Rating methodology
How reviews and measurements sit side by side without being averaged together.
A provider on VPS Advisor carries two quite different kinds of evidence, and they are kept apart on purpose.
Reviews: what customers report
People who have used a provider rate it across performance, support, value for money and reliability. This captures everything measurement cannot: whether a ticket got answered, whether the billing was honest, whether the control panel works.
It is also self-selected, unevenly distributed, and sometimes wrong. A provider with three reviews has three opinions, not a rating, and the site says so rather than printing an average of three numbers as though it were the same kind of fact as an average of three hundred.
Monitoring: what the network shows
Independent measurement of reachability, from many places, by machines the provider does not control. It captures what customers are worst at reporting, because an outage at 3am that lasted twenty minutes is invisible to a customer who was asleep and obvious to a probe network.
It is also narrow. It says nothing about support, billing or whether the hardware is what was advertised.
Why they are not combined into one number
Because they are not the same kind of claim, and a single blended score would hide which half it came from. "4.2" made of excellent uptime and poor support means something entirely different from "4.2" made of the reverse, and the reader deciding between two providers needs to know which one they are looking at.
So a listing shows both, side by side, each labelled with what it is and what it was measured from.
What is never invented
- A provider nobody has reviewed shows "not yet rated", never
0.0. A zero is a rating, and nobody gave it one. - A provider that has not opted into monitoring shows no uptime figure, not an estimated one.
- A rating is never adjusted, boosted or weighted by anything commercial. There is no arrangement under which a provider's score can be changed by talking to us.
Replies are not reviews
A provider responding to a review is a reply, not a second review. It does not count toward the review total, the monthly figures, or the rating. This sounds obvious and is a common way for review counts to become quietly wrong.
When something looks off
Reviews can be reported, and are moderated by people. Monitoring verdicts are traceable back to the measurements behind them. Both routes are open to providers and to readers.
Was this page helpful?