A password-reset request can succeed in your application while the email is accepted by a recipient server and sorted into spam. Transactional email deliverability is a product reliability issue: this guide gives SaaS teams a practical path from DNS setup to diagnosing missed resets, receipts, and alerts.

Transactional email deliverability is the ability to get product-triggered messages to a useful mailbox location, not merely to submit them successfully. Start by distinguishing recipient-server acceptance from inbox placement. Authenticate the domains your messages actually use, separate operational mail from campaigns, and monitor events by provider and message type. When placement changes, compare those signals with your own baseline before changing templates or sending infrastructure.

What transactional email deliverability means

Transactional messages follow a user action or account event: a password-reset link after a reset request, a receipt after a purchase, or a shipping update when an order changes status. Each should arrive when the user needs it and make clear which action it supports.

The distinction to keep in your logs is between a send request, delivery, and mailbox placement. An API response can show that a request was accepted by your email service. A delivery event generally indicates acceptance by the recipient’s mail server; it does not tell you whether the message went to the inbox, spam, or another folder. Resend makes this distinction explicit in its webhook event names: email.sent means it will attempt delivery, while email.delivered means the recipient’s mail server accepted the message (Resend event types).

That difference helps narrow a support report. If your service accepted a reset request but you have no delivery event, investigate the sending path and provider events. If the recipient server accepted it, investigate placement, authentication, and the user’s mailbox. Neither case is resolved by treating “API send succeeded” as proof that the user saw the message.

Why transactional emails fail to reach the inbox

A receipt or security alert can be missing because of a problem at several points in the path from your application to the mailbox. A broken DNS record can prevent authentication; an invalid address can bounce; a confusing message can reduce trust. Distinguish those possibilities with evidence instead of changing the template, DNS, and provider at once.

  • Authentication or alignment failure: SPF or DKIM may fail, or may pass for a domain that does not align with the visible From domain. A DKIM selector lookup failure, for example, prevents the receiver from retrieving the key needed to validate the signature.
  • Domain or sending reputation: Reputation is one factor that mailbox providers can use when evaluating mail. Complaints and repeated attempts to invalid addresses are useful warning signals; stop sending to addresses that have hard-bounced or complained rather than retrying a reset indefinitely. Amazon SES supports account-level suppression lists scoped to an AWS account in a Region (SES suppression lists).
  • Unexpected sending patterns: A legitimate backlog release can change the shape of your traffic. Review application releases, retry queues, and delayed jobs when a volume change and a placement change happen together; do not assume that every burst has the same cause.
  • Message or link problems: A reset link that points to the wrong environment, an unrecognizable sender, or a receipt padded with unrelated promotion can undermine user trust. Verify the rendered message and its destination links, not only the template source.
  • Address and suppression state: A hard bounce can mean the recipient address does not exist. A complaint or unsubscribe should also affect whether your application attempts another message; make suppression state part of the send decision.

These causes are hypotheses to test, not interchangeable explanations. A placement issue limited to one mailbox provider calls for a different investigation than a DKIM failure affecting every provider.

Set up SPF, DKIM, and DMARC correctly

Authentication works only when DNS records correspond to the domains used in the message. Before editing DNS, record the visible RFC5322 From domain, the SPF SMTP MAIL FROM domain, and the DKIM signing domain in the message’s d= tag. DMARC compares the SPF and DKIM identifiers with the visible From domain; a passing SPF or DKIM result contributes to a DMARC pass only when that identifier also aligns (DMARC alignment overview).

SPF: authorize the envelope sender

SPF authorizes hosts for the HELO and/or MAIL FROM identity; it does not directly authenticate the visible From header. An SPF record is a single TXT record for an owner name, begins with v=spf1, and multiple SPF records at that name are not permitted under RFC 7208.

