Relevance OS Dashboard
Four tiers (relevance, engagement, revenue, adaptation) along the grid with Maya's contribution isolated by holdout and one-click actions.
Relevance OS is the analytics layer that treats relevance as the leading indicator of revenue. It shows how well each feed matched its viewer, what that did to engagement and revenue, and what to change next. A standing holdout isolates the contribution of Maya, Genuin's relevance engine, so you read lift directly.
This capability arrives as a proposal when your network needs it, filled in from your brand context, and goes live when you confirm it. It's part of Octo, and Relevance OS: reading the analytics dashboard explains each tier and its measures.
When to use it
Use it for a weekly review, and any time a parameter changes. It shows whether a change helped engagement before revenue moves.
| Occasion | What to read |
|---|---|
| Weekly review | The Adaptation tier, where the recommended changes wait for your approval |
| A feed setting changed | That change's effect against the holdout, set in Player Controls and Feed Settings |
| An experiment finished | The lift results; you promote the winner with one click |
| A partner or floor question | The Revenue tier by partner, next to the floor and pacing recommendations |
How it works
- Every tier rolls up along the Hierarchical Grid. You filter it by cohort, experience, content source, and partner.
- Each placement keeps a standing holdout that isolates Maya's contribution. Lift is read directly rather than inferred.
- The Adaptation tier lists recommendations, each with its expected effect. Examples include re-cutting a video for product pages, raising a floor, running a Quest in a group, or pausing a partner.
- One click sends a recommendation to Studio, Canvas, or Monetize. Under Canvas guardrails, it can also run automatically.
- Relevance OS reports for US customers by default. Other regions follow the data residency settings in Security, data residency, and compliance.
- You can ask any question about the numbers in plain language from the report, through Octo Ask.
Set it up
Octo proposes the dashboard with these steps already done. You review the result and confirm it.
- Open Relevance OS from Brand Control Center (BCC).
- Filter by placement, cohort, experience, or partner. Audience segments: onsite, offsite, and offline signals explains how cohorts are built.
- Send recommendations to Studio, Canvas, or Monetize. Canvas: run your network with growth and monetization agents covers what runs automatically.
Settings
| Setting | What it controls |
|---|---|
| Holdout share | The share of each placement held out to isolate Maya's contribution; 5 percent by default |
| Recommendation auto-apply | Whether recommendations run without a click, within Canvas guardrails |
Change the holdout share per placement in Placement Experiments and Holdouts. Auto-apply follows the guardrails and approval mode you set in Canvas.
What it reports
Relevance OS reports all four tiers for each level of the grid. Measure engagement explains how the engagement measures are counted.
- Relevance: match score, exploration share, and repeat rate. Did the right video reach the right viewer in the right context?
- Engagement: watch time per user per minute, lift against the holdout, completion, and follows.
- Revenue: eCPM, fill, revenue per minute, and attributed commerce.
- Adaptation: Studio adaptation suggestions, Canvas actions, and floor and pacing recommendations.
Connect your systems
The engagement events behind these tiers reach your own stack today. Receive engagement events in your own analytics captures what members watch and do in Genuin views, live from the SDKs and in batches to your warehouse. Query Genuin event data in BigQuery covers the genuin_data table and first queries on engagement.
The relevance report calls in the Business Account API are proposed, not live. Live API today and the Business Account API roadmap maps what you can call now to what's planned.
