Email Campaign Reporting: What to Track in Braze

Apple Mail Privacy Protection changed what an open means, so Braze now splits opens into machine opens and other opens and adds a modelled estimate of real opens. Clicks are the steadier signal, and delivery figures need time to settle. Read delivery first, clicks second and opens last, and hold your campaign review until the retry window has closed.

‍

Key Takeaways

• A message Braze can't deliver on the first attempt is deferred, and temporary soft bounces are retried for up to 72 hours, so early delivery figures are provisional.
• Machine opens flag Apple Mail Privacy Protection users, while other opens are human opens that the feature doesn't affect.
• Estimated Real Opens is a modelled figure from Braze that approximates unique opens as if machine opens did not exist.
• Clicks can outnumber opens when a message is clipped, images are blocked or security software follows links.
• The spam figure counts delivered messages that recipients marked as spam, and it excludes mail that lands in the spam folder.
• Campaign Analytics defaults to the last 90 days, so an older campaign shows zero until you widen the range.

‍

When should you read the report?

Not on send day. Braze treats a message that can't be delivered on its first attempt as deferred. Temporary failures, usually an SMTP 4xx code, are retried until the message is delivered or 72 hours have passed.

Read a report inside that window and you're reading a send that hasn't finished. Deliveries can still climb while retries succeed, so the rate you saw at breakfast may not be the one you end up with.

We hold campaign reviews until the retry window has closed and treat anything pulled earlier as provisional. That's our rule rather than a Braze instruction. It exists because an early delivery figure tends to get quoted later as though it were final.

One more timing trap. By default, Campaign Analytics shows the last 90 days, so a campaign launched before that shows zero until you widen the range. The report isn't empty: the date range just doesn't reach back far enough.

‍

What do deferrals and bounces do to the numbers?

Braze counts a delivery when the receiving server accepts the message request. That's acceptance, not inbox placement, and the gap between the two is where many misreadings start.

Hard bounces are permanent failures, typically a 500-series SMTP reply, and Braze marks those addresses as invalid. Soft bounces are temporary failures that get retried, which is why a send can look worse early and better later.

Spam rate needs the same care. Braze defines it as the rate at which recipients mark a message as spam while viewing it, and it does not include mail that lands in the spam folder.

So a clean spam figure tells you how recipients reacted, not where the message landed. We read it alongside delivery and click behaviour rather than on its own.

For a rough delivery check, the Braze email FAQ describes a typical "magic number" of around 98% of messages delivered, with a bounce rate no higher than 3%. Treat that as a sanity check on your own baseline rather than a target.

‍

Which opens should you believe?

Braze publishes four open metrics, and they answer different questions.

‍

The distinction deserves a slow read. Machine opens include human opens as well as automated ones. The split separates opens that Mail Privacy Protection touches from opens it doesn't, so don't read the machine figure as a count of bots.

Estimated Real Opens exists to give you a comparable number when the raw open count has changed meaning. We use it as the baseline when comparing variants across audiences with different device mixes, and we keep clicks as the primary success signal. That's our practice, not Braze guidance.

‍

Why are clicks higher than opens?

Total Clicks counts every click within the delivered email, while Total Opens counts opens, so the two can diverge without anything being broken. The Braze email FAQ lists several causes.

• A user clicks more than once within a single open, or clicks in a preview pane without a full open being recorded.
• The message is clipped by the recipient's provider, such as Gmail clipping messages over approximately 102 KB, which can cut off the tracking pixel at the bottom.
• Images are disabled, so the open can't be tracked.
• Corporate email security tools scan incoming messages by clicking every link, which logs clicks the recipient never made.

‍

Of these, clipping is the one you control, because message size is yours to manage. The others sit with the recipient or their security team. That split is our reading, not Braze's.

Unique clicks are tracked over a seven-day period for email and measured per dispatch_id, which is a single send attempt.

‍

What else distorts the panel?

Audience size changes how a figure is produced. In a workspace with 50,000 users or fewer, Braze runs a full scan of workspace profiles for the default statistic. Above that, audience size is estimated from a sample and extrapolated, and Braze expects margins of error, especially for segments that are small relative to the workspace.

For a small segment in a large workspace, we treat the audience figure as indicative and check it against a count we can reproduce before quoting it to anyone.

‍

How does this change what you send?

Reporting shapes the next send or it is decoration. Three connections are worth drawing.

If clicks are the figure you trust, the link becomes the unit of measurement, and dynamic content changes what a click means for each recipient. Plan the links in a message as deliberately as the copy around them.

Personalised content needs a comparable baseline to prove itself, which is what Estimated Real Opens gives you. Comparing a personalised variant on raw opens across audiences with different device mixes compares two different things.

Lifecycle work is where the timing rule bites hardest. An onboarding journey and the activation campaigns around it send continuously, so there's never a moment when every retry has finished. Pick a reporting cadence and hold it. The same applies to retention campaigns, where a win-back send read too early looks worse than it is.

‍

Frequently Asked Questions
‍

1. Why does my campaign show zero analytics?

Campaign Analytics displays the last 90 days by default. A campaign launched before that shows zero until you widen the time range.

2. How long should I wait before judging an email send?

At least until the retry window closes. Braze retries soft bounces that look temporary for up to 72 hours, so we hold reviews until then.

3. What's the difference between machine opens and other opens?

Machine opens indicate an Apple Mail Privacy Protection user and include both human and non-human opens. Other opens are human opens that the feature doesn't affect.

4. Why are my clicks higher than my opens?

The tracking pixel may not have loaded. Clipping, disabled images and security tools that follow links can all produce clicks without a matching open, and a user can click several times within one open.

5. Does the spam rate include mail that lands in the spam folder?

No. Braze defines it as the rate at which recipients mark a message as spam while viewing it, and it excludes mail that lands in the spam folder.

‍

Sources

• Braze, Email reporting. Retrieved 01 Oct 2026.
• Braze, Email analytics glossary. Retrieved 01 Oct 2026.
• Braze, Email deliverability glossary. Retrieved 01 Oct 2026.
• Braze, Email FAQ. Retrieved 01 Oct 2026.
• Written by CustomerIK, a Braze implementation partner.

‍

CustomerIK is a Braze implementation partner working across onboarding, technical integration, marketing operations and customer data management.

If your email reports never quite agree with each other, let's talk.

‍