What We Actually Test in a Phone Review
We may earn a commission if you buy through our links. It does not change what we recommend or what you pay. How this works.
Anyone can restate a spec sheet. This is the procedure behind every phone review on this site, published so you can judge whether our testing resembles how you would use the phone.
The short version
- Fourteen days as the only phone. No parallel device.
- Screen-on time logged every day, not estimated.
- The same five camera scenes on every phone we compare.
- A twenty-minute load test, because sustained performance is what you feel.
Setup
- SIM moved across, so calls, data and messages are real.
- The same twenty applications installed on every review unit.
- Brightness on automatic, because that is how phones are used.
- No performance mode, no battery saver, unless we are specifically testing those.
- Three days of settling before measurement starts — initial indexing distorts battery readings.
Battery
Screen-on time recorded at the end of every day, with a note on what that day involved. We publish the average, the worst day and the best day, because an average alone hides the variance that actually matters.
We do not run video-loop battery tests. They produce a tidy number that does not correspond to anything a person does.
Performance
A benchmark on a cool phone establishes a tier. Then a twenty-minute sustained load, with frame rate or score recorded at minute one and minute twenty, plus surface temperature. The gap between those two numbers is the review’s performance story.
Display
- Peak brightness measured, not quoted.
- Outdoor readability assessed in direct sunlight.
- Colour accuracy in the default profile, because most people never change it.
- Flicker behaviour at low brightness, which matters to a minority a great deal.
Camera
Five fixed scenes, shot on every phone in a comparison, in the same conditions on the same day:
- Daylight landscape with high dynamic range.
- Indoor scene under mixed lighting.
- Low light without flash.
- A portrait with hair and a complex edge.
- A moving subject.
Full-resolution crops are published so you can judge rather than take our word for it.
Software
The dimension that affects daily life most and appears on no spec sheet:
- How many preinstalled applications, and how many can be removed.
- Whether system applications display advertising, and whether it can be disabled.
- Notification behaviour out of the box.
- The stated update policy, and the manufacturer’s record of meeting it.
- How many settings you must change before the phone behaves reasonably.
What we cannot test
Being straightforward about this matters as much as the procedure:
- Long-term durability. Fourteen days says nothing about year three.
- Battery degradation over time.
- Network performance beyond our own location and carrier.
- Every regional software variant.
Where we revisit a phone after six months, the review is updated with a dated long-term note.
Common questions
Why fourteen days?
Shorter and you are still in the honeymoon period. Most of the irritations in our reviews surfaced between day four and day ten.
Do you keep the phones?
Retail units we buy are kept or sold. Supplied units are returned or disclosed.
Why not more benchmarks?
Because benchmark scores correlate weakly with daily experience, and sustained performance tells you more.
Sumit
Writes everything on TechMozz.