Campaigns analytics roll-up

The campaigns insights tab in one response for the whole account (every campaign the token owner can see): the email trend and the LinkedIn trend — each as { totals, series }, where totals is the exact whole-window figure per metric and series one zero-filled point per bucket carrying the same fields — plus the top 10 LinkedIn campaigns by acceptance rate. Derive rates from totals; contacted is the distinct-contact count the per-contact metrics divide by (email and LinkedIn reply rate = replied / contacted, LinkedIn acceptance rate = invitesAccepted / invitesSent). Email open counts cover only emails sent while the campaign's open tracking was on, so open rates divide by the tracked sends: uniqueOpened / contactedWithOpenTracking per contact, emailsOpened / sentWithOpenTracking per email. Every count is event-date attributed — each row lands in the bucket its event happened in — so the series describe volume over time and sum exactly to totals; do not divide one point's engagement by the same point's sends as a per-bucket rate, because opens, acceptances and replies lag their send and the two counts describe different people. Buckets are daily for 7d and ISO-weekly for 30d and 90d; each point's date is its bucket's start date as a display label ('Jan 1') with the same date as YYYY-MM-DD in isoDate. Each panel is computed independently and is null when its aggregation fails, so always null-check them. Per-campaign detail stays on GET /business/campaigns/{campaignId}/stats. Rate limit: 20 requests per minute, 1000 per day.

Errors:

• 422 VALIDATION_FAILED — range was sent and is not one of 7d, 30d or 90d.
• 429 RATE_LIMITED — Per-user quota for this endpoint family exceeded (20 per minute, 1000 per day); retry after the Retry-After / RateLimit-Reset seconds.

Requires one of the following token scopes: analytics.