Your Promo Emails Quietly Stopped Working: Fixing Campaign Deliverability Decay on Shopify

Priya runs Ember Lane, a home fragrance brand on Shopify doing about $1.1M a year, with roughly 28% of that revenue coming through Klaviyo campaigns. Last October her promo calendar looked exactly like it had for two years. Same cadence, same templates, same list.

By March, campaign revenue was down 61% while nothing else had moved. Traffic steady, ads steady, reviews fine.

On the discovery call she put it the way we’ve now heard dozens of times: “our promo emails to regulars basically stopped working.” No bounce spike, no platform warning, no angry replies. This is what it looks like when Shopify email campaigns stop reaching the inbox: the sends still go out, the dashboard stays green, and the revenue quietly walks out the back door.

The slow leak that looks like a revenue problem

Merchants find this problem backwards. They notice the money first, then go looking for a marketing explanation: weak offers, tired creative, seasonality. Priya’s team spent two months rewriting subject lines before anyone asked whether the emails were being seen at all, and that instinct is understandable, because every other channel fails loudly. A broken pixel breaks your ROAS overnight. A busted checkout throws errors you can screenshot.

Deliverability decay does neither. It’s the only revenue leak in ecommerce that gets worse specifically because it never triggers an alert.

Mailbox providers never send you a memo. Gmail doesn’t email you to say your reputation dropped from good to fair. Instead, an increasing slice of your list stops receiving you in the primary tab, then stops receiving you in Promotions, then finds you in spam, and each step is invisible from inside Klaviyo or Shopify Email because a message delivered to the spam folder still counts as delivered.

That accounting quirk matters more than any other detail in this post. Your delivery rate can sit at 99.5% while your inbox placement sits at 60%. The platforms report the first number. Your revenue reports the second.

Why placement rots gradually instead of breaking loudly

Sender reputation isn’t a switch, it’s a rolling score that Gmail, Yahoo and Outlook each calculate separately, weighted heavily toward how recipients treated your last few weeks of mail. Opens that never happen, deletes without reading, a handful of spam complaints. Each one nudges the score down a little.

And the decay compounds. Lower placement means fewer people see the mail, which means engagement drops further, which lowers placement again. A store can ride that spiral for six to nine months before the revenue chart makes anyone nervous.

We audited a kitchenware brand at about $3M GMV last spring that had lost inbox placement at Gmail specifically. Yahoo and Outlook were fine. Their overall numbers looked like a mild dip, but 68% of their list was Gmail addresses, and inside that segment placement had collapsed. Provider-by-provider breakdowns are the difference between seeing this and missing it.

The signals worth trusting after Apple broke open rates

Open rate used to be the early warning. Since Apple Mail Privacy Protection started auto-fetching tracking pixels, a big chunk of your opens are machines, which means your open rate can hold steady while real humans stopped seeing your mail months ago.

So we trend numbers that machines can’t fake. Click rate per campaign over the trailing six months, split by mailbox provider. Spam complaint rate, which needs to stay under 0.1% to keep you inside the thresholds spelled out in the bulk sender rules that Google and Yahoo tightened in 2024. Revenue per recipient, which is the number that flagged Ember Lane. And the share of campaign clicks coming from Gmail addresses, because when that share falls faster than the rest of the list, Gmail is filtering you.

Klaviyo’s own deliverability guidance is worth a read here, mostly because it confirms an uncomfortable point: the platform can’t see your placement either. Nobody’s dashboard can. You infer it from engagement patterns, or you pay a seed-testing tool to measure it directly.

Dormant subscribers are the debt that compounds

Here’s the pattern we find on pretty much every decayed account. The store built its list aggressively in 2023 and 2024, popups and giveaway campaigns and a purchased-adjacent “partner” list nobody wants to talk about. The list grew to 40,000. Active clickers in the last 90 days: maybe 6,000.

And every campaign still goes to all 40,000.

Mailbox providers read that as a sender who doesn’t care whether recipients want the mail. The 34,000 silent addresses don’t just fail to buy, they actively vote against you with every unopened send. A few of them have gone stale and turned into spam traps. Mailing them at full cadence is the single most damaging habit in ecommerce email, and it’s also the easiest one to fix, because the fix is subtraction: send less, to fewer people, and watch placement recover.

We ran exactly this subtraction for a pet supplies store at around $2.4M last year. Cutting their sendable list from 51,000 to 14,000 engaged profiles felt like burning money to the founder, right up until the next month’s campaigns outearned the previous quarter’s average by 19% with a third of the volume. Fewer sends, more revenue, and a complaint rate that dropped below 0.05% for the first time in a year.

