Useful rewards research starts with evidence.

A concrete finding from our documentation review

When reviewing Freecash alternatives on 15 October 2026, we found that Freecash’s dedicated first-withdrawal guide describes a region-dependent threshold, while its beginner guide describes a fixed first-withdrawal threshold. This is an observation about public documentation, not a measurement of member outcomes.

The practical implication is to confirm the requirement in the account’s current Cashout flow and ask the provider about conflicting guidance. We do not resolve the discrepancy by publishing either number as a universal minimum.

• Dedicated first-withdrawal guide
• Beginner signup-to-cashout guide

We separate observations like this from verified activity results. A source can document a rule without establishing how often someone qualifies or how quickly a reward is fulfilled.

Survey availability by country

Measure one successful inventory snapshot per participating member per UTC day. Count distinct provider-and-survey combinations in that snapshot. Include valid zero-inventory snapshots; exclude failed requests and report their coverage separately in the source review.

Divide the total listings across snapshots by observed member-days in each country. Count distinct participating members separately. This describes availability for the observed cohort, not every member in the country, unique studies across the whole period, qualification rates or earnings.

Missing-credit causes

Use resolved offer-credit cases with documented investigation outcomes. Deduplicate repeated tickets about the same activity. Separate confirmed causes from unresolved cases and partner decisions that lack enough evidence to classify.

A support-ticket sample represents people who reported problems. It cannot establish the failure rate of all offers. A future release must show its case denominator, unknown outcomes and classification method before describing a cause as common.

Reward processing and fulfillment

Measure submission, approval and fulfillment as separate events. Use a stated observation window and show pending, cancelled and failed requests separately from completed requests. A median based only on completed requests cannot predict how quickly every new request will finish.

Privacy and publication checks

Publish aggregates only. Do not release account identifiers, emails, IP addresses, device identifiers, ticket text or individual transaction records. Require at least 30 distinct participants and 30 counted survey listings per published country row, and review combinations of releases for possible inference before approval.

Small or sparse groups are withheld rather than treated as zero. An approved production export, reproducible metric definition and valid dates are required. Development snapshots are used for implementation checks and are not presented as current production results.

Comparisons need their own evidence

Public help pages can establish documented requirements. Claims about earnings, reliability or speed require matched activity logs, recorded costs and verified outcomes. Our current comparison pages disclose their documentation-based scope.