The most expensive thing about a misbehaving device is rarely the repair. It is the two days spent fixing the wrong layer: backing up, reinstalling, restoring, reconfiguring — and then watching the same freeze happen on a clean system, because the problem was a failing drive the whole time.
The takeaway up front: answer hardware or software first, and everything after it gets cheaper. It is a diagnosis you can usually make in twenty minutes with tools already on the device. Sites that organise their guides around the symptom rather than the error code — FixTech publishes Android, Windows, AI and security walkthroughs that open with the problem as you would describe it — are useful here precisely because you rarely start out knowing which layer you are in.
Why the distinction is the whole diagnosis
Software faults are reversible and cheap. Corrupted updates, driver conflicts, background processes, full storage, malware — all fixable without opening anything or spending money.
Hardware faults are progressive and get worse. A drive with reallocated sectors, a swollen battery, a fan clogged past the point of cooling, a charging port with worn contacts. These do not improve, and every hour spent reinstalling an operating system over them is an hour the underlying failure continues.
The two produce overlapping symptoms, which is why guessing fails. A laptop that freezes could have a failing SSD or a bad graphics driver. A phone that dies at 30% could have a worn battery or a rogue background app. Same complaint, opposite responses.
Check 1: Does it fail in a clean environment?
This is the highest-value test, because it isolates the layer directly.
On Windows: boot into Safe Mode. Safe Mode loads a minimal driver set and no third-party startup items. If the freezing, crashing or slowness disappears in Safe Mode, you are almost certainly looking at software — a driver, a startup program, or something malicious. If it persists in Safe Mode, suspicion shifts firmly to hardware.
On Android: boot into Safe Mode (usually by long-pressing Power, then long-pressing "Power off"). This disables every app you installed. If the phone behaves, an app is responsible and you can bisect by uninstalling recent installs. If it still overheats, reboots or lags, the problem is below the app layer.
This single check resolves a large share of cases and costs one reboot.
Check 2: Does the device report its own health?
Modern hardware self-monitors. Most people never look.
Storage. Drives track their own degradation through SMART attributes — reallocated sectors, pending sectors, wear levelling. A drive reporting reallocated sectors is failing, regardless of whether it currently works. This is the single most common cause of "my computer freezes randomly" on machines more than a few years old, and it is the one where continuing to use the device costs you data.
Battery. Both major desktop platforms and both major mobile platforms expose a battery health or design-capacity figure. A battery at 70% of design capacity is not broken, but it explains a device that dies early, throttles under load, or shuts down at double-digit charge. Our guide to why phone batteries drain fast covers the software-side causes worth ruling out first.
Thermals. A device that is hot and slow is usually thermally throttling, and thermal throttling is a physical problem: dust, dried thermal paste, a failed fan, or a blocked vent. Slowness that arrives after ten minutes of use and clears after a rest is a temperature story, not a software one.
Check 3: Is the failure consistent or contextual?
Hardware faults tend to be tied to physical conditions. Software faults tend to be tied to actions.
Ask what the failure correlates with:
- Correlates with movement, angle, pressure, temperature, or a specific cable → hardware. A screen that flickers when the lid moves, a phone that only charges at one angle, a laptop that crashes when warm.
- Correlates with an app, a file type, a website, or a time of day → software. A crash that only happens in one program, or slowness that appears after every startup and settles later.
- Started immediately after an update or an install → software, until proven otherwise. This is the cheapest lead to follow and the most commonly ignored.
- Started gradually over months with no triggering event → hardware wear or accumulated storage pressure.
A charging port illustrates the pattern well: intermittent charging that depends on cable angle is mechanical wear, and it is progressive — it never gets better, and catching it early is the difference between a port repair and a dead device.
Check 4: Swap the cheapest variable
Before concluding anything, change one external component at a time — a different cable, a different charger, a different port, a different network.
This sounds trivial and it resolves an embarrassing share of "device failures". Cables fail internally with no visible damage. Chargers degrade. A USB-C cable rated for power but not data will charge a phone and refuse to transfer a file, which looks exactly like a broken port.
Change one thing, retest, change it back. The discipline matters more than the order: swapping three things at once tells you nothing about which one mattered.
What each answer means for what you do next
If it is software, you have a ladder to climb, cheapest first: uninstall the recent change, update or roll back drivers, clear storage to leave real headroom, scan for malware, reset settings, and only then consider a full reinstall. Most people start at the top of that ladder and skip everything cheaper.
If it is hardware, the question becomes economic rather than technical: what does the component cost against the value of the device? A battery or a drive is usually worth replacing. A logic board usually is not. Our guide to repairing or replacing a laptop walks through where that line typically falls.
The habit worth keeping
Run the four checks in order — clean boot, self-reported health, correlation, cheapest swap — and write down what you observe. Twenty minutes of diagnosis routinely saves a wasted day of reinstalling, and occasionally saves the data on a drive that was quietly telling you it was dying.
The reset is not the fix. It is what people do instead of finding one.
