Find changes, outliers, raid influence and data-quality problems before drawing conclusions.
Turn your Twitch data into direction.
Wayfinder is not a second Twitch dashboard. It turns your own historical data into evidence, experiments, guardrails and next decisions while separating confirmed external audience events from your normal baseline.
Twitch connection is optional
Connect your own Twitch account to match supported VOD, clip, and planned schedule context. Verified raid events can be added later when EventSub persistence is enabled; Wayfinder requests only the access it needs.
Prioritize the signals that matter for your goal instead of staring at every available metric.
Run controlled creator experiments and compare similar broadcasts instead of changing everything at once.
What are you trying to improve?
What to do, not just what happened
Wayfinder ranks actions by repeat evidence, excludes confirmed external distortions, and tells you when the dataset is too weak for a conclusion.
Claims earn confidence
No hidden creator score. Each claim shows its effect, sample size, comparison method and evidence level.
Turn promising signals into experiments
Wayfinder proposes controlled tests from your own data instead of telling you to blindly stream more.
What not to chase
One-off spikes, thin samples and unexplained anomalies are leads to investigate, not instructions to rebuild your channel.
Reference metrics
These numbers are here for context. Wayfinder's primary product is the decision layer above them.
Signals with receipts
Every signal has a confidence label and a Show the Math breakdown.
Is the audience changing, not just the count?
Wayfinder compares returning/new engaged viewers, chat depth and clip creation when those fields exist. These are behavioral proxies, not value judgments about viewers.
Wayfinder scorecard
No mystery 0–100 score. Each status is calculated independently.
Analyze broadcasts in context
External audience events remain visible, but confirmed distortions are removed from organic recommendation baselines.
| Date / title | Category | Avg | Peak | Follows/hr | Influence | Comparable | Context |
|---|
Compare repeatable conditions
Confirmed raid/front-page/promotion streams are excluded from organic pattern recommendations but remain visible in Stream Lab.
Organic average viewers
Organic average viewers
Organic average viewers
Organic followers per hour
Different games can do different jobs
Wayfinder avoids declaring a single “best game” when reach, conversion and consistency tell different stories.
Change one thing on purpose
Recent period vs earlier period
Wayfinder surfaces conditions that changed alongside performance. It does not claim those conditions caused the result.
Did the audience lift persist?
Available only when Wayfinder has a confirmed raid event matched to a dated stream.
Know what Wayfinder can actually support
Missing fields reduce confidence instead of being silently treated as zeros.
Know where every layer came from
Wayfinder uses an explicit source contract. Official Twitch data adds verified context; TwitchTracker may corroborate 30-day aggregates but never overrides Twitch data.
Twitch Analytics CSV
WaitingRequired source; processed locallyTwitch API + EventSub
OptionalSupported Helix VOD/clip metadata + verified EventSub contextTwitchTracker
OptionalSanitized 30-day corroboration onlyWhat each source is allowed to prove
Unsupported fields are discarded. Planned data is never treated as observed history, and third-party data cannot override first-party data.
Do independent sources agree?
TwitchTracker is used only when its 30-day window can be compared fairly with the uploaded Twitch data.
Sanitized TwitchTracker 30-day summary
No TwitchTracker data loaded.