Every repair culture war - did the tweak work, was the part worth it, was the PC even faulty - is settled by the same cheap instrument: the same measurement, taken twice. This page is the capstone of the testing method: how to pick three numbers worth tracking, change one thing at a time against them, and keep the five-line log that professionals use to prove their work and amateurs never have.
The habit feels almost insultingly simple, which is why it is skipped. Before touching anything on a slow PC, you record three numbers: boot time to a usable desktop, the launch time of the app the user lives in, and the peak Task Manager readings during their slow task. Every later claim - about a cleaner, a setting, a new part - must survive comparison with those three numbers, or be withdrawn.
Choose measurements the user actually feels, taken the same way every time:
Write them in the symptom line from your first five minutes. Unwritten before-numbers do not exist, and after-numbers without them are advertising.
The discipline that gives the numbers meaning: one change, then re-measure. Trim the startup list or stop the stuck service or clean the fan - one of them, not all three in an enthusiastic afternoon. With one change, a moved number has exactly one explanation. With five changes, a moved number has five suspects and a still-moved number has no information at all.
This is also the honest answer to just do everything: doing everything destroys the diagnosis and usually buries the one change that mattered under four that did nothing but add risk.
Comparisons are only honest when the conditions match. Time the boot from the same cold state, launch the same file, run the same task, at a similar point in the update cycle - Windows installing updates in the background forges and destroys more before-and-after claims than any tool.
If the number improved, record by how much. If it did not move, the change was not the cause - undo it where practical, restore the suspect list, and test the next one. A change kept after failing its measurement is how PCs accumulate folklore settings for years.
The whole record fits on an index card or one phone note, per machine:
For a working technician this log is quietly the most professional object in the room: it is what you show the customer, it is your own memory when the same PC returns in six months, and it is the difference between we fixed the cause and we did some things and it seems better.
Watch the video take both sets of numbers live - the before and after of the same machine, same tasks, is the whole argument.
Watch Full VideoBecause fairly sure is where the expensive mistakes live. One change keeps the evidence clean: if the number moves, your suspicion graduates to knowledge; if it does not, you learned the truth cheaply. Five changes at once teach you nothing except that afternoons are short.
A phone stopwatch, the same start and finish lines, and the same file or task are plenty. You are comparing minutes and tens of seconds, not timing Olympic sprints. What ruins comparisons is changed conditions - measuring the after-boot while updates install, for example - not stopwatch precision.
Nobody is lying; the feeling is real and the fault was elsewhere. Placebo speed is a genuine effect - a cleaned desk and a watched PC both feel faster. Trust the numbers to locate the fault, and enjoy the feeling as a bonus, not as evidence.