Lyra

Lyra joined her current ship as Bosun when their engagement rate had cratered to 12%. The captain was confused. The list was clean. The content was good. Authentication was solid. And yet only 12% of the subscribers were doing anything with the emails.

She ran the numbers and found the culprit immediately. The list had been growing for years, but the engagement-tier breakdown was upside-down. The Unicorns (the top-engagement segment) were 7% of the list. The Engaged were 19%. The Stale were 31%. The Dormant were 43%. Forty-three percent of the list had been completely silent for over six months. Sending to those subscribers was suppressing the deliverability of the engaged ones, and the captain hadn't noticed because the list size was still growing.

She did three things in sequence.

First, she ran a single re-engagement attempt on the Stale segment. One email. A clear ask. Anyone who clicked moved back to Engaged. Anyone who didn't moved to Dormant.

Second, she suppressed the Dormant segment in full. The captain balked at "losing" forty-three percent of the list, the way captains always do. Lyra explained that the forty-three percent had stopped being subscribers a long time ago, the system had just been pretending they weren't gone.

Third, she set up a monthly engagement audit so this never happened again.

After three months, the engagement rate had climbed from 12% to 24%. After eighteen months, it was 34%. The list was smaller in absolute terms, but the active list was three times more engaged than the old version. Revenue per send doubled. Domain reputation recovered fully.

The lesson she teaches now: you can rebuild engagement, but only by being honest about who's still actually rowing. The captains who refuse to suppress lose more in the long run than they would have lost by suppressing.

---

A small B2B sender insisted that opens were the metric that mattered. They had a 28% open rate and were proud of it. Lyra told them that since Apple's Mail Privacy Protection, open rates had become unreliable. The sender disagreed, citing their numbers. They wanted to keep optimizing for opens.

Lyra proposed a test. The next campaign would include a single direct ask in the body: "Reply with yes or no, do you find this newsletter useful?" Track the replies. See what they correlated with.

The campaign went out. The open rate was 28%, in line with the usual. The reply rate was 1.2%. That meant only 1.2% of the list was paying enough attention to actually read and respond to a direct question. The rest of the "opens" were either MPP auto-opens or distracted scrolls.

The follow-up was the interesting part. The sender's sales team tracked which subscribers had replied, then matched those subscribers against converted leads over the next 90 days. Reply-rate subscribers converted at 11% over the period. Open-rate subscribers (those who'd been marked as "opened" by MPP but hadn't replied) converted at 0.8%. The reply was a 13x stronger signal of actual interest than the open.

The sender shifted their primary engagement metric from open rate to reply rate within the month. They built a lead-scoring system around it. They started writing subject lines that asked questions, and content that invited reply. Their conversion-attributed-to-email metric tripled within six months.

The lesson she teaches now: opens are noise. Clicks are signal. Replies are gold. Track what actually predicts revenue, not what's easy to measure.

Lyra's intro:

Pulse this. Two senders, same list size, same content quality. Sender A has a 28% open rate and feels good about it. Sender B has a 12% open rate and is panicking. Both senders are wrong about what's actually happening on their lists. Engagement isn't an open rate. Engagement is a heartbeat. Here's how to read it.

---

When marketers talk about engagement, they usually mean opens. Sometimes they mean clicks. Occasionally they mean conversions. None of those alone is engagement. Engagement is the pattern of attention your subscribers are paying to your sends, and that pattern shows up across multiple signals at once.

The reason this matters is that any single metric can be gamed, broken, or misread. Open rates have been broken since 2021 by Apple's Mail Privacy Protection. Click rates can be inflated by bot-clicking link scanners. Conversion rates depend on offers that change campaign to campaign. No single metric is reliable on its own. Engagement is what you see when you look at all of them together over time.

There are four signals that, together, give you a real engagement read.

Signal 1: Click rates

Did the subscriber actually do something with the email beyond have it sit in their inbox? Click rates measure that. They're harder to game than opens (a bot can mark an email as opened, but it's harder to fake a meaningful click). They correlate strongly with actual interest.

The healthy click rate range varies by industry, but for most senders, 1-5% on a regular newsletter is normal. Below 1% suggests content that's not landing or a list that's lost interest. Above 5% suggests strong content-to-audience fit.

