of the 85 available when every component is measured
Evidence held
41%
2 of 5 components
Penalty
None
Weight in overall
30%
of the 7.2/10 score
Nothing on this page can be bought. Sponsorship, affiliate status, paid listings and advertising add exactly zero points. That claim is checkable rather than rhetorical: the formula is a single exported function, rendered term by term on the methodology page, and none of its inputs is a commercial relationship.
Where the points came from
Each bar is one component. A component we have not observed shows a dash, not a zero and not a neutral share: it is dropped from the average entirely, which costs coverage rather than points. Verification is listed as a bonus because not having bought from a panel ourselves is a fact about our workload, not about the panel.
Domain agenot measured
— / 20
WHOIS lookup pending. Left out of the average until we have a date.
How long the domain has been registered, from WHOIS. Weighted heavily because it is the one signal an operator cannot buy retroactively: a panel that exits and reopens under a new name starts this component at zero.
Uptime, last 30 days
+22.1 / 25
96.8% of 31 checks succeeded.
The share of our scheduled probes that got a healthy response, Laplace-smoothed — one imaginary success and one imaginary failure are added to the sample, so a short run cannot claim a perfect record or be condemned for a single miss. The smoothed rate is then squared, so 95% uptime is worth much less than 99% rather than slightly less: an SMM panel that is down when you need to place an order is down.
Catalogue freshness
+10.0 / 10
Prices synced within the last 24 hours.
How recently we last read this panel’s price list successfully. A catalogue we cannot re-read is a catalogue whose prices we cannot stand behind, so staleness we can observe costs points even when the panel itself is healthy. A panel we have never synced is a different case and is not scored here at all: that is a missing API key on our side, so it costs coverage instead of points.
Verified buyer reviewsnot measured
— / 20
No verified reviews yet. Unreviewed is not the same as badly reviewed, so this is left out of the average and counted against coverage instead.
Average rating over reviews where a moderator saw proof of purchase, pooled with 8 imaginary reviews at the index-wide average — so a thin run of five-star reviews is pulled back toward the market and cannot outrank a long record. Below 3 verified reviews the component is not scored at all. It moves in both directions: a rating under the market drags this score down, not merely fails to lift it.
Independent verificationbonus
+5.0 / 15
Verified: owner verified control of the domain.
Ten points if we placed a real order and it delivered, five if the owner proved control of the domain. The test purchase is the only signal on this page that we generate ourselves, which is why it carries the most weight per point.
Published policiesnot measured
— / 10
Not crawled for terms, refund policy or a support channel yet.
Terms, a refund policy and a support channel that answers. Cheap to satisfy and cheap to fake, so it is worth little, but a panel with none of them has told you something.
Recent health checks
Raw probe results. We store the observations rather than a rolled-up percentage, so the score can be recomputed if the formula changes.