Two moves, in order. First, stop trusting the metric: open rates have been unreliable for years, inflated by privacy proxies that auto-load pixels and deflated by filtering, so the number moves for reasons that have nothing to do with readers. Read replies, clicks, and unsubscribes instead; they are behaviors, not artifacts.
Then fix what actually decays: list health and email worth. Inboxes are flooded with AI-written sameness, which means the bar for being opened rose while most senders' distinctiveness fell. The durable answer is not subject-line tricks. It is becoming one of the few emails a reader would notice missing, which is a substance decision, made weekly, in your own voice.
- The metric is broken before the list is: privacy proxies inflate opens and filters deflate them, so the number moves without readers changing.
- Replies, clicks, and unsubscribes are the real instruments: behaviors survive measurement changes that artifacts do not.
- The inbox got flooded with sameness: AI-written email volume raised the bar for attention while lowering the average email's distinctiveness.
- List decay is normal and fixable: pruning cold subscribers improves every real metric and your deliverability with it.
- The durable fix is being missable: emails with a voice and a position get chosen, and chosen is the only open that matters.
Join the Collective Wisdom™ Newsletter
Written by us from real client builds and our own operation, not recycled AI news. What worked, what didn’t, and what to do about it. Read in a few minutes, useful the same day.
Join the newsletterReady to talk? Book a Rapid Transformation Call.
Why are open rates an unreliable metric now?
Because the open was never a behavior; it was a tracking pixel loading, and the machinery around that pixel changed under everyone's feet. Privacy features in major email clients now pre-fetch images on the server, registering opens nobody performed, while other filters strip or delay the pixel, hiding opens that happened. The same list, the same emails, and the reported number can drift either direction as client market share shifts.
What this means practically:
- Trends are noise until proven otherwise. A drifting open rate reflects your subscribers' mix of email clients as much as their interest.
- Comparisons across time mislead: your 2021 baseline was measured with a different instrument than today's number.
- Decisions made on opens compound the error: resending to 'non-openers' who actually read you, pruning 'cold' subscribers who open invisibly.
The replacement instruments are behaviors: replies, clicks, unsubscribes, and forwarding, each an action a human chose. They run smaller and truer, and every one of them survives whatever the next privacy change does to pixels. Calibrate on those, and the panic about the drifting open number usually resolves into a much more specific, fixable picture.
What changed in the inbox itself?
Volume and sameness, together. AI collapsed the cost of producing email, so everyone produces more of it, and most of it draws from the same generative well: competent, warm-ish, position-free prose that readers have learned to recognize and skim past. Harvard Business Review documents the broader 'workslop' flood, and research in Science Advances measured the underlying convergence, AI-assisted writing drifting toward mutual sameness.
The reader's adaptation is rational triage:
- The noticed tier: a handful of senders whose emails get opened on name recognition alone, because the reader would miss them.
- The someday tier: content that seemed worth keeping, archived unread, forever.
- The reflex tier: everything with template energy, deleted or filtered on sight, including much of what the sender's dashboard counts as delivered.
Every email you send is competing for tier one, and the competition is not other experts in your field; it is the reader's finite willingness to choose. The senders holding or growing engagement in this environment share one property: their emails are recognizably from someone, carrying a voice and a stance the reader could not get from the flood. Distinctiveness became the deliverability that matters.
How do I fix the list itself?
By treating decay as maintenance rather than tragedy. Lists rot at a steady rate, addresses churn, interests move on, and a list never pruned accumulates dead weight that damages the living part:
- Prune on behavior, not opens. Subscribers with no clicks, replies, or site visits across several months are gone regardless of what the pixel claims. A goodbye sequence, one honest 'still want these?' email, then removal, converts a vanity number into a working asset.
- Expect the metrics to jump: every real rate improves when the denominator stops including the departed, and mailbox providers read the higher engagement as sender quality, which lifts inbox placement for everyone remaining.
- Fix the intake while you are there: subscribers arriving from a clear promise, what you write about, how often, in whose voice, stay engaged; subscribers harvested by generic magnets churn immediately. The list's health is set at the front door.
- Own the channel's purpose: the list is the audience you can reach without an algorithm's permission, which makes it the most valuable marketing asset you hold, worth maintaining like one.
A smaller list that reads you beats a large one that has forgotten who you are, on every metric that pays.
What makes an email one that readers would miss?
Three properties, none of them subject-line engineering:
- A voice they recognize. Your emails should be identifiable with the sender line covered: your phrases, your directness, your way into a topic. In an inbox of generative sameness, a real register is the strongest open-trigger there is, because it promises the one thing the flood cannot: you.
