Block: channel-performance

A 6-row table showing how each traffic channel is performing — sessions, engagement, conversion, and revenue per session.

What it shows

Column	Format	Notes
Channel	String	One of six buckets (see channel mapping below)
Sessions	Integer	Channel volume
Engagement	Percentage with benchmark indicator	vs site-wide engagement
Conversion	Percentage with benchmark indicator	vs site-wide conversion
RPS	Currency with benchmark indicator	vs site-wide RPS

The last three columns each carry a centered-bar benchmark indicator showing how the channel compares to the site-wide average for that metric.

Channel mapping

Noibu returns sessions grouped by UTM_SOURCE (and optionally UTM_MEDIUM). The skill buckets these into six rows:

Channel	Match rule
Direct	`utm_source` is empty/null
Paid	`utm_medium` ∈ (cpc, ppc, paid, paid_social, paidsocial)
Organic	`utm_medium` = organic OR referrer matches a search engine domain (google.com, bing.com, etc.)
Email	`utm_medium` = email OR `utm_source` matches known email tools (klaviyo, mailchimp, shopify_email, etc.)
Social	`utm_medium` = social OR `utm_source` matches social platforms (facebook, fb, instagram, ig, tiktok, twitter, etc.)
Referral	Anything else with a `utm_source`

The skill applies this mapping after the query returns. Order rows by sessions descending.

Data source

Noibu only. UTM data and revenue both come from Noibu's session-level fields.

MCP calls

One query per period, grouped by UTM_SOURCE and UTM_MEDIUM:

{
  "queryInput": {
    "measures": [
      { "aggregate": { "measureAlias": "sessions", "measureFunc": "COUNT", "target": { "field": "SESSION_ID" } } },
      { "predefined": { "measure": "BOUNCE_RATE", "measureAlias": "bounce_rate" } },
      { "predefined": { "measure": "CONVERSION_RATE", "measureAlias": "cvr" } },
      { "predefined": { "measure": "REVENUE_PER_SESSION", "measureAlias": "rps" } }
    ],
    "groupBy": { "fieldSegments": ["UTM_SOURCE", "UTM_MEDIUM"] },
    "orderBy": { "measureAlias": "sessions", "direction": "DESCENDING" },
    "limit": 50
  }
}

After the query returns, apply the channel mapping above. Aggregate rows that fall into the same bucket: sum sessions, weighted-average engagement / CVR / RPS by sessions.

Site-wide reference (for indicators)

The benchmark tick on each row's indicators is the site-wide value of that metric for the same period. If the core-kpis block is enabled, read those values from its response. If not, run a separate site-wide query (same shape as the per-channel query but without groupBy).

Methodology tooltips

Direct (the row label) → "Sessions with no UTM source. Includes typed URLs, bookmarks, in-app browsers, and links with stripped referrer data — so it may absorb some untagged paid or social traffic."
RPS (column header) → "Revenue per session — channel revenue divided by channel sessions."

The Engagement and Conversion column headers don't need tooltips — they're spelled out, and the engagement definition is already covered in the core-kpis tooltip if that block is on the dashboard.

Empty state

If a channel bucket has zero sessions in the window, omit the row entirely — don't render an empty row with dashes. If ALL non-direct buckets are empty (the user has no UTM tracking set up), render the block with just the Direct row and a small note: "No tagged traffic in this window."

Polarity

All four metrics are higher-is-better. Above-site indicators in success-green; below in warning-orange.

Block: channel-performance