For a sender with one fixed IP, this illustrates the record shape: v=spf1 ip4:192.0.2.10 -all. The address 192.0.2.10 is a documentation example, not a sending IP to copy. Replace it with the actual authorized sender, or use the exact mechanism specified by your provider, and merge all authorized sources into the one record. Count DNS-query-causing mechanisms across nested includes: the SPF evaluation limit is 10, and exceeding it returns permerror.

DKIM: publish the key for the signing domain

A DKIM public key is published as a TXT record at <selector>._domainkey.<d-domain>. For example, a message signed with d=example.com and s=mail looks up mail._domainkey.example.com. A record has a tag-list structure; p= is required and contains the public-key data in base64, while k= identifies the key type and defaults to rsa. For RSA, the p= value encodes an ASN.1 DER-encoded public key. An empty p= value revokes the key (RFC 6376).

A record’s shape can be shown as v=DKIM1; k=rsa; p=<base64-public-key>, but the key itself must come from the signing provider; do not publish the placeholder. DKIM verification checks the canonicalized message body against the signature’s bh= value and checks the selected headers listed in h= against the signature. Amazon SES Easy DKIM creates and rotates a 2048-bit key for you (SES sending guide).

DMARC: align authentication with the visible From

A starting DMARC record can look like _dmarc.example.com TXT "v=DMARC1; p=none; rua=mailto:dmarc-reports@example.com". Use a real reporting mailbox you control. p=none requests reporting without asking receivers to quarantine or reject messages; use reports to identify sending sources and alignment problems before choosing a stronger policy.

Relaxed alignment allows the identifier and visible From domains to share an organizational domain; strict alignment requires an exact match. The current DMARC specifications listed by dmarc.org are RFC 9989 for the core protocol, RFC 9990 for aggregate reporting, and RFC 9991 for failure reporting, published by the RFC Editor on May 20, 2026 (DMARC specifications).

DNS edits are not an instant test: Amazon SES says changes to domain identity records can take up to 72 hours to propagate. SES domain verification lets you send from addresses on the verified domain, and requires access to its DNS settings (SES domain identities). Volanea’s domain endpoint returns the DNS records for DKIM, MAIL FROM MX/SPF, and DMARC, which gives developers a concrete set of records to publish; its guided setup also checks SPF, DKIM, and DMARC configuration (Volanea domain records, Volanea Amazon SES comparison).

Separate transactional and marketing sending

A reset link and a newsletter have different triggers and recipient expectations. Use distinct subdomains or sending streams so that operational messages and campaigns can be identified and monitored separately. For example, notify.example.com can be used for product messages and news.example.com for campaigns. This separation does not guarantee that one stream cannot affect another, but it gives the team a clearer boundary for reporting and reputation management.

Resend recommends sending from one or more subdomains to help isolate reputation, including protecting transactional email from negative recipient responses to marketing mail (Resend domain guidance). You can also separate event reporting: Amazon SES configuration sets determine where selected events go and which event types are published. The sending application specifies the configuration set for a message (SES event publishing).

Keep the message itself consistent with its stream. A receipt can contain details of the purchase and links needed to manage the account; it should not turn into a campaign with an unrelated discount. Likewise, a marketing unsubscribe should not be confused with an account-security alert. Volanea puts transactional email, campaigns, and automations behind one API, but teams still need to preserve these message boundaries in their domains, data, and workflows (Volanea).

Build sending volume and reputation safely

For a new domain or sending setup, build volume from real product activity rather than manufacturing traffic to meet an arbitrary schedule. Keep sending patterns tied to expected user actions, and stage planned migrations or releases so you can see which change corresponds to a change in events. There is no single ramp-up schedule that fits every application, recipient mix, and sending history; use your provider’s events and mailbox-provider feedback to pace changes.

Triggered mail still depends on valid recipient data. A password reset requested from an address does not make that address deliverable, and a retry loop can turn one failed request into repeated sends. Before launching a new workflow, confirm it respects unsubscribes, complaints, hard bounces, and suppression state. For SES, its account-level suppression list applies within the current AWS Region, so make sure your sending path accounts for the relevant regional scope (SES suppression lists).

We send through shared Amazon SES infrastructure at Volanea and do not offer dedicated IPs. That means our practical focus is domain setup, message streams, and the event signals available to the team—not a dedicated-IP warm-up plan. Volanea’s reputation guard slows sending before a bad list hurts a domain (Volanea Amazon SES comparison). It is a control for a risky list, not a replacement for preventing stale or unwanted contacts from entering an automation.

Write messages that support deliverability and user trust

Make the sender and subject recognizable. “Acme” from receipts@example.com and “Reset your password” tell the recipient more than an unfamiliar sender name and a vague subject such as “Important update.” Keep that identity consistent across account alerts, receipts, and sign-in links.

Keep the message focused on the event that triggered it. A reset email should identify the service, explain why it arrived, provide the correct reset link, and say what to do if the recipient did not request it. A receipt should show enough purchase detail for the customer to identify the transaction. Test links in the rendered email, including on mobile; a broken link can make a technically delivered message fail its product task.

Promotional material can distract from the reason an operational message arrived. If the user needs to act on a security alert or confirm a purchase, make that action clear before any secondary content. Volanea’s sending endpoint processes suppression checks, contact upserts, template rendering, tracking, and delivery in one pipeline (Volanea API details); we still recommend testing the final rendered message and its links as part of the application release.

Monitor deliverability and troubleshoot incidents

Instrument events by provider and message type. Amazon SES event publishing supports sends, deliveries, opens, clicks, bounces, complaints, rejections, rendering failures, and delivery delays, with destinations including CloudWatch, Amazon SNS, and EventBridge. Configuration sets control the destinations and event types for a send (SES event publishing). Treat these as different stages: an API request succeeding does not prove server acceptance, and server acceptance does not prove inbox placement.

Use a baseline from your own application instead of a universal industry cutoff. For example, alert if a provider’s hard-bounce rate doubles against its own recent baseline, then inspect whether the change is limited to a campaign, a new workflow, or all transactional messages. The threshold is a local operational trigger, not a claim about what every mailbox provider considers acceptable.

A worked incident makes the sequence concrete. Suppose password-reset delivery falls after a release that changed the sender from auth.example.com to example.com. Compare the affected messages’ authentication results and inspect the DKIM d= and SPF MAIL FROM domains against the visible From. If DKIM passes for auth.example.com but DMARC alignment is strict against example.com, the signing domain may not align; correct the domain setup or alignment policy, then send a small test and watch the same provider-specific delivery events before restoring normal traffic.

For an incident, proceed in order:

  1. Scope it: Compare API requests, server delivery events, and mailbox-provider results by message type and recipient domain. Find out whether the issue affects one template, one provider, or all sends.
  2. Check authentication: Inspect headers and DNS for SPF, DKIM, and DMARC results, including alignment with the visible From domain. Check whether a recent DNS or selector change is still propagating.
  3. Check sending and recipient signals: Review bounces, complaints, delays, and suppression events. Stop retries to hard-bounced or complained-about addresses, and inspect application queues for duplicated triggers.
  4. Compare the message: Test the affected rendered template against a known working version. Confirm the sender identity, subject, links, and environment-specific URLs.
  5. Recover cautiously: After fixing the cause, send a test to the affected provider and compare event results before resuming broader traffic. Record the root cause so the next release can catch it earlier.

Volanea provides send records, delivery, open, click, bounce, and unsubscribe statistics, alongside event logs and contact activity timelines. Those records can help connect an individual reset or receipt to the rest of the sending path, rather than treating every missing-message report as the same failure.

Choose infrastructure for reliable transactional sending

Choose based on the controls your team wants to operate and the evidence it needs when a message fails. Amazon SES is a fit for teams that want AWS API or SMTP sending and configurable event destinations. Postmark offers a transactional REST API using a server token and SMTP using its server endpoint; it also requires a confirmed sender signature and recommends domain verification, DKIM, and Return-Path authentication. Resend has a direct email API and SMTP, while SendGrid offers its Mail Send API and SMTP Relay.

ServiceSending and authentication detailsUseful operational details
Amazon SESAPI and SMTP; domain identity verification uses DNS; Easy DKIM uses a provider-managed 2048-bit key.Configuration sets can publish delivery, bounce, complaint, delay, and other events to AWS destinations. Account-level suppression is scoped to the AWS Region.
PostmarkREST API uses an X-Postmark-Server-Token; SMTP uses smtp.postmarkapp.com on ports 25, 2525, or 587 with STARTTLS.Requires a confirmed sender signature; recommends verified domain, DKIM, and Return-Path authentication. SMTP can use a server token or stream-specific SMTP token.
ResendEmail API requires from, to, and subject; SMTP host is smtp.resend.com, with username resend and API-key password.Requires a verified domain; SPF and DKIM are set up during domain verification, while DMARC is not automatic. Webhooks distinguish email.sent from email.delivered.
SendGridMail Send API accepts message details in JSON; SMTP Relay is also supported.Domain authentication uses DNS records supplied by SendGrid. Its Event Webhook documents processed, delivered, and bounced events.

The integrations differ in the specifics, not in a universal ranking. If your team is already AWS-oriented and wants to route events into AWS destinations, SES exposes that control. If you want a transactional REST API and SMTP option with sender signatures and message streams, Postmark documents both paths. Resend’s event distinction is useful for separating API acceptance from recipient-server acceptance. SendGrid provides both API and SMTP Relay plus an Event Webhook. Evaluate the workflow you will operate, not just the number of send methods.

For a unified system that also handles campaigns and automations, Volanea sends through Amazon SES and puts sending, contacts, templates, segments, and event visibility behind one API. We include suppression checks in the sending pipeline and a reputation guard, but sending is on shared SES infrastructure rather than dedicated IPs. Plans differ by monthly volume and domain count; all features are on every plan. At the monthly email limit, sending pauses, with no overage billing. For example, Volanea Growth is $20/month for 50,000 emails, while Scale is $149/month for 200,000 emails (Volanea plans and sending details).

Before moving production resets, test the integration end to end: send a test message, consume an event, confirm how a bounce affects future sends, and verify that a retry does not create duplicates. Volanea’s sk_test_ keys render and log messages but never deliver, which makes them useful for checking the integration without sending to a real recipient.

FAQ

Does the 30/30/50 rule apply to transactional email?

No. The 30/30/50 rule is discussed as a cold-email guideline, not as a transactional-email deliverability standard. A password reset is triggered by a specific user request, so the relevant checks are whether the address is valid, the message is authenticated and aligned, and the application avoids sending repeated requests or ignoring suppression state.

What SMTP login does Volanea use?

Use volanea as the SMTP username and the project’s secret API key as the password. Amazon SES SMTP has separate requirements, including a Region-specific SES SMTP endpoint and SMTP credentials, so do not reuse those instructions for Volanea. For Volanea test keys, sk_test_ messages render and log but do not deliver.

How long can DNS setup take before I test a domain?

Amazon SES says domain identity DNS changes can take up to 72 hours to propagate. Query the records and inspect an actual test message’s SPF, DKIM, and DMARC results rather than assuming that a saved DNS change is already visible to every receiver. Check alignment as well as whether each mechanism passes.

What does Volanea cost for 50,000 emails a month?

Volanea Growth costs $20/month for 50,000 emails. The plans differ by monthly volume and domain count, and features are available on every plan. Sending pauses at the monthly limit rather than generating overage billing. You can compare those details on Volanea’s plans page.

If you want to test the sending path with a small monthly allowance, start free with 1,000 emails a month.