Why I Stopped Chasing Open Rates

The number I used to lead with

I'm Hlib Storchak. I build outbound systems for B2B founders and sales teams, and I've booked 2000+ meetings for B2B clients doing it. For years, the first line of every weekly report I sent a client was the open rate. It was the easiest number to produce, it moved week to week so it felt like progress, and clients liked seeing a percentage go up. I kept doing it long after I half suspected the number was lying to me, because nobody was asking me to stop.

What finally stopped me was not a hunch. It was going back through a client's numbers after a subject-line test I was sure had worked, and finding that the "lift" lined up almost exactly with a jump in the share of that list reading on an iPhone, not with anything I had changed in the copy.

What open rate actually measures in 2026

Open rate was never a perfect metric, but it used to mean roughly what it said: a tracking pixel loaded because a person opened the email. That stopped being reliably true once Apple started pre-fetching images for privacy reasons, and the scale of that shift is bigger than most people running a cold email program realize.

Litmus's Email Analytics data, calculated from over 1 billion tracked opens in May 2026, puts Apple at 64.66% of all email opens, with Gmail a distant second at 24.11%. Apple's Mail Privacy Protection sits inside that Apple share, and it works by routing images, including the invisible tracking pixel every sending tool relies on, through Apple's own proxy servers before the message ever reaches the recipient. The pixel can fire, and your dashboard logs an "open," without a person having looked at anything.

The benchmark that quietly stopped reporting it

The clearest evidence that the industry already knows this is what the benchmarks themselves choose to publish. Instantly's 2026 Cold Email Benchmark Report, built from a large volume of real send data and one of the most cited benchmarks a cold email operator will run into, centers entirely on reply rate: an average of 3.43%, elite campaigns clearing 10.7%+, first-touch replies accounting for 58% of the total, and follow-up sends contributing the remaining 42%. It does not lead with, or even report, an open rate figure. A benchmark report built by a company whose product literally displays an open rate column chose not to make that column the headline number. That is a stronger signal than any single opinion piece, mine included.

Doing the math on a 44% open rate

Say a campaign reports a 44% open rate, a figure in the range several 2026 cold email benchmarks cite as average. Apply Litmus's 64.66% Apple share to that list, and a large share of the "opens" behind that 44% never had a human decision behind them at all: MPP fired the pixel the moment the message hit an inbox that supports it, often before it was ever displayed. I am not going to hand you a precise "real" percentage, because Litmus's data does not tell you what share of MPP-eligible opens are pixel-only versus genuinely read later, and neither does any other public source I have found. What the data does tell you, reliably, is the direction: the more Apple-heavy your list, the less a reported open rate has to do with actual attention, and a 44% headline number could be hiding anywhere from a healthy list to one nobody is reading at all.

Open rate vs the metrics that actually predict pipeline

Here is the same argument as a straight comparison, the version I actually put in front of a founder who wants one slide.

Notice that reply rate, positive reply rate, and meetings booked all require a person on the other end to do something. Open rate is the only row in that table that a proxy server can produce on its own.

The one job open rate still does

I have not deleted the open rate column from my dashboards, I demoted it. It still does one job well: catching a sending setup that is completely broken. If a domain that used to sit at a normal open rate suddenly falls to near zero across a whole campaign, that is rarely a sign that your subject lines got worse overnight. It is far more often a sign that the domain landed in spam, a DNS record broke, or a mailbox got suspended, and reply rate will not tell you that fast enough because there was nothing to reply to in the first place.

Used that way, open rate is a trip wire for deliverability disasters, not a performance number. The distinction matters: a trip wire only needs to tell you something is badly wrong, it does not need to be precise, and it does not need to move week to week in a way that means anything about copy or targeting.

What I report to clients instead

The mistake I see most often when I take over an account is a reporting habit built entirely around whatever a sending tool shows by default, which is usually open rate front and center because it is the easiest number to make go up with a subject-line tweak. On every account I run now, the weekly number is reply rate first, positive reply rate second, and meetings booked third, with open rate kept in a footnote purely as the deliverability trip wire above. Clients push back on this less than I expected, once they see one side-by-side week where the open rate held steady and the reply rate quietly dropped. That gap is the whole argument.

The cost of a week spent defending a vanity metric

Here is a rough model, with every input labeled as an assumption you should replace with your own numbers. Assume a weekly reporting review with a founder and a GTM lead runs 45 minutes, and assume both people bill their time internally at a blended €70 to €90 an hour. Assume roughly a third of that meeting, historically, has gone to explaining why an open rate moved, rather than deciding what to change in targeting or copy. That is somewhere in the range of €35 to €65 a week spent narrating a number that a proxy server can move on its own, before you count the larger cost: a team that spends its attention defending or celebrating open rate is not spending that same attention on the ICP, list, and message decisions that actually move reply rate. Run the same math with your own meeting length and your own blended rate. The exact euro figure matters less than the fact that the time is going somewhere, and right now, for a lot of teams, it is going to a number that a spam filter can inflate for free.

How to make the switch without a fight

If your team or your client is attached to seeing open rate move, here is the order I actually use to make the switch stick, rather than triggering a debate about whether open rate "still matters":

1. Keep open rate on the dashboard, just smaller. Removing it outright reads as hiding something. Shrinking it to a footnote next to the real numbers does the same job without the fight.
2. Pick one definition of "positive reply" and write it down. Before you can report positive reply rate as a headline number, everyone reading the report needs to agree on what counts, so it does not drift person to person.
3. Show one week where the two numbers disagreed. The argument for reply rate over open rate is more convincing as a single real example from your own list than as a general claim, because it turns an abstract argument into something the reader already half suspected.
4. Set the open rate trip wire threshold explicitly. Decide in advance what drop counts as "something is broken" so the metric still has a job, instead of just disappearing from the conversation.
5. Re-baseline reply rate expectations once, then leave them. A team moving off open rate for the first time often has no feel for what a normal reply rate week looks like. Give it four to six weeks of real data before treating any single week's number as a verdict.

The stack that makes this easier

This is the stack I run for clients: Salesforge for sending and sequencing defaults its own reporting toward reply rate rather than leading with opens, which is one less habit I have to fight when I take over a new account. That is a personal preference from running it day to day, not a claim that it is the only tool that gets this right. If your current sending tool still leads with open rate on the main dashboard, you can get the same effect by simply choosing which number you look at first, the tool does not have to change for the habit to change.

The pushback I still get

The most common objection is that open rate still tells you something about subject lines, so dropping it loses a useful signal. It is true that a catastrophic subject line can suppress opens, but a proxy-inflated, MPP-heavy open rate is a noisy instrument for testing that, since the noise floor moves with your list's device mix, not your copy. A/B testing subject lines against reply rate directly, even though it takes a larger sample to reach significance, tests the thing you actually care about instead of a number a server can move without your prospect doing anything.

The second objection is that reply rate is a lagging indicator compared to open rate, so teams lose an early read on a campaign. That is fair, which is exactly why the trip wire use case above still exists: a collapse in open rate is still worth checking same day, it is only the week-to-week performance narrative that needs to move to reply rate.

FAQ