7 Problem Solving and Analysis Techniques
In working on my next book, Supporting Enterprise Applications, I am trying to help folks like me whose sole task in life is to troubleshoot and diagnose product problems. I’ve noticed many of the techniques used in supporting software can be applied elsewhere and indeed a lot of what I am including in the book originate from outside the I.T. industry.
Since all Customer Support teams are solving problems, here are some tips that apply to most types, including technical, business, billing, and general customer management.
Frame It Up
You cannot troubleshoot or solve something you don’t fully understand. So break the problem down carefully, ask the what, where, when, how type questions and document the answers to create a complete picture. Consider the scope and extent of the problem and its various effects and impacts. Also consider what is and isn’t affected as this helps identify key relationships, dependencies, and of course validate any assumptions.
Also for problems that are not total failures, consider why this is deemed a problem, since sometimes inappropriate expectations are at the root cause.
Recreate The Scenario
Try to reproduce the problem yourself. Getting a true visual understanding of the issue adds context and subtle nuances that are often very important in understanding the issue fully and identifying its causal factors.
Study The Evidence
Most problems leave a trail of how they came about. To analyze this effectively you need to know what to look for, where to look for it, and the real meaning of what you are looking at. This comes with experience, and by looking at similar problems, but may also originate through good training and detailed reference material. Be patient, very methodical, and thorough.
Process of Elimination
Take time to list out what could be the cause of the problem. Prioritize these based on likelihood, then validate the top ones, collecting either useful details or simply eliminating possibilities. Top Tip: Only get sidetracked during validation if that line of investigation is more likely to yield something meaningful than the next potential cause on your list. If not, note it down and move on.
Divide and Conquer
Define the problem area (or process) into logical sections. Divide this list into two halves. Check one half for evidence of the problem and causes. Discard that half if nothing is found. Continue with the second half by first splitting it in two and choosing which half to investigate. Continue splitting and eliminating until you have pinpointed the precise problem area. This is best suited to long complex processes that result in a vague problem outcomes.
Don’t Jump In
It’s easy to dive right into complex problem analysis and lose sight of the real goal. This is especially true for technical problems, and often a simple replacement component fixes an issue without the need to take everything apart. Similarly check all the basics before going in-depth, make sure the usage and setup of the product or service fits with what is was designed and implemented to do.
A modern mechanic just plugs your car into a computer to diagnose problems – it’s quicker, more thorough, very consistent, and generally more reliable. Well designed products and services come with similar management tools that allow for checking key information to help spot and prevent problems. Don’t forget to use these to save you time and effort.