Building Confidence Scoring for Email Open Tracking (Engineering Notes)

Nate S — Mon, 15 Jun 2026 19:46:38 +0000

Most email open tracking in 2026 is broken. Apple Mail Privacy
Protection fires a fake open within seconds of delivery, before any
human sees the email. Corporate scanners do the same. Open rates run
2-3x inflated.

The engineering problem: how do you tell a real human open from a
machine pre-fetch given only the HTTP request metadata of the pixel
load?

*Signals available
*
Every open event arrives at the tracking endpoint with:

Request IP
User-Agent string
Request timestamp (relative to email send)
Accept-Language, Referer, other headers

Patterns by source
**
**Apple MPP pre-fetches:

IP from Apple-attributable ranges (17.0.0.0/8 mostly)
User-Agent: Mac/iOS native with Apple's tracking-relay format
Timing: typically 30 seconds to 5 minutes after delivery
No subsequent click activity

Corporate scanner pre-fetches (Defender, Mimecast, Proofpoint):

IP from known scanner ranges (each vendor publishes these or they are discoverable via reverse lookup)
User-Agent: scanner-specific signatures
Timing: sub-5-second from delivery
Multiple link and image requests within 1-3 second window from same IP

Gmail image proxy:

IP from googleusercontent.com range
User-Agent: Google bot signature
Timing: variable

Real human opens:

IP from residential or generic corporate range
User-Agent: actual mail client used by a human
Timing: rarely sub-30-second from delivery; clusters at typical inbox-check times
Often followed by click activity within 30 minutes

*Model approach
*
A gradient-boosted classifier on the feature set above gives 95-98%
agreement with human-rated labels on a held-out test set. Output is a
confidence score (0-100%) which maps to Tier 1-5.

False-positive rate (Tier 1 graded when it's a machine): <2%
False-negative rate (Tier 4-5 graded when it's a human): <5%

The productionized version of the model + dashboard surface is at:
https://outsolvi.com/features/confidence-scoring

*The retraining problem
*
Apple keeps shifting MPP's IP block allocation. The model needs
retraining every few months as patterns drift. We've automated the
labeled-data collection so retraining is a 1-day job rather than a
1-week job.

Anyone else working on email signal filtering? Curious about your
approach to the drift problem specifically.

— Nate Summers
Co-Founder, Outsolvi

The Three Email Metrics That Stopped Working in 2026

Nate S — Mon, 15 Jun 2026 19:41:46 +0000

There are three email metrics that almost every B2B sales team still
tracks and almost every B2B sales team should stop trusting in the way
they currently do. The metrics aren't broken; what's broken is how
they're measured.

*1. Open rate
*
The textbook definition: share of delivered emails where the recipient
opened the message. In 2025 and earlier, this was approximately what
was being measured. In 2026 it isn't.

Apple Mail Privacy Protection (MPP), launched with iOS 15 in September
2021, pre-loads every remote image (including tracking pixels) on
Apple-controlled relay servers before the recipient opens the email.
With Apple Mail at roughly 58% global email-client market share
(Litmus), MPP affects half or more of every B2B list.

Corporate email scanners (Microsoft Defender for Office 365, Mimecast,
Proofpoint) pre-fetch every link and image on every email for malware
scanning. Every enterprise B2B prospect generates a fake open within
seconds of delivery.

Gmail's image proxy can register the same pixel as multiple opens.
Smaller problem, still real.

The result: a reported 70% open rate typically maps to a 25-35% actual
human-read rate.

The fix is confidence scoring: grade each open from Tier 1
(high-confidence human) to Tier 5 (bot/scanner) based on IP block,
User-Agent, and timing signature. Filter below a 25% confidence
threshold.

*2. Click-through rate
*
The same problem reaches clicks. Microsoft Safe Links, Mimecast URL
Protect, and Proofpoint TAP pre-fetch every link on every email for
malware scanning. Every link gets clicked automatically by the scanner
within seconds of delivery.

Unfiltered click-through rate in 2026 is contaminated 2-3x by scanner
auto-clicks. The fix is unique-clicks-per-recipient with scanner-IP
filtering and a time-window filter (any click within 5 seconds of
delivery is almost always a scanner, not a human).

*3. Time-to-first-open
*
When the open is a machine pre-fetch firing 30 seconds after delivery,
the average time-to-first-open metric becomes meaningless. The measure
that actually matters is time-to-first-Tier-1-open: when did a real
human first engage with the email?

The structural lesson: 2026 email tracking only works if it filters
MPP, scanner pre-fetches, and proxy duplicates. Tools that count raw
pixel loads are reporting noise, not engagement.

I wrote up the technical breakdown of how MPP actually works here:
https://outsolvi.com/glossary/apple-mail-privacy-protection

Nate Summers Co-Founder, Outsolvi

DEV Community: Nate S

Building Confidence Scoring for Email Open Tracking (Engineering Notes)

The Three Email Metrics That Stopped Working in 2026