You send an important message to a company address and, almost before you have closed the compose window, the tracker reports an open. It is tempting to think the recipient was waiting for it. Far more often, a machine got there first.
Who opens email before people do
- Secure email gateways. Many organisations route incoming mail through a filtering service that inspects attachments, links and sometimes images before delivery.
- Link and image sandboxes. Some security products fetch every remote resource in a message to check it is safe, which requests the tracking image as a side effect.
- Privacy proxies. Some consumer mail apps download remote images in advance through their own servers, whether or not the message is ever read.
How automated opens look
- Timing. Within a few seconds to a couple of minutes of sending, often before the message could realistically reach a person.
- Network. Requests from data-centre or security-vendor address ranges instead of an office or consumer connection, when the log shows a network owner.
- User-agent. Generic library strings or headless-browser markers rather than a recognisable mail client.
- Grouping. Several fetches at almost the same moment, sometimes alongside requests for every link in the message.
Reading the rest of the log
Set the early automated fetch aside and look at what comes after. A later open during the recipient's working hours, from a normal network and client, is the one that more likely reflects a person. If there is nothing after the scanner, treat the message as delivered but not yet confirmed as read.
Keep each message on its own image so that one company's scanner cannot be confused with another's. A per-message generator like MailPixel makes that easy, and the raw request details it records are what let you tell the gateway apart from the person behind it.
A quick worked example
Suppose you send a quote to a buyer at a large company at 10:02. The log for that message shows:
| Time | Network | Client string | Likely source |
|---|---|---|---|
| 10:02:07 | Cloud provider range | Generic HTTP library | Mail gateway or scanner |
| 10:02:08 | Same range | Same string | Same scan, second pass |
| 14:31 | Corporate office range | Desktop mail client | A person reading |
The first two rows arrived before the message could plausibly have reached an inbox and been read, and they share an automated-looking client string. The third row has a normal client, a normal network and a sensible time. That last row is the one worth acting on, and it suggests that a sensible follow-up is the next morning rather than at 10:05.
What not to do
- Do not reply within minutes because of an instant "open". You will look as if you are watching the inbox, and the person probably has not seen the email yet.
- Do not count the scanner's fetch as engagement in any report you share with others.
- Do not try to hide the image from scanners with tricks. It rarely works for long and makes the message look more suspicious to the same filters.