Email opens and clicks: how to spot bot activity
Check inflated opens and clicks in your event campaigns: signals to investigate, provider filters and verifiable human actions.
To investigate suspicious opens and clicks in an event campaign, first check how your provider filters automated activity. Then compare the report with actions such as completing a registration or replying to a question. The open rate alone does not prove that anyone read your message.
Imagine you have sent a reminder and the report looks good, but questions about the time and tickets keep coming in. Before concluding that attendees ignored the instructions, check what the platform counted as an interaction. It may be showing different signals from the ones you expect.
Distinguish an open from an automated link visit
Apple Mail’s privacy protection can preload the tracking pixel even if the person does not open the message. Mailchimp explains this effect in its Mail Privacy Protection FAQ.
A security filter can also inspect links to check their destination. Klaviyo documents these machine-generated clicks in both email and SMS. Loading an image and visiting a link are different events, although both may appear as recipient activity.
A report with plenty of clicks is not enough to diagnose the problem. You need to know what counts as a click, whether automated activity has already been excluded and what action you expected afterwards.
A table to review the report with your team
This table suggests checks, rather than a foolproof way to classify recipients. One signal alone does not justify labelling someone as a bot or removing them from a list.
| Report signal | Possible explanation | What to check |
|---|---|---|
| Opens without other actions | Reading without taking action, preloading or insufficient information | How the platform measures and filters opens |
| Clicks without completed registrations | Automated inspection, abandonment or a problem on the page | Click filtering and the actual registration journey |
| A drop after changing report settings | The activity included has changed | Comparable settings and dates before assessing the message |
| Questions about information already sent | The content did not meet those people’s needs | The specific question and the recipient’s circumstances |
The last row matters even if the other metrics look good. If Clara asks which coach she should take, an email listing all transfers may not have answered her question. Campaign data and a question within a conversation tell you different things.
Check your provider’s filters
In Mailchimp, review the options for filtering automated activity and what they cover in the report. Its documentation on bots and filtering explains which signals it aims to exclude. Check the actual settings on your account before comparing campaigns.
Klaviyo’s documentation describes the “Bot Click” property within click activity. You can use the reporting and exclusion options the platform provides to distinguish this activity. Follow its current documentation; do not assume your tool has the same controls under the same names.
Record the campaign, reporting period, filter settings and target action. If an agency is preparing a report for a client, this brief note prevents two rates from being presented as comparable when they were calculated differently.
If the provider does not distinguish automated activity, explain that limitation. Do not invent a percentage to subtract or adjust the figure by guesswork until it looks reasonable.
Look at what happened after the message
For an event reminder, you can compare completed registrations, replies to a confirmation request or attendance recorded at check-in. Each source demonstrates something different. Downloading a ticket does not confirm attendance; neither does a question about parking.
With EVA, you can start a conversation and see what the person needs to clarify. If they reply “I’m taking the hotel transfer” or ask about a particular entrance, you have an interaction with content that helps you review the communication. This does not make WhatsApp, SMS or chat links immune to automated activity.
In how event communication changes when you move from sending messages to having conversations, we explore that difference. For this review, start with a narrower question: what result did you want, and which source can confirm it?
EVA combines messages prepared by the organiser with answers to attendee questions. You can assess those conversations alongside your campaign metrics, without treating an open as sufficient proof of attention.
Frequently asked questions
Are all clicks without a registration made by bots?
No. A person may visit and leave, encounter an error or decide to register later. Review the filtering and the journey before attributing the click to a machine.
Does a fall in opens mean the email is performing worse?
Not necessarily. What the report counts may also have changed. Compare campaigns using a consistent method and look at the action you care about.
Can I tell who actually read an email?
A recorded open does not, on its own, demonstrate human reading. A reply or completed action provides other evidence, but cannot reconstruct all the attention that person paid.
