The Email Deliverability Crisis Nobody's Warning Mid-Market Marketers About

A marketing director we work with ran a campaign in February. Her list was clean. Her copy was on-brand. Her platform reported a 38% open rate, well above her usual benchmark. Two weeks later, her CEO called to ask why their largest customer hadn't received the renewal-window email she'd sent them. She checked the platform. Sent. She checked the customer's inbox. Nothing in primary, nothing in promotions, nothing in spam. Just gone.

That story is now the modal experience of mid-market email marketing. The platform reports look fine. The numbers look fine. The actual messages, increasingly, are not arriving. Most marketing teams have not been told the rules changed. The platforms they use are not incentivized to tell them, because the failure shows up downstream of the dashboard.

Email deliverability quietly became the most expensive unsolved problem in mid-market marketing in 2024 and 2025. Google and Yahoo rewrote the rules. Apple changed how opens are measured. Spam filters got stricter. Cold email response rates collapsed. And almost none of this made it into the marketing-conference talk track, because the people running the talk track sell tools that depend on you not realizing how broken the underlying channel has become.

This paper covers what changed, what it costs, and what a mid-market marketing team should actually do this quarter. None of the answers are expensive. Most of them are an afternoon of work.

What changed in 2024 that broke everyone's email program?

Three things changed in roughly twelve months, and they compound on each other in ways most marketing platforms have not surfaced clearly.

The first is Google and Yahoo's bulk sender requirements. Google announced the new rules in October 2023 and started enforcing them in February 2024. Any sender pushing 5,000 or more messages per day to Gmail or Yahoo accounts now has to authenticate with SPF and DKIM (with DMARC alignment), provide one-click unsubscribe in the message header, process unsubscribes within two days, and stay under a spam complaint threshold of 0.3%. The threshold is per day, per Gmail recipient. Most mid-market marketing teams cross it on every campaign send and never realized they did.

Senders who fail any of those requirements do not get a warning email. They get throttled. Then their messages start landing in promotions. Then in spam. Then they stop landing at all. The descent is fast and the platform reports underneath the surface keep saying the messages were "sent" the whole way down.

The second is Apple's Mail Privacy Protection. MPP, introduced in iOS 15, automatically pre-fetches images on behalf of Apple Mail users before the user opens the message. The tracking pixel fires whether or not the human ever sees the email. Apple Mail is a meaningful share of every B2C list and a non-trivial share of B2B lists too. This means open rates are inflated for any sender with Apple Mail recipients, and the inflation is uneven across segments, which makes A/B testing on opens unreliable. The Mailchimp all-industry benchmark currently shows a 35.63% open rate, 2.62% click rate, and 0.22% unsubscribe rate, and Mailchimp itself notes that opens are inflated by MPP and clicks are the more reliable signal. Most platforms still default open rate as the headline metric.

The third is the slow strangulation of cold email. The average B2B cold email response rate is roughly 4 to 5% in 2024 and 2025, down from around 8.5% in 2019, and roughly 17% of cold emails never reach any inbox. The decline is a combined effect of every prospect's filter learning faster, every outbound tool's templates collapsing toward the same patterns, and the bulk-sender rules applying with full force to high-volume cold outbound. Cold email is not dead. It is, however, much closer to broken than the SaaS pitching it admits, which is why the anatomy of a 5-figure cold email campaign looks nothing like the spray-and-pray playbook of 2019.

These three shifts arrived at the same time. The mid-market team running the same email program in 2026 it ran in 2022 is now sending into a fundamentally different inbox.

What does broken deliverability actually cost?

The cost is invisible, which is what makes it dangerous.

A traditional marketing problem (poor open rates, weak copy, bad list segmentation) shows up clearly on the dashboard. A deliverability problem hides behind the dashboard. The platform records the message as sent. The recipient never sees it. The team plans the next campaign on the assumption the previous one reached its audience. The math underneath every forecast is wrong, and nobody can tell.

Three patterns recur across the teams we audit.

Renewal and retention messages disappear into spam. This is the expensive one. A B2B customer who does not see your renewal-window email does not renew on time. The CSM finds out a month later and has to start a save conversation. A B2C customer who does not see your reactivation email churns silently. Deliverability damage shows up in revenue retention numbers, but nobody traces it back to the inbox because the platform said the email was delivered.

Lead nurture sequences run on empty. A sales rep follows up on a lead that was supposed to have received a five-message nurture sequence. The lead has received zero of them. The conversation starts cold instead of warm. Conversion rates slip. The team blames the SDR or the offer. The actual problem was that the sender domain reputation collapsed three months ago.

Cold outbound torpedos the brand domain. This is the most preventable one. A team runs cold outbound from the same root domain as their transactional and marketing email. The cold outbound generates spam complaints. The complaints poison the entire root domain's reputation. Now the password-reset emails are landing in spam too. By the time the team notices, the damage takes weeks to undo.

