Open rates are now a vanity metric. Here are the metrics to trust.
The question comes up for every email marketer. “Our opens are at 45 percent, why is revenue flat?” The answer, four years after Apple launched Mail Privacy Protection, is that the opens in your dashboard are no longer the signal you think they are. They are a noisy mix of real human reads, Apple’s pre-fetching bots, corporate firewalls pinging every image, and caching layers doing exactly what they were designed to do.
This is not a reason to give up on measurement. It is a reason to rebuild your measurement stack around signals mailbox providers cannot game and pre-fetchers cannot inflate.
What your open rate is actually measuring in 2026
Opens were always a proxy. In 2021, roughly 40 percent of email traffic lived in Apple Mail on iOS and macOS, all of which now pre-loads tracking pixels whether the recipient opened the email or not. Gmail’s image caching and several corporate email gateways add similar inflation. Litmus has tracked the downstream effect on open-rate reliability since MPP launched, and the picture has not improved.
The result is an open-rate metric that runs 15 to 40 percent higher than actual reads, and that correlates weakly with clicks, conversions, or revenue. The metric is not broken in the sense of “completely random.” It still moves in the right direction when your subject lines improve or your sender reputation drops. But the signal-to-noise ratio is bad enough that using opens as a primary KPI leads teams to optimize for the wrong things. You end up chasing subject-line micro-wins while your actual conversion funnel quietly erodes.
The three-layer conversion stack
A modern lifecycle measurement stack is three layers deep. Each layer answers a different question.
Layer 1: Engagement signals you can trust
Clicks are still the most reliable “this person actually did something” metric you can pull straight from your ESP. Inside clicks, watch the ratio of unique clickers to unique openers. If that ratio is rising, your audience is getting more engaged per email. If it is falling, your list is decaying. Scroll depth and dwell time, if your ESP supports them, are even better because they cannot be pre-fetched.
Complement this with a short list of behavioral events outside the email itself: site visits after a send, product views, add-to-carts within the attribution window. If you are on Klaviyo, Braze, or Iterable, these events are already flowing, you just need to surface them in campaign reports.
Layer 2: Downstream conversions and attribution
This is where the real ROI conversation happens. For every send, you want two numbers: revenue attributed to the email within your chosen attribution window, and revenue lift compared to a holdout group that did not receive the send. The holdout is the uncomfortable but essential part of the stack, because it separates “revenue that happened after the email” from “revenue that happened because of the email.”
Multi-touch attribution has also gotten a lot more sophisticated in the last 18 months. Klaviyo’s attribution documentation is a good starting point if you are reconsidering your current model, and most major ESPs now support some version of it natively.
Layer 3: Program health and list quality
The third layer is the slow-moving stuff that predicts whether your program will still work in six months. Subscriber growth, subscriber quality at acquisition, unsubscribe reasons, complaint rate, sender reputation. It is recommended to run this analysis quarterly.
Reporting dashboards that actually get read
A measurement stack is only useful if someone reads it.
Weekly campaign performance. Clicks, click-to-conversion rate, revenue per send, revenue per recipient, and the holdout comparison. Opens go on this dashboard, but as context, not as a headline metric.
Monthly flow performance. Every automation broken out by step, with conversion and revenue at each stage. This is where you catch decaying welcome series and abandoned cart flows that quietly stopped converting.
Quarterly program health. List growth, subscriber quality, reputation, complaint rate, and unsubscribe reasons. Reviewed alongside a forward-looking roadmap.
The consistent thread across all three is that opens appear nowhere in the lead metrics. They are available as a diagnostic when something looks off elsewhere, but they do not drive the narrative.
Final takeaway
Opens are not worthless. They are just no longer a primary KPI, and using them as one is how lifecycle programs end up optimizing the wrong inputs for a full year before anyone notices.
Swap opens for a conversion stack that starts with clicks and behavioral events, anchors in attributed revenue with holdouts, and uses program health as the quarterly backstop. The design of your emails still matters enormously, which is why we wrote the anatomy of a high-converting email as a companion to this measurement piece. But measurement is what turns good design into a program you can defend to a CFO.
If your current dashboard still leads with opens, our Free Email Marketing Audit includes a measurement-stack review. We will tell you exactly which metrics in your current setup are noise and which are signal, and give you a dashboard template to replace what is not working.
