Deliverability, and the numbers on your dashboard
Where your email actually lands, provider by provider; choosing which numbers the overview shows; and which yardstick a send is judged against.
EmailAnalytics → Deliverability. An overall open rate hides the thing that matters — one mailbox provider quietly starting to refuse you — so this counts every send by the recipient’s own domain: how many were accepted, how many bounced (and how many for good), how many marked it spam, and how many opened and clicked. If one domain is bad and the rest are fine, it is that provider’s opinion of you, not your writing.
The thresholds are theirs, not ours. Gmail and Yahoo’s bulk-sender rules ask for a spam-complaint rate under 0.1% and start filtering at 0.3%; a permanent-bounce rate over about 2% reads as a bought list. Under fifty sends a domain is not judged at all, because a rate on a handful of emails is noise. What to do when one goes red: stop sending to people who never open, take the hard bounces off, and make sure the address you send from can receive a reply.
Choosing your numbers. The Analytics overview shows four figures by default; “Choose the numbers” lets an owner or admin pick up to eight from conversations, sales, email and people — first reply time, open and click rates, revenue from email, unsubscribes, open tickets, new contacts, average review, and more. They appear in the order you tick them, and each is fetched only when it is asked for, so a dashboard nobody customized costs what it always did. A number that needs a connection you have not made shows a dash and says which, never a zero: on a dashboard those look the same and mean opposite things.
Which yardstick a send is judged against. A results page compares your rates to a benchmark, and which benchmark matters more than the benchmark itself. An automated email — a welcome, an abandoned checkout — is expected and arrives when the person is already thinking about you, so it opens far higher than a newsletter; holding a newsletter to a welcome email’s numbers is how people conclude their newsletter is broken. And a 200-person list opens at rates a 50,000-person list never will. So the comparison uses the narrowest cohort it can name: the kind of send first, then your industry blended with your list size. The figures are rounded from the public 2025 Mailchimp and Klaviyo reports — a yardstick, not a target.