Field data · published May 2026
Email Marketing Benchmarks 2026 — by Industry and by Deliverability Constraint
An operational view of where engagement metrics actually sit in 2026 across the industries we see most on Cloud Server for Email managed infrastructure, cross-referenced against the latest published datasets from Mailchimp, MailerLite, Klaviyo, Brevo, HubSpot, Litmus, Campaign Monitor, Validity, ActiveCampaign and Salesforce. The piece is built around two arguments. First, that traditional benchmarks based on open rate have become structurally unreliable since Apple Mail Privacy Protection (MPP), and the metrics that actually decide deliverability outcomes — complaint rate, bounce rate, inbox placement — deserve more attention than they get. Second, that “industry average” numbers should be read against the gating thresholds that Gmail, Yahoo and Microsoft now enforce, not against each other.
This note replaces the 2021 industry-by-day benchmark posts that used to live in the community forum. The 2021 dataset predates Apple MPP (April 2021), the Gmail/Yahoo bulk sender enforcement (February 2024), and the Microsoft Defender for Office 365 reputation overhaul (2025). Open rates from 2021 are not comparable to anything you measure today. We retire them and publish this in their place.
On this page
01
Why the 2021 numbers stopped working
Three structural changes between 2021 and 2026 have made every benchmark from that earlier period a misleading reference point.
1. Apple Mail Privacy Protection (April 2021)
MPP pre-fetches every email image on Apple’s servers, registering a “open” whether the recipient ever sees the message. Apple Mail accounts for roughly half of US email opens. The visible consequence: open rates across all ESPs jumped to artificial highs, then settled at MPP-inflated levels that do not reflect real human engagement. Open rates declined sharply from 48.69% in 2022 to 26.9% in 2025 after the wave of MPP normalisation, before stabilising under stricter measurement. Anyone presenting a flat 2021 open rate as a benchmark today is comparing two different metrics.
2. Gmail and Yahoo bulk sender enforcement (February 2024)
The joint Gmail/Yahoo policy introduced three hard requirements for senders above 5,000 messages per day: SPF + DKIM authentication aligned with the From domain, DMARC policy at p=none minimum, and one-click List-Unsubscribe headers. Most importantly, it codified a spam complaint rate ceiling of 0.3%, calculated per send (not as a rolling average). Above 0.3%, Gmail issues 5.7.x permanent failures and removes you from inbox until you stay below 0.3% for seven consecutive days. The 2021 dataset has no concept of this ceiling because it did not exist.
3. The infrastructure cost of dirty lists became fatal
In 2021 a bounce rate of 5% was unremarkable. In 2026, Gmail rejects domains that consistently bounce above 2% with permanent 5xx codes, and the rejection accumulates on reputation files that take weeks of clean sending to recover. The same is true at Microsoft Outlook.com and at Yahoo. List hygiene moved from a marketing nice-to-have to an infrastructure compliance requirement.
When a campaign “underperforms” against the 2021 benchmarks you find on older blogs, the most likely explanation is that you are measuring a 2026 reality (post-MPP, enforced thresholds) against a 2021 reality (no privacy protection, no enforced thresholds). The campaign is probably fine. The benchmark is broken.
02
Headline metrics across providers
Each major ESP publishes its own benchmark from its own customer base. The numbers diverge because the customer bases differ — Mailchimp leans SMB, Klaviyo leans ecommerce, Brevo leans European, MailerLite leans creator/SMB. We list them together so you can see the spread:
Three reasons. First, methodology differs — some sources count bot-prefetch opens, others filter them. Second, customer base composition differs — ecommerce, B2B and creator businesses have structurally different engagement profiles. Third, the calculation base differs — CTR can be calculated on sent, on delivered or on opened, and each gives a different number for the same campaign.
For your own programs, the most useful single benchmark is the median across your vertical from your ESP — not a cross-source average.
