is important to step back a bit and realize that not everything is suited to be data driven. Data-driven works well when the entire data-to-value chain can be implemented with flow of data and processing by computers. But many of the things that we do in business can’t be entirely handed off to technology. They need human participation and human judgment. These are the things that I consider to be data informed instead of data driven.

The Data-To-Value Chain

The ultimate purpose of data analysis is to create value – to earn more, spend less, reduce risk, retain good customers, acquire new customers, recruit and retain skilled employees, improve quality, adapt to change, and so on. There are many ways to create value in business. To make the analytics connection, we need to understand and enable the data-to-value chain. (See figure 1.)

Figure 1. The Data-To-Value Chain

The data-to-value chain, viewed in context of analytics and data science, describes a sequence of processes that begins with data creation and progresses through multiple stages that lead to actions, outcomes, and value creation. 

Data is the starting point. It all begins with data—the digital collection of descriptive, qualitative, and quantitative facts that are the raw materials used to create digital information.

Information is produced when data is processed and analyzed. We use data as basic units or building blocks, grouping multiple units together. Creating information is a process of connecting related data elements and putting them in context that has greater meaning than individual facts.

Knowledge is created through the  flow of information. Both digital and experiential information are received, absorbed, understood, and retained by the human mind. Synergistic knowledge is particularly powerful—new information that relates to and readily connects with knowledge already held. Both existing knowledge and newly created knowledge may be individual or organizational. Individual knowledge is held in relation to the experiences, recall, and beliefs of a single person. Organizational or social knowledge is held in relation to the consensus beliefs and cultural norms of a group of people. 

Decisions flow from knowledge. More precisely, conscious and informed decisions come from knowledge. The most basic decisions are whether to act or to not act on information received and knowledge held. The questions are: Should I do something? And if so, what should I do?

Actions are shaped and driven by decisions. Actions are simply the behavior of doing something. Start doing something not previously done. Stop doing something that has been routine. Change the way that an existing activity or process is done. Continue doing something without making any changes. Start, stop, continue, or change – at a macro level, those are the actions that we take to produce outcomes.

Outcomes are the results of actions – the changes that are produced in people, processes, products, services, behaviors, relationships, etc. 

Value is directly connected to outcomes. Ideally, our actions produce positive outcomes that increase value, and we are able to avoid negative outcomes that inhibit or erode value. Positive outcomes may optimize processes, mitigate risks, innovate processes or products, etc.

The Data-Driven Value Chain

The driven value chain begins with data as already described – descriptive, qualitative, and quantitative facts that are processed and analyzed to create information. (See figure 2.) The information tends to be primarily the results of predictive and prescriptive analysis that is understood and retained by artificial intelligence and machine learning models. 

The models are the decision makers.
Everything up to the point of action is computerized.

The models are the decision makers—producing recommended or prescriptive actions that do things such as approve or deny applications and activate or terminate services. Everything up to the point of action is computerized. Action is the responsibility of people when implementing recommended actions, and the responsibility of technology when automating prescriptive actions. 

Figure 2. The Data-Driven Value Chain

The Data-Informed Value Chain

The data-informed variation of the data-to-value chain also begins with data. Typical information includes trends, patterns, diagnoses, and predictions that flow to decision makers to expand and complement their knowledge. In the data-informed chain this is the point where the activities shift from computerized to human. In this variation, data and information are computerized, but knowledge, decisions, actions, and outcomes depend on people. (See Figure 3.)

Experiential knowledge of decision makers—recall, experiences, and beliefs—are as central to decision processes as knowledge created from data.

Experiential knowledge of decision makers—recall, experiences, and beliefs—are as central to decision processes as is the knowledge created from data. It takes both digital and experiential knowledge to provide insight, guide judgment, and align with decision-maker beliefs and organizational values as part of the decision-making process.

Figure 3. The Data-Informed Value Chain

Visually, the differences between data-informed and data-driven are subtle, but they are important. The roles of people, and the variation in knowledge creation and decision making directly affect both the upside and the downside of each approach. 

Comparing the Approaches

Each approach has both an upside and a downside. The tables in figure 4 summarize some of the pros and cons of each.

Figure 4. Data-Informed vs. Data-Driven

There are implications for speed of decision making, biases and influences in decisions, explainable decisions, disruptions, sustainability, and more. It makes sense to be both data-driven and data-informed with the choice dependent on the use case. The value of data-driven, automated online insurance quotes is obvious. Dependency on people would slow the process and potentially introduce errors and inconsistency of decisions. The risks of automating hiring, promotion, and compensation decisions are equally apparent. Impersonal and potentially biased decisions are sure to bring complaints and resistance, and to take a toll on employee morale.

The important point here is that data-informed or data-driven isn’t a one-size-fits-all concept. The right choice varies depending on the use case. Ultimately, every organization needs to have a mix of data-informed and data-driven.

Making It Actionable

Data to value, data-informed, data-driven … It sounds so compelling, yet many organizations struggle to get from analytics to impact. Simply having data and processing it for analytics is not a guarantee that value will be created. Getting to value requires a few good practices that make data analytics actionable:

• Connect analytics with people
• Connect analytics with business processes
• Connect analytics with decision making
• Frame business problems
• Identify business requirements
• Translate requirements to questions
• Use the right data
• Turn data into answers
• Turn answers into decisions
• Turn decisions into actions

These practices hold true regardless of which value chain variation is used. I have only introduced them as a list of ten practices here. Perhaps I’ll dig deeper in a future blog.