A future view could compare aggregate viewing before, during, and after a
documented Drops campaign. It must expose coverage and uncertainty and must not present
correlation as campaign-caused growth.
Comparison frame
no line, bars, or axis values
Aggregate viewership around a reward window
Values omitted
measurement withheld
Comparable baselineaggregate pending
Sourced campaign windowaggregate pending
Comparable follow-throughaggregate pending
Illustrative layout only. Values intentionally omitted. The campaign boundary would come from cited reward facts; the aggregate series would come from a separately governed source.
No causal claimA visible change would not prove that rewards caused
it. Game releases, front-page promotion, creator schedules, esports, and coverage gaps can all
confound the comparison.
Evidence that must travel with the view
planned disclosure set
Campaign
What window is being compared?
Campaign identity, earning start and end, eligible-channel rule, source, observation time, and corrections.
Aggregate
What was actually observed?
Measure, grain, platform scope, collection intervals, missing intervals, and whether replays or raids are represented.
Baseline
Is the comparison defensible?
Comparable periods, category continuity, time-zone treatment, and known releases or broadcasts that could distort context.
Result
How uncertain is it?
Coverage statement, suppression rule, correction status, and language that distinguishes association from causation.
Fail-closed states
No chart is better than an unsupported chart.
Withhold the view when a campaign boundary is unknown or conflicted.
Mark incomplete observation windows instead of interpolating them.
Suppress slices that violate the approved publication threshold.
Retract or annotate results when an upstream campaign fact is corrected.
Production boundary
Aggregates wait for a governed data posture.
Real measurement requires GOV-003 to define permitted Twitch Helix aggregate collection,
source attribution, observation windows, retention, corrections, freshness, and safe public
thresholds. Until that posture is approved, StreamStats publishes no measured result.