Do Personalized Images Improve Open Rates?

No. They cannot.

This is worth stating plainly because a great many vendors in this category claim otherwise, and until recently some pages on this site did too. We have removed them.

The reason is mechanical rather than debatable: an email image is fetched when the message is opened. Nothing inside the email exists to the recipient until they open it. A personalized image cannot influence the decision to open, because at the moment that decision is made, the image has not been requested.

What determines opens is the sender name, the subject line, the preview text and whether the message reached the inbox at all. That is the entire list.

So what do personalized images affect?

Everything after the open.

Click-through rate. This is the metric personalization is aimed at. Once someone is reading, a graphic carrying their name, their company, or the product they were looking at gives them a reason to act that a generic banner does not.

Time spent and attention. Harder to measure, but a personalized visual interrupts skimming in a way stock imagery does not.

Conversion. Downstream of clicks, and the number that actually matters.

Reply rate, in cold outreach specifically, where the image is evidence of effort.

If a vendor tells you their images will lift your open rate, they either do not understand how email rendering works or they are hoping you do not. Neither is a good sign.

The larger problem: open rate is broken anyway

Even if personalized images could affect opens, you would struggle to measure it, because open rate stopped being a reliable metric in 2021.

Apple Mail Privacy Protection prefetches remote images on delivery, through Apple's proxy servers, whether or not a human ever opens the message. The tracking pixel fires regardless. Prefetching can happen hours after delivery, while the device sits on Wi-Fi with Mail running in the background - entirely disconnected from any actual person reading anything.

The effect is not marginal. Reporting after the rollout showed average open rates rising by several percentage points with no change in real engagement, and for lists with heavy Apple Mail usage, open rates can read 15-40% higher than reality.

Five years on this has not been solved, and iOS 18 added further complications - Link Tracking Protection stripping UTM parameters, AI-generated previews, and inbox categorization all affecting what you can measure.

The practical consequence: open rate is now a directional signal at best. Comparing your open rate to an industry benchmark is close to meaningless, because you do not know what proportion of either number is Apple's servers.

What this means for image-based metrics - including ours

Here is the part most vendors will not tell you.

If Apple prefetches the images in an email, the image server sees a request. That request looks exactly like a human opening the email, because from the server's point of view it is identical.

So any metric based on image loads - including the open counts OKZest reports per design - includes prefetches that no person ever saw. We would rather say that plainly than let you build a case on numbers that are softer than they look.

Treat image load counts as an upper bound on human attention, not a measurement of it. They are useful for spotting relative differences between campaigns, and unreliable as an absolute.

What to measure instead

Click-through rate, calculated on delivered rather than opened. Clicks require a human, and using delivered as the denominator sidesteps the inflated open number entirely. This is the single most useful change most teams can make to their reporting.

Conversion and revenue per send. The end of the funnel is the least corrupted part of it and the part your business cares about.

Reply rate, for outreach.

Unsubscribe and complaint rate, as a health check. Rising unsubscribes alongside rising opens is a strong sign the opens are not real.

Your own trend over time, rather than industry benchmarks. Your Apple Mail proportion is roughly stable, so period-on-period comparisons within your own data remain meaningful even when the absolute number does not.

If you must report open rate to someone, report it with the caveat attached. "Opens are inflated by Apple's prefetching and are not comparable to pre-2021 figures" is a sentence worth having in your reporting template.

Frequently asked questions

Do personalized images improve email open rates? No. The image is fetched when the email is opened, so it cannot influence whether someone opens it. Opens are determined by sender name, subject line, preview text and deliverability. Personalized images affect what happens after the open - primarily click-through.

Why do so many tools claim they improve open rates? Usually loose writing rather than deliberate deception - "engagement" gets used as a catch-all and open rate gets swept in. But it is wrong, and it is worth treating as a signal about how carefully a vendor makes its other claims.

What do personalized images actually improve? Click-through rate mainly, and conversion downstream of it. In cold outreach, reply rate. All of these are post-open metrics.

Is my open rate accurate? Probably not. Apple Mail Privacy Protection prefetches images on delivery whether or not anyone opens the message, so any recipient using Apple Mail may register an open they never made. Depending on your audience mix, your open rate can read 15-40% higher than reality.

Does Apple MPP affect click tracking too? Clicks require a human, so click data remains far more reliable. iOS 18's Link Tracking Protection does strip some URL parameters, which affects attribution rather than whether the click is recorded.

Should I stop reporting open rate? Not necessarily, but stop treating it as a performance measure and stop comparing it to benchmarks. Use it as a directional signal within your own data over time, with a note explaining the inflation.

What is the best single metric for email? Click-through rate calculated on delivered rather than opened, then conversion. Using delivered as the denominator avoids the inflated open figure entirely.

Do OKZest's image open counts include Apple prefetches? An image load looks the same to a server whether a person triggered it or a proxy did, so counts based on image loads should be read as an upper bound rather than a measurement of human attention. Useful for comparing campaigns, unreliable as an absolute number.

Measure the thing that moved

If you take one change from this: calculate click-through on delivered rather than opened, and compare against your own history rather than an industry figure. That gives you a number that means something, and it makes it possible to tell whether a change - personalized images or anything else - actually did anything.

OKZest generates personalized images from one template, rendered per recipient when they open the email. They will not improve your open rate. They give people who have already opened a reason to click.

Create a free account, no credit card required.

Sources

Related reading: do personalized images actually improve click-through rates, improving click-through on a budget, personalized images explained.