The metric to actually watch is click-to-open rate (CTOR), which is clicks divided by opens. CTOR tells you what percentage of the people who actually saw the email did something with it. A 28% open rate with a 5% CTOR is healthy. A 28% open rate with a 1% CTOR means most of the "opens" were either MPP auto-opens or distracted scrolls that never resulted in a real read.

Signal 2: Reply rates

Did the subscriber take the time to write back? Reply rates are the gold standard of engagement, and almost nobody tracks them.

Replies are hard to fake. They take real attention. They take real intent. A subscriber who replies has done significantly more work than a subscriber who clicked, who has done more work than a subscriber who "opened." The signal is correspondingly stronger.

Healthy reply rates depend on whether you're inviting replies. For a regular newsletter that doesn't ask for replies, even 0.1% is meaningful. For a B2B sender who actively invites replies in the body, 1-3% is normal and 5%+ is exceptional.

If you've never tracked reply rates, the simplest way to start is to send from a real, monitored mailbox (not a no-reply address) and count the replies. You'll learn more about your list from one month of reply data than from a year of open data.

Signal 3: Conversion rates

Did the subscriber do the thing the email was trying to get them to do? Sign up. Buy. Book. Download. Conversion rates measure the bottom of the funnel.

The complication is that conversions depend on the offer, the landing page, the purchase friction, and a dozen other things outside the email itself. A clean A/B test on subject lines might show one variant winning on opens but losing on conversions. The "winning" subject line could be the one that attracts non-buyers.

Conversion rates are the final reality check. They tell you whether the engagement you're seeing actually translates into the outcomes the business cares about.

Signal 4: Reverse signals (negative engagement)

Engagement isn't only positive. Unsubscribes, complaints, and silent drift are also engagement signals. They tell you about subscribers who've decided you're not worth their inbox space.

A healthy unsubscribe rate is 0.2-0.5% per send. Below 0.2% might suggest the unsubscribe link is hard to find (a problem). Above 0.5% suggests content or cadence that's pushing people away.

Complaint rates should stay below 0.1% per send. Above 0.3% is a deliverability problem in the making. Above 0.5% is an active fire.

Silent drift is the one most senders miss. It's the subscriber who used to engage and stopped, but didn't unsubscribe. They're still on the list, still being sent to, still being counted as a "subscriber," but they've effectively left. The engagement signal is their absence.

How to actually read engagement

Engagement isn't a number. It's a story told by the four signals above, watched over time.

The story you want to see:

- Click rates stable or rising over weeks - CTOR consistent across campaigns - Reply rates measurable and growing for senders who invite reply - Conversion rates correlating with click and reply rates - Unsubscribes and complaints stable and low - Silent drift rate (% of subscribers who haven't engaged in 60+ days) stable or shrinking

The story you want to avoid:

- Click rates falling slowly week over week - CTOR falling even when opens stay flat - Reply rates dropping or going to zero - Unsubscribes climbing - Silent drift rate growing

When the story turns negative, something on the list has shifted. Could be content fatigue, could be cadence drift, could be audience saturation, could be that the wrong segment is getting the wrong sends. The signal tells you to investigate. The investigation finds the cause.

What engagement isn't

Engagement isn't a vanity metric to brag about. Senders who post their open rates on LinkedIn are bragging about a number that's largely fictional since 2021. Engagement isn't what the dashboard shows by default. It's what the dashboard shows when you've calibrated it to the metrics that actually predict revenue and retention.

Engagement isn't fixed. Subscribers move between engagement tiers constantly. A Unicorn subscriber can drift to Stale in a quarter if your content stops landing. A Stale subscriber can return to Engaged after a single great send. The pattern is dynamic.

Engagement isn't the whole story. A perfectly engaged list of 500 subscribers might generate less revenue than a moderately engaged list of 50,000. Engagement quality matters, but so does scale. The job is to maximize quality at the scale you've got.

Where to start

If you've never measured engagement beyond opens, start here.

One, set up CTOR tracking. Most ESPs will calculate it for you if you ask. If yours doesn't, calculate it manually: clicks divided by opens, expressed as a percentage. Track it per campaign.

Two, send from a real mailbox and count the replies for one month. Don't change anything else. Just count. You'll learn what your reply rate baseline actually is.