03
Open rate by industry
Data consolidated from MailerLite 2026 (the broadest published dataset, 3.6M campaigns), with cross-reference to WebFX/Campaign Monitor aggregates where vertical names differ. These are MPP-influenced raw open rates as ESPs report them — for the human-engagement view, see CTOR in section 5.
These verticals reach audiences who actively opted in to a community they identify with. The opt-in is high-intent and the unsubscribe friction is low; the audience self-selects to people who want the mail. Commercial categories converge on the lower band because the opt-in is transactional (discount, account creation, free trial) and the audience-to-mail relationship is utilitarian rather than identity-driven.
Range, not single number
Treat any single number as the midpoint of a wide range. MailerLite’s data of 3.6 million campaigns shows industry open rates ranged from 30.1% to 55.71% on their methodology; WebFX’s consolidated view has government at 30.5% and daily deals at 13%. Same year, same broad question, very different numbers depending on the dataset.
04
Click-through rate by industry
Click-through rate is calculated as clicks divided by emails sent (or delivered, depending on source). Because CTR is not affected by image-prefetch the way open rate is, it is the more reliable engagement metric in 2026.
Two patterns stand out. First, flows beat campaigns by 3× on click rate in ecommerce (Klaviyo 2026: 5.58% vs 1.69%), because flows are triggered by intent signals (cart abandonment, post-purchase, browse abandonment) while campaigns broadcast to everyone. Second, B2B clicks higher than B2C in absolute terms when the offer is informational (research, whitepaper, webinar) — the click cost is lower than the click cost of a B2C purchase, so the threshold to click is lower.
05
CTOR — the metric that matters post-MPP
Click-to-open rate (CTOR) is clicks divided by opens. It controls for the audience that actually saw the message, so it measures content quality rather than subject-line quality or list quality. Because the denominator is opens (an MPP-influenced number), CTOR has its own caveat — but it is less corrupted than raw CTR because both numerator and denominator move together with bot opens.
If raw opens are bot-inflated by MPP but clicks are real, CTOR systematically understates true human engagement — the inflated denominator pushes the ratio down. That is a useful direction of bias: a CTOR target like “stay above 8% in B2B” is a conservative target that controls for bot noise. If your CTOR is dropping over time, the most likely real-world cause is content fatigue with your existing list, not Apple changing the rules again.
06
Bounce rate and the new 2% ceiling
Bounce rate is the percentage of messages that fail delivery. The 2021 conventional wisdom of “keep it under 5%” is obsolete. A bounce rate above 2% triggers permanent 5xx rejections from Gmail as of November 2025, and Microsoft and Yahoo apply similar (less publicised) thresholds.
The 2026 bounce-rate tiers
Hard vs soft breakdown
From the consolidated 2026 data: hard bounces average 0.4% (permanent: bad address, dead domain, mailbox closed) and soft bounces average 0.7% (temporary: full mailbox, server unavailable, greylisting). Total bounce should land under 1.5% on a well-maintained list; under 0.5% on a list that runs verification before sends and removes 30-day inactives.
Industries with high subscriber turnover — restaurants (staff churn), daily deals (signup-and-abandon), and very-large-volume promotional senders — show bounce rates 1.5–2× higher than other categories. If you operate in one of these, your list needs verification as a continuous process, not a one-off cleanup.
07
Spam complaint rate — the 0.3% gating line
This is the most important section in the page and the one most often missed in marketing benchmark reports. Spam complaint rate is the only engagement metric that directly triggers enforcement at the inbox provider. Everything else — open rate, click rate, even bounce rate — is a lagging indicator. Complaint rate is the trigger itself.
The 2026 enforcement thresholds
A common mistake we see in postmortems: a sender averaging 0.08% across the month assumes they are safe, but one bad campaign at 0.45% triggers the Gmail enforcement signal anyway. The threshold applies per-campaign, not monthly average, meaning a single bad send can trigger filtering. The recovery clock starts from the bad send, not from the rolling average.