- A position per email: something believed, argued, or called out, not a roundup of the agreeable. Readers keep senders who help them think, and forward senders who say what others hedge. One honest stance outperforms five curated links, every time.
- Usefulness at the reader's altitude: the judgment layer, what to do, what to skip, what actually happened when, rather than the informational layer they can generate themselves. Your email competes with free AI answers now; the winning material is what those answers cannot contain.
The production discipline that sustains all three: derive emails from your real work, this week's client pattern, the call you made, the thing you corrected, rather than composing 'content' from nothing. The substance was already earned during the week. The email just delivers it, in your voice, to people who chose to hear from you.
Should I change how often I send?
Only after the substance question is settled, because frequency is a multiplier on whatever the emails are. Cutting the cadence of unmissed emails just slows the decay; raising the cadence of missed ones would grow the list's engagement instead.
The honest sequencing:
- Fix worth first. Voice, position, usefulness, until replies tell you the emails land. Frequency debates before that are rearranging deck chairs.
- Then send at the rhythm you can sustain with substance: for most expert businesses, weekly is the sweet spot, frequent enough to stay a habit in the reader's week, spaced enough that each email has something real inside. A rhythm you can hold through busy seasons beats an ambitious one you abandon visibly.
- Keep the rhythm honest with derivation: emails drawn from your actual work and captured material take an hour, not an afternoon, which is what makes weekly sustainable without the quality sagging.
- Let consistency itself compound: readers calibrate to your rhythm, and reliability becomes part of the relationship, the email that always comes carries more weight than the one that sometimes does.
Watching this whole discipline practiced in public, voice, position, weekly rhythm, is part of what the Collective Wisdom newsletter is, which is also the fastest way to see whether the approach earns its opens.
The PLB Perspective
The open-rate panic usually walks in carrying the wrong villain: the owner suspects her subject lines, her send times, her platform, when the honest culprits are a broken instrument and an inbox that changed species. You cannot optimize your way back to 2019 opens, because 2019's inbox and 2019's measurement no longer exist. The senders thriving now never got the memo to panic, because their emails were always the kind readers choose, and choosing survived every change that broke the metrics.
Here is the audit I actually run on a struggling newsletter, and it takes five minutes: read the last six sends and ask what a subscriber would have lost by not receiving them. If the honest answer is nothing, links they could find, encouragement they did not request, updates about the sender, then the open rate is not the problem, it is the verdict. The fix is never mechanical. It is deciding, each week, to send something with a spine in it.
And I want to defend email itself while its metrics are being mourned: the owned list remains the single most valuable channel an expert holds, precisely because everything else got worse faster. Feeds throttle, engines intermediate, platforms reprice, and the list still delivers your words to people who asked for them, at a cost of nothing, on infrastructure nobody can revoke. The channel is not dying. It is consolidating, toward the senders worth choosing, and that consolidation is an opportunity wearing a scary dashboard.
Treat any published benchmark skeptically, because privacy proxies inflate and filters deflate the number differently per audience. The working approach: baseline your own list's replies, clicks, and unsubscribes over a quarter, and measure against yourself. A list where a meaningful fraction clicks and a few reply per send is healthy at almost any reported open rate; one with silence behind a decent open number is the one in trouble.
Prune on total behavior rather than opens alone, since privacy features hide real readers: no clicks, no replies, and no site visits across several months marks a subscriber as genuinely gone. Send one honest re-permission email, then remove the silent. Every real metric improves, deliverability rises with the engagement rate, and the list becomes what it is for: people who actually want the words.
Blank-prompt AI email performs worse every month, because readers have learned its register and the inbox is saturated with it. Emails drafted by AI from your documented voice, positions, and real material perform like yours, because substantively they are. The variable was never the drafting tool; it is whether the email carries anything the reader's own AI could not have generated, which is a supply question about you.
More than ever, for exactly the reasons the environment feels hostile: the flood is consolidating reader attention toward the few senders worth choosing, and an owned list is the one channel no algorithm meters. Start with the promise clear, a voice that is recognizably yours, and a weekly rhythm derived from your real work. The senders who begin now, with substance, inherit the attention the template era is forfeiting.
Mostly the environment, not you: LinkedIn shows posts to fewer people, the feed is flooded with AI-generated content, and a growing share of buyer attention left feeds for AI answers entirely.
Because effort is flowing into channels that expired while the buyers moved somewhere your marketing doesn't reach: private research inside AI answers. More volume into a drained pond catches fewer fish, at higher cost.
Consistency was a scarcity strategy, and the scarcity ended: when AI made volume free for everyone, posting frequency stopped signaling anything. The advice survives its own obsolescence because it keeps its victims too busy to notice.