This is a phrase I use frequently when I am trying to find a stubborn or elusive problem. I am not always sure everyone understands the statement. The short definition of the phrase is to keep chipping away on the periphery issues to get to the root (dragon) of the issue. It also means sometimes you cannot solve the problem in one step and you have to plan your way to resolution.
I will explore both options in this post.
The graphic below shows the dragon in the midst of the OODA loop. This is by design. The approach to basic and pragmatic trouble shooting is to follow the loop. 
Getting to the dragon by process of elimination
This approach is used when you are not entirely sure of what or where the dragon is. An analogy would be of a dragon slayer going on a quest to find the dragon. The dragon slayer would follow the trail and the evidence till the dragon was found.You need to really focus on the OODA loop here. The first step is to clearly observe all the elements not just one the ones you are comfortable with. Look at it from different angles and from different perspectives. In the orientation phase of your observation you are likely seeing just a symptom of the deeper problem. If you could see the root problem (dragon) you would take a different path to solution. For this discussion we are going to state you are just seeing the symptom.
Once you have oriented yourself to the observation you need to list the options (decisions) you have. Play it out in your head what happens with this test or with this action. Visualize the process. This will give you what you expect to learn or gain from the exercise. This is where I see most trouble shooters fail. They tend to hack away at things and in doing so miss the relevant. They also, end up wasting time and more importantly they tend to add to the problem.
The approach of elimination is just that. Break down the issues in systematic and measured steps. You should have an expected result from the effort. Sometimes the effort is to just prove what it is not. This is OK, sometimes you need to validate the known so you can find the unknown. As Spock use to say in Star Trek, “if you eliminate the possible all that is left is the impossible”. There is a lot of wisdom in that. All too often we jump on the obvious, only to find out it did not solve the problem and we have learned little from the exercise. By taking a step back before we jump in, we can see if jumping in is even worthwhile.
Once you have played out the possible decisions you need to act on them. I am amazed at how many people freeze here. They have a solid basis for their approach, it is based on solid observation and alignment (orientation) with the decisions and yet they do not act. My favourite clue for this, becomes the accountability blame game. It manifests itself in statements like “I do not have authority to do this or I sent an email and no one responded”. This shows lack of ownership and commitment. If you do not act then how do you expect to solve the problem. Hiding behind email shows lack of personal accountability and leadership.
Once you act on your decision then you need to start all over again. What did your action do to change the situation at all? Even a nothing change is telling you something. If nothing happens, you may have to undo your action. If it did in fact change something what did it change. The exercise of “what is it telling me?” is critical in the orientation step of the repeatable loop. Force yourself to dig deeper and in doing so you will be given different decisions to draw from. Once a list of decisions is made act on your plan. Once the action is done, start all over again. Are you sensing a theme yet?
Eventually, even if it is just shear will power, you will find the dragon and reduced it to its smallest elements. The result is the dragon is gone. Now like all dragon slayers you have to go and find a new one to battle. How you defeated this dragon now becomes knowledge you can use for your next encounter.
Slaying the Dragon in steps
When you know what the dragon is and where the dragon can be found you may need to take several steps to get to it. If so, then it becomes all about the planning and execution of the plan. It is the deeper, more mature, application of the OODA loop. When you have found the root cause and you know it cannot be slain in one cut then you need to plan and prepare for its exorcism. To be fair, of late, this is lacking in so many firms and people in our industry. We have become obsessed with the tools and the analytics but little work is actually done solving the problem. We have better diagnostic tools then we ever have had and yet we still are falling behind in solving issues. This is not unique to the information technology field, all one has to do is look at our health care and you will see we are more focused on the problem then we are on the solution.In building your plan you need to break it down into its appropriate steps. Once each step is defined you need to mentally validate that the steps are in the correct order and what is the expected outcome of each step.
Doing so is critical to the OODA loop execution.
So once you have made your plan, you observe the environment. This can be done by doing a baseline test and measurement. You then orient yourself to the observations. Has this baseline highlighted anything or shown a pattern you were unaware of? Note these observations for future reference. Make a decision as to the plan execution. Execute (Act) on the first step of the plan.
Once the plan is in motion, observe what is happening. Orient yourself to your expected outcomes. Are the expected outcomes being realized? If not, what have you found out? Do you need to change the plan? Once you have made up your mind then act on your decisions and the next step of the plan. Start the loop again at the observe step.
Doing so will quickly, and hopefully, predictably slay the dragon. The key is that you must be adaptable to what you are seeing and you must be able to orient yourself to your observations. Your decisions need to be based on what you see and what the context is of what you saw. Lastly, without action you become the victim of inertia and in problem solving inertia or inaction is the enemy.