Three, define your engagement tiers (Unicorns, Engaged, Stale, Dormant) using both opens and clicks combined, not opens alone. Anyone who's only "opened" without ever clicking is a low-tier engager. Anyone who's clicked is a higher-tier engager regardless of open count.

Four, watch all four signals (click rates, replies, conversions, reverse signals) for one full quarter. Don't change strategy. Just calibrate.

By the end of the quarter, you'll know what your real engagement looks like, not the open-rate fiction. From there, every decision you make about content, cadence, and segmentation has a real signal to optimize against.

The heartbeat is there. You just have to listen.

- Lyra

Lyra's intro:

I had a sender tell me last month that their open rate was 41% and they wanted to know what subject line I'd write to push it to 45%. I asked them what their click-to-open rate was. They didn't know. I told them their open rate was a story Apple was telling them, and the click-to-open rate was the actual book. They didn't believe me. Two weeks later they figured it out themselves. Here's the version I'd written down.

---

In 2021, Apple released Mail Privacy Protection (MPP), a feature in iOS 15 and macOS Monterey that fundamentally changed what an "open" means in email marketing.

Before MPP, an open was tracked by a tiny invisible image (a tracking pixel) embedded in the email. When the recipient's mail client loaded the image, it pinged the sender's analytics and registered as an open. The system wasn't perfect (some clients blocked images by default, some users had image-loading off), but it was directionally accurate. An open meant a recipient had probably looked at the email.

After MPP, Apple Mail started pre-loading every email's images on a privacy proxy server. This happens automatically, regardless of whether the user actually opened the email. The tracking pixel pings as if the email were opened, but the user might never have seen it. They might have deleted it without looking. They might have set their phone down and walked away. The "open" registers anyway.

This single change broke open rates as a reliable metric.

How much of your open rate is real?

The exact percentage depends on your audience, but here's the rough math.

Apple Mail accounts for roughly 50-60% of consumer email reading globally. That's iPhone Mail app users plus macOS Mail users. Among those users, MPP is enabled by default and has very high adoption (90%+).

So if your audience is consumer-facing (newsletter, ecommerce, content creator), expect 50-60% of your reported "opens" to be MPP auto-opens with no actual reader behind them. Your real open rate is probably 40-50% of what your dashboard shows.

If your audience is B2B (working off corporate mail clients like Outlook or Gmail Enterprise), MPP impact is smaller, but it's still 20-40% of opens. Your real open rate is 60-80% of what your dashboard shows.

In practice: a sender showing a 30% open rate is probably seeing real human engagement on something like 12-18% of the list. The other 12-18% is Apple's privacy proxy. Both numbers count as "opens" in your dashboard, and you can't tell them apart from the open data alone.

Why senders kept treating open rate like it works

Three reasons.

Inertia. Open rate has been the headline metric in email marketing since the late 1990s. Every dashboard shows it prominently. Every blog post benchmarks it. Every report leads with it. Changing what you measure is harder than continuing to measure the wrong thing.

Plausible-looking numbers. MPP didn't make open rates obviously broken. It made them inflated in a way that looks normal. A 30% open rate looks healthy, even if it's technically half-fictional. The number doesn't trigger alarms.

ESPs didn't pivot fast. Most email service providers continued to feature open rate as the headline engagement metric, sometimes with small footnotes about MPP. The marketing world followed the dashboards.

The combination meant senders kept making decisions based on a metric that had silently lost most of its meaning. Subject-line testing optimized for inflated opens. Send-time testing optimized for inflated opens. Re-engagement decisions used inflated opens as the input. The downstream impact compounded.

What to look at instead

Five better metrics, in order of how reliable they are.

One: Click rate. Every click is a deliberate action by a real human (with rare exceptions for link-scanning security tools, which can be filtered out). A 2% click rate is harder to fake than a 30% open rate. Click rates are the new headline metric for most senders.

Two: Click-to-open rate (CTOR). Clicks divided by opens, expressed as a percentage. CTOR tells you what percentage of the people who saw the email actually engaged with it. Even if "opens" are inflated by MPP, CTOR remains directionally useful because both the numerator (clicks) and denominator (opens, including MPP) are still measured the same way. If your CTOR drops, something is wrong with content-to-promise alignment, regardless of what your raw open rate says.

