Reliability, without invented certainty
Our assessment measures documented safeguards and transparency. No numeric reliability scores have been assigned in this release.
The 20-criterion rubric
Payout safeguards carry 30 points; operational/security transparency 25; refunds, disputes and support 20; portability 15; contractual/product transparency 10. Each dimension contains four equally weighted criteria. Each criterion is met, not met, or unknown, with dated evidence.
Payout safeguards · four criteria, 30 points
Identify the legal settlement counterparty; document standard payout timing and restrictions; document reserve/hold grounds and release rules; provide a described appeal or escalation procedure for holds or closure. Each criterion carries 7.5 points.
Operational transparency · four criteria, 25 points
A dated public component/status history; a published incident response and disclosure process; current security scope and verifiable support for claimed controls; documented notice for material incompatible changes or discontinuation. Each criterion carries 6.25 points.
Refunds and support · four criteria, 20 points
Document financial responsibility; publish refund/dispute costs and deadlines; describe support and escalation channels with their limits; support claimed response commitments with an applicable SLA or a documented comparable test. Each criterion carries 5 points.
Portability · four criteria, 15 points
Document core-data export; explain subscription/payment-data migration limits; define post-termination access and actions; explain migration/deprecation paths. Each criterion carries 3.75 points.
Contractual transparency · four criteria, 10 points
Explain legal model and owner/processor dependencies; disclose tariff rules and material add-ons; publish eligibility and product-acceptance dimensions; reconcile researched material events/orders with current disclosures and remedies without an established unresolved contradiction. Each criterion carries 2.5 points. Absence of an observed incident is not positive evidence.
Unknown is not zero
The lower bound sums weights for met criteria; the upper bound adds unknown weights. Coverage sums the weight of reviewed met and not-met criteria. An interval requires at least 80% coverage and at least two known criteria per dimension. A single score requires 100% coverage. Otherwise: Not rated.
What a score cannot mean
It is not a probability of approval, uptime, bankruptcy or fund protection. Complaints, followers, funding and endorsements do not add or subtract points. Open material risks remain visible independently of a score. The complete proposed rubric is in the project specification; calibration is pending.