Reading analytics and the deliverability health score
What each dashboard view shows, how the 0–100 health score is composed, and the reading order that finds problems fastest.
The Analytics section reports everything SendCanyon observes about your sending, sliced three ways: by campaign, by mailbox, and by domain, each with a selectable date range and CSV export. The same events feed all three views — the slicing exists because problems hide in aggregates.
The three views
- By campaign — sends, delivery, opens, clicks, replies, bounces, unsubscribes per campaign and per step. This is where you compare copy: step-level reply rates show which email in a sequence earns the responses.
- By mailbox — the same metrics attributed to each sending mailbox, plus warmup stage and daily-limit utilization. This is where a single sick sender becomes visible inside a healthy fleet.
- By domain — the aggregate reputation view: rates across all mailboxes on the domain, DNS verification status, and the domain's health score history.
The deliverability health score
Every domain and mailbox carries a 0–100 health score, recomputed daily. It exists so that the two metrics that kill sending programs — bounces and complaints — can't hide inside averaged-out dashboards. The composition:
| Component | Weight | What moves it |
|---|---|---|
| Hard-bounce rate (14-day) | 35% | Full marks under 1%; zero at 5%+ |
| Spam-complaint rate (30-day) | 30% | Full marks under 0.05%; zero at 0.3%+ |
| Reply trend (30-day) | 20% | Replies per send versus your trailing baseline |
| Warmup engagement & consistency | 15% | Peer-network engagement, volume smoothness, DNS still verified |
Read the bands, not the digits: 80–100 healthy, keep going; 60–79 something is drifting — open the component breakdown and find which input slipped; below 60 stop scaling and fix — the score page lists the contributing events. The score is also the gate for a few automatic protections: a mailbox falling below 40 has its campaign sending paused and its warmup floor raised until it recovers.
A reading order that works
- Start at Domains: any domain score off its baseline? If yes, drill into its mailboxes; if one mailbox owns the damage, the problem is that sender — if all dropped together, the problem is a campaign or the DNS.
- Then Campaigns: bounce rate per campaign against the 2% line, replies per step against your baseline. A campaign with clean bounces and no replies is a copy problem, not an infrastructure one.
- Weekly, not hourly. Daily numbers on small volumes are noise; the useful signals are 7-day trends.
Exports
Every table exports to CSV from the range you're viewing, and everything on these dashboards is also available programmatically via GET /api/v1/analytics for your own BI — see API keys and authentication.