Three: Reply rate. A reply is the strongest engagement signal there is. It can't be faked by a privacy proxy. It's hard to fake by automation in any meaningful way. Tracking reply rate requires sending from a real mailbox (not a no-reply address) and either monitoring the mailbox manually or using an ESP that tracks replies automatically.

Four: Conversion rate. The bottom of the funnel. Did the subscriber actually do the thing? Conversions cut through every upstream measurement noise. They're slower to measure (you have to wait for the conversion event) but they're the closest thing to ground truth.

Five: Unsubscribe and complaint rates. The reverse signals. They're not affected by MPP at all. They're the same metric they've always been. A rising unsubscribe rate or complaint rate tells you something is wrong, regardless of what the open numbers say.

How to retrofit your reporting

If you're stuck with dashboards that lead with open rate, here's how to retrofit your thinking.

Add CTOR to every campaign report. Most ESPs calculate it. If yours doesn't, the math is clicks/opens. Display it next to open rate, not buried somewhere.

Stop comparing your open rate to "industry benchmarks." The benchmarks are also affected by MPP. They're equally inflated. The comparison is meaningless.

Compare your campaigns against each other instead. Within your own data, the MPP inflation is consistent across campaigns. So if Campaign A had a 28% open rate and Campaign B had a 32% open rate, Campaign B was probably better. The relative comparison still works even if the absolute numbers are inflated.

Track real-world outcomes. Reply rate, conversion rate, revenue per send. These are the metrics that money actually moves through. The vanity metrics live separately.

For send-time testing, use clicks as the metric of interest, not opens. MPP auto-opens don't follow circadian rhythms. Real clicks do.

For subject-line testing, use clicks if you have enough volume to be statistically significant. If you don't, fall back to opens but be aware of the noise floor.

What this means for senders

Open rate isn't useless. It's just no longer reliable enough to be the headline metric. Senders who pretend otherwise are optimizing against noise.

Senders who pivot to clicks, CTOR, replies, and conversions get a clearer picture of what's actually working on their lists. They make better decisions. They write content that earns clicks instead of clickbait that inflates opens. They know which subscribers are real and which are Apple's proxy.

The pulse is still there. You just have to listen for the right signals. Open rate is the noise. Clicks, replies, and conversions are the heartbeat.

- Lyra

Lyra's intro:

The "we miss you!" sequence is one of the most common re-engagement patterns in email marketing, and it's also one of the worst. Five emails over three weeks, each one slightly more desperate than the last, ending in a discount nobody asked for. Open rates collapse with each step. Unsubscribes spike. Complaint rates climb. The whole sequence is a slow-motion crash. Here's the version that actually works.

---

A re-engagement campaign is what you send to subscribers who used to engage and stopped. They're not gone yet (they'd be in the Lapsed tier and you'd suppress them), but they're drifting (the Stale tier, in the engagement-tier system). The window where you can pull them back is real but narrow.

Most senders run re-engagement as a sequence of multiple emails. The instinct makes sense: if one email won't catch them, maybe the second or third will. The data says no. Multiple emails to disengaged subscribers do worse than one. They burn the relationship instead of saving it.

Here's why the single-email approach beats the sequence, and how to run it.

Why a single email outperforms a sequence

Three reasons.

One: every email after the first carries diminishing returns. The first re-engagement email reaches the subscriber when they're already in a "I haven't been opening these" state. If they don't open or click, they've made their preference clear. The second email is sent to someone who has just told you (through silence) that they don't want this content. The third email is sent to someone who has now told you twice. The signal you're sending with each subsequent send is "I'm not listening to your behavior."

Two: re-engagement sequences amplify complaint risk. Subscribers who didn't open the first email are more likely to mark the second or third email as spam. The complaint rate on re-engagement sequences runs 3-5x higher than on regular sends. This is exactly the wrong direction. The point of re-engagement is to clean up the list and protect deliverability, not to torch reputation by spamming people who've already told you they're done.

Three: the single email is more honest. A single, clear ask treats the subscriber as an adult who can make a decision. A multi-email sequence treats them as someone who needs to be convinced, manipulated, or worn down. The first version is the one that produces real subscribers. The second version produces resentful ones, who unsubscribe later anyway with worse data trails.

