Open rates lied. What should you watch?Delivery, bounce, complaint and unsubscribe trend decide more than a pixel ever did.

An open is recorded when a tracking pixel loads, and for years that meant a person had opened the message and their client had fetched images. Mail clients that prefetch and cache images before a person opens anything, done for privacy and performance reasons, break that assumption: the pixel can load whether or not anyone reads the email. This is common knowledge among people who run campaigns now, which is exactly why a dashboard that still leads with a big open-rate number is showing you the least reliable metric it has.

Opens are not worthless. As a rough, campaign-to-campaign comparison within the same list and the same mail client mix, a sharp drop is still worth a look. What they cannot do is tell you, in an absolute sense, how many people actually read a given send, or predict whether your next campaign will land in the inbox.

• Delivery rate. The share of a send that reached a mailbox at all, distinct from bounces and blocks. A slipping delivery rate on a list that used to deliver cleanly is the earliest sign something changed on the sending side.
• Bounce rate. Hard bounces mean the address does not exist; a rising rate usually means the list is aging or was never fully verified. This is one of the two numbers a mailbox provider and Amazon SES both watch closely.
• Complaint rate. The share of delivered mail marked as spam by the recipient. It is the second number SES watches, and the one hardest to recover from once it climbs, because a complaint is a person actively telling a provider they did not want the message.
• Unsubscribe trend. Not the raw count on one send, but the direction over several. A steady climb usually means content or frequency drifted from what the list signed up for; a spike on one send usually means that specific send missed.
• Per-campaign engagement, relative. Comparing this week's click-through against your own list's recent average, rather than against an industry benchmark, filters out the prefetch noise and shows whether a specific send actually worked.

Two of the five metrics above are not just good practice on SES, they are the numbers your sending privileges depend on. AWS documents review territory around a 5 percent bounce rate and a 0.1 percent complaint rate, with real enforcement risk once bounces approach 10 percent or complaints approach 0.5 percent. The SES console's reputation dashboard shows both against those lines, along with an account status of healthy, under review, or paused.

Nothing in that dashboard mentions opens. A campaign with a beautiful open rate and a rising complaint rate is still a campaign heading toward a paused account; a campaign with a flat open rate and clean bounce and complaint numbers is a healthy send that a vanity metric made to look boring.

1. Delivery rate: check it dropped less than a point from the prior week; if not, look for a specific bounce spike behind it.
2. Bounce rate: confirm it is still comfortably under the review line, and remove or re-verify addresses that hard-bounced.
3. Complaint rate: read the number before you read anything else on the dashboard; a rising trend here outranks every other metric this week.
4. Unsubscribe trend: compare the last three sends, not just the latest one, and check whether frequency or content shifted.
5. Relative engagement: compare this send's click-through against your list's own trailing average, not an external benchmark.

SendBunny surfaces delivery, bounce, complaint and unsubscribe numbers per campaign, alongside the same account-level reputation view Amazon SES shows in its own console, inside a platform installed into your AWS account for a one-time $199. It does not fix list hygiene or content on your behalf; the visibility is the product, and the decisions stay yours.