Industry baselines
What complaint rate actually means operationally
The math is harsh at low volumes. If you are sending low volumes — say, 100-200 emails a day — a single complaint gives you a 0.5-1% rate. Inbox providers know this and apply different sensitivity bands by volume class, but the spirit of the threshold is the same: you cannot send at scale on dirty consent.
For our managed-infrastructure clients, the operational rules we set are tighter than the enforcement thresholds:
- Target ceiling: 0.08% (Gmail’s own recommendation, not the 0.10% policy line)
- Investigate above 0.10% by segmenting the send and isolating the offending segment
- Immediate intervention above 0.20% — pause sends, root-cause, do not just keep going
- Above 0.30%, the campaign is already in enforcement — the question is recovery, not prevention
For deeper field data on complaint behaviour, see our notes on tracing complaint-rate spikes to their source and the 2.5% club of senders one year after the DMARC mandate.
08
Flows vs campaigns — the revenue split
Klaviyo’s 2026 Omnichannel Benchmark Report quantified what every practitioner already knew: triggered flows generate disproportionate revenue relative to their send volume.
Email flows generated nearly 41% of email revenue from just 5.3% of sends in 2026. The implication is structural: a sender who under-invests in triggered automations and over-invests in broadcast campaigns is leaving most of the available email revenue on the table.
The top-10% benchmark
Within the flow category, the gap between average and best-in-class is also wide. Top 10% email flows achieve RPR as high as $7.79 and click rates over 10%, demonstrating that sophisticated segmentation, content relevance, and orchestration define best-in-class performance. RPR (revenue per recipient) is the metric ecommerce should optimise; abstract opens are a distraction once the order conversion path is measurable.
The five flows that move the most revenue
- Welcome series — the highest RPR flow in nearly every vertical; the moment of strongest intent the sender will ever have
- Browse abandonment — intent without commitment; lower RPR than cart but reaches more users
- Cart abandonment — the highest single-message RPR; 60-70% of cart starts never finish without it
- Post-purchase / replenishment — the second-highest RPR after welcome in consumables and beauty
- Win-back — targeted at lapsed subscribers; lower RPR per send but cheap because the list already exists
09
Transactional email benchmarks
Transactional emails (order confirmations, shipping notices, password resets, account alerts) operate on different physics from marketing. The recipient asked for the message implicitly by taking an action; the message is expected and high-attention.
Transactional emails have a CTR of 4.8% — nearly 2x the average, and are read by 75% of recipients. The Brevo dataset shows transactional CTR of 7.39%, even higher, because their sample skews toward shipping confirmations and account alerts where the click is essentially mandatory.
Transactional email must run on a separate IP pool from marketing. The reason is operational, not theoretical: a marketing campaign that gets throttled or filtered also throttles the password-reset and 2FA emails on the same IPs, breaking core product functionality. We isolate transactional streams by default on every managed installation. See our transactional infrastructure page for the architecture.
10
ROI per dollar by vertical
Email ROI is consistently the highest of any digital channel, but the spread by methodology and vertical is wide. The Litmus benchmark of $36 per $1 spent is the most cited figure; the HubSpot estimate puts it higher (up to $42), and Klaviyo’s top-quartile ecommerce data puts top-tier flows at $7.79 RPR.
Two contextual points. First, ROI is calculated against the marketing spend on email (platform fees, design, list acquisition), not against revenue alone — this is why ecommerce flows can show extreme RPR while the channel-wide ROI looks “only” 36×: the same dollar spent on the platform powers thousands of sends. Second, the ROI calculation assumes deliverability is solved. A campaign that lands in spam returns $0 regardless of how good the content is. The ROI numbers above are what is achievable when the infrastructure is right; they are the upper bound, not the universal.
11
Cold email vs marketing email
Cold outbound email (sent to recipients who have not opted in) operates under fundamentally different metrics from opt-in marketing. Conflating the two benchmarks is the most common mistake we see in SDR teams comparing their numbers to marketing reports.