What the single re-engagement email looks like

The format that works:

- One email - Sent to subscribers who haven't engaged in 90+ days but haven't yet hit the 180-day Lapsed threshold - Subject line that's clear, not cute. "Are you still reading?" or "We're checking in" works. "WE MISS YOU!!! 🥺" does not. - Body that does three specific things

First, acknowledge the silence. "We've noticed you haven't been opening our emails in a while. That's totally fine, life is busy and inboxes are crowded."

Second, state the deal. "If you'd like to keep getting our emails, just click the button below. If we don't hear back, we'll cut your sends way down (or remove you entirely) so we're not cluttering your inbox."

Third, give a single clear action. A button or link that says something like "Yes, keep me on the list." That's it. No discount. No incentive. No "just one more chance."

The button doesn't have to do anything fancy on the back end. Clicking it can simply mark the subscriber as re-engaged, which moves them back to the Engaged tier in your engagement-tier system.

What about the subscribers who don't click?

After 14 days, the subscribers who didn't click move to one of two places.

Option A: they go to the Dormant tier and get suppressed. This is the cleaner option. The subscriber didn't respond to a direct ask. Their preference is clear. You stop sending. Their address goes on a do-not-send suppression list permanently.

Option B: they go to a low-cadence "lite" tier. This is for senders who want to maintain optionality. Subscribers in this tier get one email a quarter (or one a month at most), each one another low-pressure check-in. The cadence is so light that the deliverability damage from the inactive subscribers is minimized, but you maintain the option that one of them might re-engage someday.

Both options work. Option A is what I usually recommend because the cleaner suppression strategy protects deliverability better. Option B works for senders with a long sales cycle or rare-purchase audiences (real estate, B2B with long evaluation periods, etc.).

What does NOT work is option C: keep sending them at full cadence anyway. That's the option most senders default to because they don't want to "lose" the addresses, and it's the option that costs them deliverability over time.

The numbers you should expect

Response rates on a single re-engagement email:

- Strong list, mostly Tier A consent: 15-30% click-through rate - Average list, mixed consent: 8-15% click-through rate - Weak list, unclear consent or old data: 2-8% click-through rate

These numbers are smaller than what most senders expect, and that's actually the whole point. The 70-95% who don't respond have already left the relationship. The re-engagement email just makes that visible. Your active list shrinks as a result, but the active list is now exclusively people who actively chose to keep getting your emails. Engagement rates on subsequent sends climb dramatically.

What about a "win-back" offer?

Win-back offers (a discount or special incentive sent to lapsed buyers) work for some industries. Specifically, they work for transactional commerce where the subscriber has bought something before and might buy again.

They do not work for content-based subscribers (newsletters, blogs, content creators). A discount code in a re-engagement email confuses the relationship. The subscriber didn't sign up for a discount. They signed up for content. Offering them a discount to re-engage signals that you don't understand what they were originally there for.

For ecommerce specifically, a win-back can be folded into the single re-engagement email as a secondary call-to-action. "If you'd like to keep getting our emails, click here. If you're feeling like a treat, here's 15% off your next order, code RETURN15." The single email still does the engagement test (the click is the signal). The discount is a bonus for the buyers in the segment.

When to run re-engagement

The cadence I recommend:

- Monthly: review the Stale tier. If there are enough subscribers in it (say, 100+), run a re-engagement campaign on that cohort. - Quarterly: full audit. Re-check the engagement tier definitions, look for systemic drift, fix anything that's broken. - After major content changes: if you've significantly shifted what your sends look like, sound like, or cover, run an off-cycle re-engagement on the segment that signed up before the shift. Their original consent might not apply to the new content.

Don't run re-engagement weekly. The cadence is too tight and creates the same problem as multi-email sequences (constant re-asking).

What this all adds up to

A single re-engagement email, well-written, does in one send what a multi-email sequence does worse over three weeks. It identifies who's still actually paying attention. It removes who isn't. It protects deliverability. It produces a list that's smaller in absolute terms but dramatically more engaged.

Most senders' instinct is to fight harder for every subscriber. The Bosun's instinct is to read the rhythm honestly and adjust the crew accordingly. A smaller crew rowing in time beats a packed ship rowing chaos. The math on email lists works the same way.

One ask. One signal. One decision. Then you move on.

- Lyra