Most email-sending advice for operators is some version of “use Amazon SES if you’re cheap, use Resend if you want a nice dashboard, use Mailchimp if you don’t want to think about it.” That advice is fine until you actually look at the inbox-placement numbers — and then you find that one provider keeps quietly outperforming everyone else for newsletter sending: Postmark.
If your last three newsletters landed in Promotions or Spam at Gmail, the cheapest fix isn’t a deliverability consultant. It’s switching to Postmark. This post is the operator’s guide: what Postmark actually is, the cost math at newsletter scale, the step-by-step setup with the One Two Three Send plugin, and the honest tradeoffs vs SES and Resend.
What Postmark is, in one paragraph
Postmark is a transactional and broadcast email service operated by Wildbit (acquired by ActiveCampaign in 2022, but run as a separate product). It’s been around since 2010 and is known in the developer community for one thing: deliverability. Where most email providers let you send anything and learn about reputation problems via your bounce rate, Postmark aggressively splits transactional traffic from marketing/broadcast traffic onto separate IP pools, monitors content patterns proactively, and has the most-paranoid abuse team in the industry. The result: their inbox placement is consistently 5–15% higher than the median provider on the same lists. That sounds small until you calculate the revenue lift on a list of 10,000 paid subscribers.
Why operators pick Postmark over SES or Resend
Three reasons that show up after a few months of running:
1. Inbox placement. Postmark’s Send Score and email Insights tools show you, per send, how Gmail/Outlook/Yahoo are categorizing your mail. SES gives you “delivered” — meaning the recipient’s server accepted the message. That doesn’t tell you whether it landed in the Inbox or Promotions or Spam. Postmark tells you.
2. Bounce + complaint dashboard, free, no SNS or Lambda. SES requires you to wire up SNS topics → Lambda → your database → suppression list before you can see why a domain is bouncing. Postmark gives you that view in their console out of the box. For an operator who values shipping over plumbing, it’s the difference between a one-hour setup and a one-day setup.
3. Strict separation of broadcast and transactional streams. When you accidentally send a marketing email through your transactional stream — every newsletter operator does this at least once — Postmark blocks it before it harms your transactional reputation. SES has nothing equivalent. The savings from not losing your password-reset deliverability after one rogue announcement are real.
The cost reality
Postmark is more expensive than SES, less expensive than Mailchimp at scale, and roughly even with Resend.
| Subscribers (sending 5×/week) | Emails/month | Postmark | SES | Resend retail |
|---|---|---|---|---|
| 1,000 | 22,000 | $15 | $2.20 | $20 |
| 5,000 | 110,000 | $80 | $11 | $50 |
| 15,000 | 330,000 | $200 | $33 | $200 |
| 50,000 | 1,100,000 | $700 | $110 | $1,000+ |
The honest reading: Postmark is roughly 7× SES at any volume. That’s the price of the deliverability premium and the dashboard. For lists under ~25,000 paid subscribers, it’s the right tradeoff almost every time. Above that, you’ll start considering SES for cost reasons and accepting the operational overhead.
Postmark also has a free tier — 100 emails/month, no time limit — that’s enough to set everything up, send yourself test emails, and verify the integration with One Two Three Send before you put a card in.
Setup with the One Two Three Send plugin
Eight minutes if you have a domain in hand. The plugin uses Postmark’s SMTP interface (no custom Postmark provider needed — the standard SMTP option in OTTS handles it perfectly).
Step 1 — Sign up and create a server
Go to postmarkapp.com and sign up. Once you’re in, click Servers → Create Server. Name it after your newsletter (“Love Ireland”, “OTTS Newsletter”, etc.). Pick the “Broadcast” server type — this is critical. Postmark separates transactional and broadcast traffic onto different IP pools and reputation profiles, and broadcast is what newsletters need. Don’t pick the default Transactional unless you also want to use Postmark for password resets and order confirmations on the same site (in which case create two servers, one of each type).
Step 2 — Verify your sending domain
From the new server, go to Sender Signatures → Add Domain. Enter your domain (e.g. yourbrand.com). Postmark gives you three DNS records to add:
- One
TXTrecord for SPF (the value includesspf.mtasv.net) - One
TXTrecord for DKIM (the long signing key) - One
CNAMErecord for the Return-Path domain
Add all three to your DNS, click “Verify.” Done within 10 minutes. Skipping any of these — especially DKIM — is the difference between landing in Promotions and landing in Inbox at Gmail.
Step 3 — Get your Server API token
From the server’s overview page, click API Tokens. Copy the Server token. This token is both your SMTP username AND your SMTP password (Postmark uses the same value for both — unusual but documented).
Step 4 — Configure One Two Three Send
In your WordPress admin, go to Newsletter → Settings → Provider. Pick SMTP. Enter:
- SMTP Host:
smtp.postmarkapp.com - Port:
587 - Encryption: STARTTLS (the default)
- Username: [your Server token]
- Password: [the same Server token]
- From Email: any address at your verified domain (e.g.
hello@yourbrand.com) - From Name: your newsletter’s name
Click Test connection. The plugin will fire a test email to your admin address. If it lands, you’re done.
Step 5 — Send your first newsletter through Postmark
Generate a draft, hit Send. Within 30 seconds, the message shows up in Postmark’s Activity log with its delivery state. Click into any send to see the message metadata, bounce/complaint state if any, and (within an hour) the recipient’s email-client breakdown.
Send a test to your own Gmail and click Show original. You should see SPF: PASS, DKIM: PASS, DMARC: PASS (the last one assumes you have a DMARC TXT record on your domain — if you don’t, see Pitfall #2 below). All three passing is the bar. Anything less and your inbox placement won’t be what Postmark is capable of giving you.
Comparison vs the obvious alternatives
vs Amazon SES — SES is roughly 1/7 the price at every scale. It’s the right choice when monthly volume crosses ~250,000 emails AND you have someone on the team who can wire up SNS for bounces, run CloudWatch dashboards, and warm up dedicated IPs. For everyone smaller, the time you spend on SES infrastructure costs more than Postmark’s bill. We have a separate post on running SES for the operators who are at that scale.
vs Resend — