How do you know if an email campaign is working, beyond the open rate shown in your tool? Six metrics matter: deliverability rate, click-through rate, CTOR, bounce rate, complaint rate, and ROI per campaign, all calculated on emails delivered rather than emails sent.
Brevo, Mailjet, or Mailchimp’s dashboard can show a decent open rate while campaigns land in spam: the two numbers tell different stories. Since Apple Mail Privacy Protection preloads tracking pixels on iPhones, the open rate misleads more often than it informs. What follows covers the exact formulas, the 2026 thresholds behind them, and the most common reason these metrics deteriorate without anyone noticing.
The 6 email marketing KPIs to track first
Each metric is calculated on a specific denominator. Confusing “emails sent” with “emails delivered” artificially inflates rates and hides a list problem.
Deliverability rate
Deliverability rate is calculated as follows: (emails delivered ÷ emails sent) × 100. This metric determines every other one, since an email that never reaches the inbox can neither be opened nor clicked.
Validity and Litmus’s 2026 benchmarks set the limit: above 2% bounces on a send, the list should be considered degraded, and the target for hard bounces alone remains under 0.5%.
Open rate, with a caveat
Open rate divides opened emails by delivered emails: (emails opened ÷ emails delivered) × 100. The average rate sits around 18.2% in France according to the DMA France, 2026 report. List quality directly affects this number: our guide on improving open rate through list quality covers the link between invalid addresses and actual opens. Apple Mail Privacy Protection is the major limitation of this metric, and it’s covered below.
Click-through rate
Klaviyo measured an average CTR of 1.69% in 2026 on standard ecommerce campaigns, versus 5.58% on automated flows triggered by behavior (abandoned cart, welcome, win-back). Formula: (unique clicks ÷ emails delivered) × 100. Flows convert better, according to Klaviyo, because they arrive at the exact moment the contact takes action, while a campaign goes out whenever the sender decides to send it.
CTOR, the click-to-open rate
The click-to-open rate divides unique clicks by opened emails, not delivered emails: (unique clicks ÷ emails opened) × 100. It measures content relevance once the email has been viewed. Our engagement rate calculation method breaks down this formula and its variants step by step. One caveat though: CTOR is still built on the open rate, so it’s partially skewed by Apple MPP. Raw click-through rate, on the other hand, isn’t affected by this mechanism and remains the most reliable metric for gauging real engagement.
Bounce rate: hard bounce and soft bounce
A hard bounce is a permanent rejection (nonexistent address, closed domain). A soft bounce is temporary (full mailbox, server unavailable). Technically, the rejection happens during the SMTP exchange: the receiving server responds after the MAIL FROM command with an error code, documented as an Enhanced Status Code in the NDR sent back to the sender. Our article on hard bounce and soft bounce covers the most common SMTP codes and their causes. Overall bounce rate formula: (emails rejected ÷ emails sent) × 100.
Complaint rate
Formula: (complaints received ÷ emails delivered) × 100. The feedback loops set up by major mailbox providers report every “Report as spam” click back to the sender. Google recommends staying under 0.10% complaints to maintain stable deliverability on Gmail; beyond that, filtering becomes automatic, regardless of the content sent. These requirements tightened in February 2024: Google and Yahoo then imposed new rules on senders sending more than 5,000 messages within 24 hours to Gmail addresses. Staying under this threshold now determines inbox placement rather than automatic filtering to the spam folder.
How to calculate and track these KPIs
Email ROI rounds out the previous six metrics: ((revenue attributed to the campaign − cost of the send) ÷ cost of the send) × 100. A send that costs €200 and generates €3,000 in attributed revenue shows a ROI of 1,400%. According to industry benchmarks compiled by Omnisend, email marketing generates an average of $36 to $42 for every dollar invested, a ratio that climbs to $45 for every $1 invested in retail and ecommerce according to Litmus State of Email data.