Authentication drift, the checkbox everyone set once

The other quiet killer is technical. SPF, DKIM and DMARC got configured once, usually by whoever set up Klaviyo or Shopify Email in the first place, and nobody has looked since.

Then things drift. The store switches theme agencies and the DNS gets touched. Someone adds a new sending tool, a review platform or a helpdesk that mails from the same domain, and the SPF record stops covering everything. DMARC sits at p=none forever, which providers increasingly treat as a shrug.

On Ember Lane, we found the brand had moved to a dedicated sending domain in 2024 but left the old shared-domain DKIM key referenced in DNS, and a transactional tool added later was failing alignment entirely. None of it broke delivery outright. All of it eroded trust, send after send, for over a year.

The six-week re-warm we ran for Ember Lane

Rebuilding placement is boring on purpose. You shrink your sending down to the people who demonstrably want your mail, prove to providers that engagement is high, then expand in steps.

Weeks one and two, Ember Lane mailed only 90-day clickers, about 4,800 people, twice a week, with plain useful content rather than discount blasts. Weeks three and four widened to 180-day openers-plus-clickers with authentication fixed and DMARC moved to quarantine. Weeks five and six reintroduced the broader engaged segment and stood up a sunset flow that gives silent subscribers a re-permission ask, then removes them.

Gmail click share recovered in week four. By week nine, campaign revenue was back to 84% of its old baseline. The list was 22,000 addresses smaller than the one that had been underperforming. Priya’s team was nervous about deleting subscribers right up until they saw that the smaller list was outearning the big one.

The discipline is the hard part, not the plan. Every re-warm we’ve run has a moment around week three where someone wants to blast the full list for a product launch, and giving in resets the clock. We put the launch in the engaged-segment sends instead and it did fine.

A maintenance routine that keeps you out of trouble

Deliverability isn’t a project, it’s a habit. The stores that never call us in a panic run the same short checklist monthly.

They check complaint rate and provider-level click trends. They run the sunset flow rather than pausing it in Q4 (everyone wants to pause it in Q4, that’s exactly when volume spikes make it matter most). They confirm SPF, DKIM and DMARC still pass after any app or DNS change. They cap list growth tools to double opt-in when signup quality drops.

One more habit worth stealing: they schedule the check against something that already happens monthly, usually the ad-spend review, so it never gets skipped. Deliverability work that depends on someone remembering it is deliverability work that stops by August.

Twenty minutes a month. Cheap insurance against a six-week rebuild.

What we keep telling clients

Deliverability decay is a trust problem wearing a technical costume. Providers stopped trusting your mail because, statistically, your recipients stopped acting like they wanted it. Every fix that works flows from taking that seriously, and every fix that fails is an attempt to route around it, usually by switching platforms and dragging the same bloated list to a new sender with no history at all.

The merchants who get burned worst are the ones with the healthiest-looking dashboards. Delivered rate high, opens propped up by Apple’s pixel-fetching, list growing every month. All three of those numbers can be excellent while your actual reach shrinks by half.

So watch revenue per recipient like you watch ROAS, and treat a sustained slide as a placement question before a creative one. Subject line tests can’t fix a spam folder.

Priya’s version of the lesson: Ember Lane now mails 17,000 fewer people than it did last October and makes more money from email than it did then. She reviews the provider-split click report on the first Monday of every month, and the promo calendar she almost blamed for the whole mess is back to running unchanged.

Questions we get every week

How do I know if my campaigns are landing in spam without paying for a tool?

Split your click rates by mailbox provider in Klaviyo and trend them over six months. A collapse at one provider while others hold steady is filtering, not fatigue. You can also check placement manually with a handful of seed addresses you control at Gmail, Yahoo and Outlook.

Will switching from Klaviyo to another platform reset my reputation?

No, and it usually makes things worse. Reputation follows your sending domain and your list more than your platform, and a new platform means starting engagement history from zero, so fix the list and the authentication where you are.

How long does a re-warm take for a typical Shopify store?

Plan for four to eight weeks depending on how deep the damage goes. Gmail tends to respond within two to three weeks of consistently high engagement, Outlook is slower. The sunset flow you build during the re-warm is what keeps you from repeating it next year.

Is it really safe to stop mailing half my list?

The dormant half of your list is generating close to zero revenue and actively suppressing the reach of the half that buys. Every re-warm we’ve run has ended with a smaller list earning more. Keep the dormant addresses in a re-permission flow for 60 days, then let them go.

If email revenue is sliding and nobody can tell you why, talk to us about a deliverability audit and we’ll show you exactly where your campaigns are landing.