HubSpot's 2026 State of Marketing report found 83.5% of marketers are expected to produce more content, with 35.7% saying "much more." The volume pressure is real, and the natural place to apply that volume is email. But applying more email volume to a poorly authenticated, badly segmented program does not produce more results. It produces a faster reputation decline. HubSpot's data also shows 86.4% of marketing teams now use AI in at least a few areas, with content creation as the top use case, which is exactly the volume accelerant pushing more sends through under-maintained infrastructure, and is the same tension at the heart of content velocity vs. quality.

Why are most marketing teams unprotected?

Three structural reasons.

The first is that email infrastructure used to be invisible because it worked. SPF, DKIM, DMARC were set up once, by an IT person who has since left, and nobody touched them again. Most marketing teams could not tell you what their DMARC policy is. Most CMOs could not tell you which subdomain their cold outbound runs from. The infrastructure has degraded quietly across many businesses because nobody owns it.

The second is that the marketing platforms benefit from the appearance of healthy delivery, not the reality of it. A platform that tells a customer "your sender reputation is collapsing, here are five things you need to fix" is a platform the customer might churn from. A platform that shows clean dashboards with healthy open rates (inflated by Apple, but not flagged as such) is a platform that gets renewed. The incentives do not line up.

The third is that the bulk-sender requirements went live without a coordinated industry communication push. The vendors that did warn customers tended to be the deliverability-specialist tools (Postmark, SparkPost, Mailgun, dedicated IP providers) whose audience is technical. Mid-market marketing teams using HubSpot, Mailchimp, Constant Contact, ActiveCampaign, Klaviyo, or one of the dozen sales-outbound tools mostly received an in-app banner they swiped past.

The result: a large share of mid-market email programs are running on infrastructure that does not meet the new minimum bar, and nobody has told the marketing leader in plain language what the bar is or how to clear it.

What does a healthy 2026 email program actually look like?

We work with mid-market and SMB marketing teams across a wide range of sending volumes, and the healthy programs share the same operational pattern. None of it is exotic.

Authenticated, aligned, and policed. SPF, DKIM, and DMARC published correctly for every sending domain. DMARC at minimum at p=quarantine, ideally at p=reject for any domain that is not actively in transition. DMARC reports monitored monthly, even briefly, to catch unauthorized senders or misaligned subdomains.

Volume segmented by subdomain. Transactional email runs from one subdomain (e.g. mail.brand.com). Marketing email runs from a second (e.g. news.brand.com). Sales outbound runs from a third (e.g. team.brand.com or, increasingly, an entirely separate root domain set up for outbound). The reputations are isolated. A bad cold campaign cannot kill the password-reset deliverability.

List hygiene as a recurring practice, not a one-time event. Anyone who has not engaged (opened, clicked, replied, purchased) in 12 months gets removed or moved to a re-engagement-only segment. Hard bounces removed within days. Spam complaints reviewed weekly. The team sends to a list that wants to hear from them, not to the list they wish they had.

Clicks, replies, and conversions as the headline metrics. Opens are watched as a directional signal but not used for testing or forecasting. The Mailchimp benchmark of 2.62% click rate is a more useful comparison point than the 35.63% open rate, because clicks are not inflated by MPP.

Spam complaints under 0.1%. Google's published threshold is 0.3%. Healthy senders sit closer to 0.1% as a working ceiling. The teams that monitor this proactively can catch a deliverability problem in the week it starts. The teams that don't catch it three months later when revenue numbers slip.

Re-engagement sequences before list-pruning. Before removing a dormant subscriber, send a short re-engagement sequence with a clear opt-in or opt-out signal. Subscribers who reactivate are gold. Subscribers who explicitly opt out are doing you a favor. Subscribers who silently never engage again are dead weight that hurts your reputation.

Cold outbound separated and contained. Anything resembling cold outreach (sales prospecting, partnership outreach, link-building, PR pitches) lives on a separate domain or subdomain with its own warm-up, its own monitoring, and its own complaint thresholds. The brand domain stays clean.

This is not a complicated playbook. It is a basic set of operational hygiene practices that maybe a third of the mid-market teams we audit have fully implemented.

How does cold email fit into this?

Honest answer: cold email is harder than it was, the rules are stricter, and the response rates are lower. It still works for tightly targeted, well-personalized outreach to a specific persona with a clear reason to care. It does not work for the spray-and-pray model that defined 2018 to 2022 outbound.

The teams making cold email work in 2026 do five things differently than the teams that aren't.

They send less, to fewer prospects, with more personalization. The math has flipped. A 50-message-per-day campaign with a 20% reply rate is now beating a 500-message-per-day campaign with a 2% reply rate, and the difference is reputation damage avoided.

They warm up sending domains for at least three to four weeks before scaling. This is not a guess. It is how Gmail's reputation system works.

They keep cold outbound entirely off the brand's primary root domain. A new domain, properly authenticated, dedicated to outbound, with a redirect to the main brand site if a prospect clicks through.

They write like a human. Patterns that read as obviously template-driven get filtered before a human ever sees them. The spam filter has trained on millions of outbound-tool templates and recognizes them faster than ever. The teams winning at cold use AI for the research and personalization layer (industry context, recent triggers, role-specific framing) and human-or-near-human writing for the actual message, which is why personalization at scale is mostly a lie without a real voice and segmentation layer.