The complaint rate row is the operational story. A cold-email programme running at 2% complaint rate is normal and well within the operating envelope of a cold-email programme. The same complaint rate from a marketing programme would trigger Gmail/Yahoo/Microsoft enforcement within a single send. The two streams cannot share infrastructure: a 2% complaint rate on a shared IP pool destroys the marketing programme’s reputation overnight.
This is the architectural argument for dedicated cold-email infrastructure: not because cold email is inherently spam, but because the structural complaint rate it produces is incompatible with the thresholds that protect marketing and transactional streams. Our cold email infrastructure exists precisely to absorb the structural complaint volume that cold outreach generates without contaminating anything else the client sends.
12
How to read benchmarks correctly
Three reading rules we recommend to clients who bring us “here is a benchmark report, are we underperforming?” questions.
1. Match the benchmark to the email type
An SDR running 3.2% CTR on cold outreach is in the top quartile of cold B2B; the same 3.2% on a marketing list would be excellent. Comparing across these is meaningless. An SDR manager reviews their cold outreach numbers: 3.2% CTR, 0.4% bounce rate, 0.07% spam complaints. They pull a B2C benchmark report showing an industry average CTR of 2.0 to 2.5% and assume their 3.2% is only slightly above average — that comparison is structurally wrong; the benchmark and the actual program measure different things.
2. Match the benchmark to the methodology
Mailchimp counts opens including bots; some reports filter bots. Klaviyo measures CTR on delivered; HubSpot measures on sent. A 6% click rate in one methodology is a 5% click rate in another. Look at the methodology footnote before treating any single benchmark as a fixed line.
3. Match the benchmark to your vertical, not your audience
An apparel brand selling to fashion-conscious millennials should not benchmark against “e-commerce average”; it should benchmark against apparel specifically. Klaviyo’s vertical breakouts and Mailchimp’s industry reports are useful precisely because they are vertical-specific. Cross-vertical “industry averages” mostly average two structurally different audiences (B2B vs B2C) and produce a number that fits neither.
Your own historical performance, at the same time of year, against the same audience, on the same infrastructure. That comparison controls for everything that varies across published benchmarks and tells you whether you are improving. Published benchmarks are useful for direction-setting and for management conversations, but the trajectory of your own metrics is the ground truth.
13
Methodology and sources
This piece consolidates published 2026 data from the following primary sources. Numbers shown above are either drawn from a single source (cited inline) or aggregated where multiple sources agree on the direction and the spread is narrow enough to support a range.
Cloud Server for Email proprietary data
Where the numbers come from our own infrastructure rather than a published source (the spam complaint operational rules in section 7, the cold-email vs marketing comparison in section 11, and the transactional infrastructure rule in section 9), the data is drawn from aggregated observations across managed PowerMTA installations operated for clients during 2024–2026. We do not publish per-client data and do not include any client-identifying information in these numbers.
What this piece is not
This is not a substitute for the source reports. If you are a marketing operations lead building an internal benchmark deck, read the primary reports (links above) for full methodology footnotes, segmentation tables and regional breakouts. This piece is an operational synthesis — what the numbers mean for senders operating against the 2026 enforcement environment, written from the infrastructure side of the conversation.
Where benchmarks meet infrastructure
Benchmarks describe outcomes. Infrastructure produces them.
The hardest number on this page is the 0.3% Gmail complaint threshold. Hitting it ends conversations with the inbox until you rebuild reputation over weeks. Our managed PowerMTA + MailWizz infrastructure is built to keep you below that line by default — dedicated IPs, daily Postmaster Tools surveillance, complaint-rate alerting per campaign.
See infrastructure Talk to an engineerRelated operational notes
- Complaint Rate Spikes — Tracing the Source — field methodology for finding the segment driving an inbox-provider warning
- DMARC p=reject One Year After the Mandate — what happened to senders who moved vs the ones who stayed at p=none
- Q1 2026 Permanent Rejections — Field Data — the actual rejection codes we logged during the first quarter of enforcement maturity
- Inbox Placement Is a Lagging Indicator — why placement reports tell you yesterday’s story
- Six Months with Postmaster Tools v2 — what changed and what to monitor