Open Rate Isn't a Real Metric: How Apple MPP Broke Cold Email Tracking

If you're still using open rate as a primary cold email metric, you're optimizing for a fiction. Apple Mail Privacy Protection (MPP), which rolled out in 2021 and reached broad adoption by 2024, fundamentally broke the metric — and most cold email teams haven't adjusted.

Across our Growtoro client base, "open rates" are routinely inflated by 25–60% relative to actual human opens. The teams that have updated their measurement stack have a much clearer view of campaign health; the teams still chasing open rate as a north star are making decisions on bad data.

What Apple MPP Actually Does

When an iOS or macOS user opens any email in Apple Mail with MPP enabled (the default since iOS 15), Apple's privacy proxy:

1. Pre-fetches every image in every email — whether the user actively viewed it or not.
2. Loads the pre-fetched content through Apple's proxy servers (so the sender's IP-based location tracking is also defeated).
3. Renders the email locally only when the user actually opens it.

From the sender's perspective, every MPP-enabled recipient registers as having "opened" every email — within minutes of delivery, regardless of whether the human ever looked.

The Inflation Math

The percentage of "fake" opens depends on your audience's Apple ecosystem penetration:

• Heavy iPhone audience (creative, executives, consumer): 50–70% of opens are MPP artifacts.
• Mixed B2B audience: 35–55%.
• Enterprise/Outlook-heavy audience: 15–30%.
• Specific verticals (developer tools, legacy enterprise): Sometimes under 15%.

The implication: a campaign showing "62% open rate" might have an actual human open rate of 25–40%. Two campaigns showing the same reported open rate could have wildly different real engagement, depending on audience composition.

What Open Rate Now Can and Cannot Tell You

What it can still tell you:

• Massive deliverability problems (open rate near zero usually means the email isn't being delivered at all).
• Coarse subject line effectiveness when comparing similar audiences against each other.
• Trend changes over time within the same audience composition.

What it can't tell you:

• Whether a specific recipient actually read your message.
• Comparative performance across audiences with different Apple ecosystem density.
• Engagement quality.
• The actual subject line lift across A/B tests with small sample sizes.

The Metrics That Replaced Open Rate

1. Reply rate. The clearest signal that a human read, processed, and responded to the message. Hard to fake, hard for MPP to inflate.

2. Click rate (on the rare cold email with a tracked link). Less MPP-inflated than opens because clicks require a real action. Still partially inflated by some security scanners.

3. Sequence-completion-to-reply ratio. What percentage of recipients who received all 4 emails ultimately replied? A direct measure of sequence effectiveness.

4. Time-to-reply distribution. Real human replies follow specific time-of-day patterns. A reply graph that doesn't follow human work-hour patterns suggests automation or bot replies.

5. Conversion to meeting per 1,000 sends. The metric that maps directly to revenue. The only true north star.

The A/B Testing Problem

Apple MPP didn't just break the metric — it broke A/B testing on subject lines. Comparing two subject lines by their open rates is meaningless if the audience compositions differ even slightly, because MPP inflation is correlated with iOS share.

The fix: A/B test on reply rate, not open rate. Requires larger sample sizes (you need 2–3x the volume to reach statistical significance on a less-frequent event), but produces real signal.

For high-volume campaigns, this is doable. For low-volume campaigns (under 500 sends per variant), subject line A/B tests on open rate are essentially noise — be conservative about reading anything into them.

The Filtering Mistake

One common but wrong response to MPP: "filter out Apple opens and use the remaining data as ground truth." This doesn't work, because:

• The user-agent strings MPP uses are not reliably identifiable.
• Even non-Apple opens include pre-fetching by some corporate security scanners, ad-blockers, and email clients.
• The "real" open data left after filtering is itself biased toward non-Apple users — useful for some signal but not generalizable.

Better to accept that open rate is broken and shift to engagement-action metrics (replies, clicks, meetings).

What This Means for Cold Email Strategy

Three implications for how to run cold email in 2026:

1. Optimize subject lines for inbox sorting, not open rate. A subject line's job isn't just to drive opens — it's to position the email correctly in the inbox triage. "Quick q on [Company]'s outbound" sorts as "looks like a peer/colleague email" regardless of whether MPP pre-fetches it.

2. Optimize body copy for the readers who actually open. If 40% of "opens" are MPP and 60% are humans, the humans matter and the MPP artifacts don't. Write for the humans.

3. Track reply rate as the engagement north star. Build dashboards around reply rate, qualified-reply rate, and meetings-per-1000-sends. Open rate becomes a secondary signal, not a primary one.

The Tooling Implications

Most sending platforms still report open rate as a primary metric. Some now offer "MPP-adjusted" open rates — these are approximations and shouldn't be treated as ground truth.

The right tooling stack for engagement measurement in 2026:

• Primary metrics dashboard: reply rate, meetings booked, pipeline generated.
• Secondary metrics: bounce rate, complaint rate, sequence drop-off.
• Tertiary metrics: open rate (for trend monitoring only, not optimization). The same MPP-driven measurement noise affects newsletter open rates — for the diagnostic pattern when newsletter opens drop suddenly, see the newsletter open rate decline guide.

The Bottom Line

Open rate is a metric that lost its meaning when Apple MPP rolled out. Cold email teams that still optimize for it are making decisions on substantially inflated data. Shift the measurement stack to reply rate, meeting rate, and pipeline conversion — the metrics that actually map to revenue.

For the full 2026 benchmark set, see the cold email benchmarks guide. For an outbound program with reply-rate-first measurement built in, build a campaign and we'll track the metrics that actually matter.