How does proxy caching distort open rates?
Still have a question, spotted an error, or have a better explanation or a source we should cite?
When Gmail routes images through its proxy, it doesn't fetch them fresh every time. It caches them. So when a subscriber opens your email multiple times, the second and third opens might not generate new pixel requests at all. The cached image loads from Google's servers directly without re-contacting your ESP. Your open count for that subscriber stays at one.
This means proxy caching tends to undercount opens for Gmail users. You might see a subscriber as a single-open when they actually read your email three times. Total open counts (as opposed to unique opens) are particularly affected, since caching absorbs the re-reads.
Apple's proxy does the opposite. It pre-fetches images before the user opens, which inflates open rates. Apple doesn't cache in the same way, or rather the pre-fetch behavior creates artificial opens where Gmail's caching suppresses real ones.
The combined effect on your open data is an audience-dependent distortion: subscribers using Gmail tend to have undercounted re-opens, while Apple Mail users have overcounted opens. If your audience is mixed, these effects can partially cancel out in aggregate numbers, making the data look cleaner than it actually is.
The more reliable signal is still click rate. Clicks bypass proxy caching because they require navigating to a new URL via a redirect through your tracking domain, and that redirect is harder to cache reliably.
Contributors
Who worked on this answer
Every name links to their profile. Every company links to their site. Real people, real accountability.
