Panic makes people poke at random. A method makes you calm and fast. This is the loop experienced techs run without thinking — learn it and you can troubleshoot things you've never encountered, because the process doesn't care what the problem is.
1. Define the problem precisely
"It's broken" isn't a problem statement. Get specific: what exactly fails, what's the error, when did it start, and what changed around then? The single most useful question in all of IT is "what changed?" — an update, a new device, a config edit. Most problems are a recent change wearing a disguise.
2. Reproduce it
If you can make it fail on demand, you can test whether it's fixed. Confirm the scope while you're at it: is it one user or everyone? One device or all of them? One app or the whole machine? That answer alone often points straight at the cause.
3. Isolate the layer
Work the stack from the bottom up and cut the problem in half at each step:
- Physical — is it plugged in, powered, connected? (Yes, really. Check first.)
- Network — can it reach the gateway, an external IP, resolve DNS?
- System — is the service running, the disk full, a process hung?
- Application — logs, config, permissions, dependencies.
- User — is it the account, the input, or a misunderstanding?
Each layer you clear removes half the suspects. That's how you get to the cause fast instead of guessing.
4. Form a hypothesis and test it
Based on what you've isolated, make a specific guess — "the DNS server is unreachable" — and test that exact thing. If you're right, you've found it. If you're wrong, you've eliminated it and learned something. Either way you move forward, never in circles.
5. Fix it, then prove it
Apply the fix and then reproduce the original failure to confirm it's gone. "It seems fine now" is how tickets get reopened. Make it fail the old way — if it won't, you're done.
6. Document what happened
One line — symptom, cause, fix — saves the next person (often future you) an hour. Over time these notes become the runbook your team wishes it had, and they turn one-off firefights into repeatable procedures.
Get the flowcharts that walk you through it
The IT Pro's Troubleshooting Flowchart Pack turns this method into printable decision trees for the problems you actually face — no network, slow machine, login failures, and more. Follow the arrows, find the fix. Instant download.
Get the flowchart pack →