The DIKDAR Principle
SCIO Decision Intelligence Framework
“Data informs decisions,” we exclaim! Oh yeah? How exactly?
I’ve learned to listen carefully to what’s said right after data informs decisions. Inevitably, I’ve found the conversation drifts back to data and analytics. But what about the decision? You mentioned a decision. What’s the decision look and feel like? And that’s where the conversation typically stops. We know intuitively there’s a decision in the value chain to business results, but it’s elusive. We can’t point to it. We can’t say exactly how it’s achieved.
How good is the organization’s current decision-making? It is reflected in the organization’s decision track record. Unfortunately, most organizations do not track their decisions, actions, and results well. Sadly, many managers would prefer their decisions not to be tracked, as transparency can be uncomfortable. Ultimately, decision quality is reflected in operational performance. This presents a massive risk and opportunity to the organization’s ability to deliver value from its assets.
In the words of W. Edwards Deming, “Your system is perfectly designed to give you the results that you get. We need a system for better decision-making. I’ve learned DIKDAR, a principle behind improving the organization’s decision-making capability.
DIKDAR is an acronym that stands for Data > Information > Knowledge > Decision > Action > Result. It is a progression rather than a business process.
DIKDAR is an adaption from DIDAR in data quality management (DIDAR) that I learned from an operational leader many years ago. The source is surprising and ironic because this leader was famous for making big decisions using his experience and gut feeling, not using evidence or following his advice to apply the DIKDAR progression. This principle has other variants, but the DIKDAR format suits asset management and operational excellence well.
How does DIKDAR work?
Let’s start with the end in mind, the “R” for Results. We all want great business results — it is the one thing we can all agree on. Unfortunately, we don’t agree on what it takes to earn those great results. Deming also tells us, “Every system is perfectly designed to get the result that it does,” which means we get the results we deserve, not necessarily what we want. If we want better results, we need a better system. Our aspiring results are not the problem.
That brings us to “A” for action. Many people in the organization are ready, willing and able to take action — they must be directed toward the best action. Sure, execution can be challenging, but the organization will execute to its best ability. Action is also not our problem.
Let’s go back to the beginning, the first “D” in DIKDAR, Data. We are taught that data is an important asset, the new currency of business. Much of the digital transformations asset-owning organizations undertake are centred on increasing data and quality data management. Our organizations are facing an avalanche of data, some of which is useful and some of which is not. Data is vital, but it is not actionable until context is added. I contend all the perfect data is not the objective, but I’m fighting a losing battle.
When you add technical context to data, you get the “I,” as in Information. Information is obtained with the analysis of data. Digital transformations have also provided organizations with analytics platform tools and services that enable almost any data manipulation. Information is useful to understand what is going on in our complex systems. What’s changing, and by how much, and is it significant? Is the situation normal or abnormal? Is there a problem developing? Technical context is important, but Information is still not directly actionable.
When you add business context to information, you get “K” Knowledge. Knowledge includes what we think we know and our uncertainty around that knowledge. With knowledge you can understand if the situation is a real issue, problem or risk to the operations business plans. Is the issue worth addressing? How do we frame the problem? How do we model solution alternatives? Knowledge is actionable.
There are two sources for knowledge synthesis. First, the intuitive model is based primarily on qualitative experience, subject matter judgement or intuitive gut feel. Second, the evidence-based model is derived from quantitative data evidence. Both sources of knowledge composition are valid. Practitioners and leaders often find it difficult to combine those two sources — they haven’t been trained to deal with disparate sources of knowledge. It can be done — you can combine intuitive knowledge with evidence; or combine evidence knowledge with intuitive.
Actionable knowledge can trigger a “D”, Decision when the value to the organization is threatened or opportunities arise. We should apply sound problem-solving and decision-making practices to arrive at a decision. The potential consequences and the complexity of the problem will determine the appropriate level of rigour to be applied. The analysis may evaluate several solution alternatives against multiple objectives. The decision can be one of satisficing or optimizing.
Ultimately, the decision will be made in only one of those two domains. If the final decision is made in the intuitive model domain, the decision-maker uses their experience and intuition which may be influenced by other’s expert professional judgement or evidence presented. I call this the warm & fuzzies generating good feelings or the queasies generating uneasy feelings. If the final decision is made from the evidence model domain, the decision-maker may integrate intuitive knowledge but will make the call based on what the evaluation is recommending.
How should DIKDAR be applied? Backwards.
Most people want to start with the data and end with the decision, Data<Info<Knowledge<Decision. They focus on the data and information with some notional idea of a decision. This discovery method yields mediocre results. A better way is to work from the end to the beginning. Given that we all want a great Result (R), we will commit the organization to Action (A). So, let's centre ourselves on the decision to be best made. What Knowledge (K) is required to make a quality decision? What Information (I) and analytics are required to feed the required knowledge? Finally, what Data (D) supports the information requirements?
A better progression is Decision>Knowledge>Info>Data. This more deliberate method provides superior results. Some data is more important than other data. The deliberate approach allows the distinction of valuable data to support the decision, while the discovery approach does not.
Don’t be a DIKDAR, be a RADKID
The real challenge individuals and organizations have in quality decision-making is not in the Action and Results, Data and Information; it is distilling the available Knowledge into Decisions. We can trust that once the decision is made, the organization will execute to earn the results we deserve. We will also have the required data and analytics skills to distill into knowledge with our uncertainty. We have more data than we think and don’t need as much data as we think.
The domain we must work in is the Knowledge>Decision, or more precisely, the Decision>Knowledge.
What’s needed is an appropriate decision-making framework to bring those elements together, resulting in more, better decisions. The Decision Intelligence Framework provides the platform and canvas to distill Knowledge (K) into quality Decisions (D), the second D in DIKDAR.
Book a Decision Intelligence Framework demo with SCIO.