Google Postmaster Tools (postmaster.google.com) shows your domain’s reputation on the Gmail side. Connecting it requires verifying domain ownership via a TXT record in your DNS: allow 15 minutes if the zone is already accessible, and up to 24 hours for DNS propagation to complete. The first data then takes up to 48 hours to show up in the interface. Microsoft SNDS does the same for Outlook and Hotmail. Neither covers smaller webmail providers: an inbox placement test (seed testing) sends the email to a panel of addresses spread across multiple providers to check where it actually lands, spam included. It’s the only method that shows the result before a customer reports it.
| Metric | 2026 Benchmark (France / Ecommerce Sector) | Alert Threshold |
|---|---|---|
| Open Rate | 18.2% on average (DMA France) | Reliability reduced by Apple MPP above 40% |
| Click-Through Rate (campaign) | 1.69% (Klaviyo, ecommerce) | Below 1% over several consecutive sends |
| Click-Through Rate (automated flow) | 5.58% (Klaviyo, ecommerce) | Not isolated in 2026 benchmarks |
| Overall Bounce Rate | Below 2% recommended | Above 2% (EmailAddress.ai) |
| Hard Bounce (isolated) | Target below 0.5% | A single send at 8% can tank reputation |
| Complaint Rate | Below 0.10% recommended | Above 0.10% (Google) |
Thresholds vary from one sending platform to another, however: Brevo, Mailjet, and Mailchimp don’t apply exactly the same tolerances before suspending an account.
Why your KPIs deteriorate
Three causes come up most often, and they’re rarely visible in the email tool’s main dashboard.
The first is sender reputation (sender reputation). A domain or IP address that suddenly sends a large volume after a pause gets its emails deferred (deferrals) by receiving servers before being outright blocked. A gradual IP warm-up schedule prevents this blocking, and it matters more on a dedicated IP than on a shared pool used by other senders. A single send with an 8% hard bounce rate can tank this reputation overnight, on Gmail Postmaster as much as on Microsoft SNDS.
Apple Mail Privacy Protection, launched with iOS 15 in September 2021 and enabled by default in Settings > Mail under “Mail Privacy Protection,” plays a role too. The mechanism, described by Mailchimp in its Apple Privacy documentation, preloads the email’s content through a proxy server as soon as it’s received, before the user even opens it. The tracking pixel fires, then, for emails that were never read. Our article on the impact of Apple Mail Privacy Protection on email marketing covers the consequences for open-based segmentation, often unusable for campaigns sent mostly to Apple addresses.
Invalid addresses piling up in the list itself are the easiest cause to fix. Deactivated addresses or old typos that were never cleaned up are enough to skew every calculation. Catchall addresses make the problem worse: they accept every message received without ever confirming that a mailbox actually exists behind the address. Each of these addresses inflates the denominator of every calculation and drags down every rate, even when the campaign content hasn’t changed.
The role of list verification in the reliability of your KPIs
Adding a verification layer to a stack that already includes a sending tool and an IP warmup service can seem redundant: the hard bounce rate stays under 2%, so why complicate things further? The answer comes down to when the address gets checked. A sending tool detects the hard bounce after the send, once the damage to reputation is already done. Upstream verification removes the address before it goes out, which prevents the bounce rather than just measuring it.
CaptainVerify documented the case of Groupe Obiz in 2026, which designs loyalty programs and ecommerce platforms for its B2B clients: integrating verification via API made it possible to secure lists upstream of each send instead of dealing with hard bounces after the fact, once reputation had already taken a hit. It’s this list hygiene approach that builds a stable deliverability rate over time, rather than a one-off adjustment triggered after a reputation alert. A clean list starts at collection: our newsletter definition covers opt-in and the GDPR requirements that determine the quality of the addresses collected.
The only way to know whether a list is dragging down your KPIs is to have it verified before the next send, rather than waiting for the next bounce spike to find out.
Best practices for improving your email marketing KPIs
These actions follow a logical order, from the most urgent to the most structural.
- Verify the list before every campaign, not just on initial import: at least every 30 days if send frequency is low, since a B2B database degrades continuously as people change jobs and domains shut down.
- Segment inactive subscribers at 90 days and launch a re-engagement campaign before excluding them permanently from future sends.
- Monitor Google Postmaster Tools and Microsoft SNDS every week, not just when a bounce spike shows up.
- Spread the volume ramp-up over a new IP or domain: start under 500 emails per day in the first week, then double it gradually.
- Track raw click-through rate as your primary reference rather than open rate on a list containing a significant share of Apple addresses.
A KPI that recovers after this cleanup confirms the diagnosis. A KPI that stays low despite a clean list shifts the problem toward content or send timing, away from deliverability. Mail servers, for their part, don’t care about any of these debates: they just count.