They measure replies, not opens. Cold email opens are even more polluted by privacy and pre-fetching than marketing-email opens. Replies are the only signal worth tracking.

For most mid-market businesses, cold email should be one channel in a mix that includes content, organic social, and inbound, not the lead engine the entire pipeline depends on. Teams that overweight cold are the most exposed to the reputation collapse this paper is about.

What should a mid-market marketing team do this quarter?

Five concrete moves. None of them require a vendor change. All of them can be done in weeks, not months.

1. Audit your authentication. Pull your SPF, DKIM, and DMARC records for every domain you send from. Use a free tool like MXToolbox or Google Postmaster Tools. Confirm SPF passes, DKIM passes, and DMARC is at least at p=quarantine. If any of those is broken, the rest of this list does not matter yet. Fix authentication first.

2. Set up Google Postmaster Tools. It is free. It tells you your domain reputation, your spam complaint rate, and whether Google considers your sender behavior healthy. Most marketing teams have never logged into it. Twenty minutes of setup and you have a real-time deliverability dashboard for the largest inbox provider in the world.

3. Separate your sending streams. If transactional, marketing, and outbound are all running from the same root domain, set up subdomains for each. Migrate cold outbound off the brand entirely if you can. Reputations are isolated by domain. Take advantage of that.

4. Clean the list. Remove anyone who has not engaged in 12 months. Send a re-engagement sequence to anyone in the 6 to 12 month window with a clear stay-or-go question. The list will get smaller. The deliverability will get better. The dollar-per-recipient will go up. This is a trade most teams should make.

5. Switch your headline metric to clicks (and replies, for outbound). Update your reporting. Train the team to look at clicks as the leading indicator. Watch opens as a directional secondary signal. Make sure nobody is making A/B testing decisions on opens alone, because the data is too polluted to be useful.

If you do these five things this quarter, you will land more email than you did last quarter. That is the entire goal.

Where does AI fit into all of this?

Useful: AI is good at writing higher-quality, more varied marketing copy that does not pattern-match to spam-flagged templates. It is good at segmenting and personalizing at scale, which makes the smaller, tighter list that healthy 2026 email requires actually feasible. It is good at drafting re-engagement sequences, summarizing engagement data, and surfacing patterns in why specific subscribers churn.

Not useful: AI cannot fix authentication. It cannot lower a spam complaint rate that is already too high. It cannot rebuild a poisoned domain reputation. The infrastructure layer is human work and engineering work, not content work, and no AI tool will solve it for you.

Most mid-market teams do not have a content quality problem. They have a content volume problem (driven by HubSpot's finding that 83.5% of marketers are expected to produce more content) layered on top of an infrastructure-hygiene problem the marketing team has never been told to think about. There is also a credibility tax to think about: the Nuremberg Institute found that 52% of consumers reduce engagement with content they believe is AI-generated, so AI volume into email needs to be paired with editorial oversight or it accelerates spam complaints rather than engagement. AI helps with the first problem. It does not help with the second.

What does the next two years probably look like?

The privacy ratchet does not loosen. Apple will keep tightening. Google will likely raise the bulk-sender bar again, particularly around DMARC enforcement and complaint thresholds. The walled gardens of consumer mail (Gmail, Yahoo, Outlook, Apple) are increasingly the entire delivery surface, and each has its own rules, monitoring, and reputation system.

Cold email will keep getting harder. Tighter targeting, smaller volumes, better personalization, and proper domain separation will become the baseline rather than the advanced practice. Teams running 2019-style outbound will see response rates fall further toward zero.

Marketing teams that treat email as a critical infrastructure (the way they treat the website or the CRM) will outperform teams that treat it as a campaign-by-campaign tool. The teams that staff a person, even part-time, with deliverability ownership will catch problems early enough to fix them. The teams that don't will keep being surprised by quiet revenue retention slips that nobody can quite explain.

The good news: this is one of the few problems in marketing where the moves are clear, the science is settled, and the work is bounded. A team that decides to take it seriously can be in a healthy state in one quarter.

A short, honest soft sell

FUEL is a marketing platform built for the parts of this problem that are content problems. We do not sell email infrastructure. We do not pretend to fix authentication, sender reputation, or the underlying deliverability layer. That work is yours, and the playbook above is the one we recommend.

What we do help with is the content side. Better-written, more varied, more on-brand email copy is the layer that sits on top of healthy infrastructure and decides whether the inbox readers who do see your messages actually engage. Teams using FUEL produce email sequences in their own voice across welcome, nurture, re-engagement, and win-back, in an afternoon rather than a quarter, which makes the smaller, tighter, healthier list approach actually viable.

If you read this paper and recognized your own program, the most useful next step is probably not to buy more software. It is to spend a half-day on the five-step audit above. Once your infrastructure is healthy, then the question of how to feed it with better content becomes worth asking.

Run the Foundation Report on your business. If the output surprises you, that is the point.

If you're an agency, generate a Foundation Report on a client you have worked with for years. If the output does not challenge your thinking, walk away. If it does, the team plans are priced for agencies ready to scale what works.

Generate My Foundation Report

If a different paper in the series is more relevant to where you are right now, the full list is at /white-papers.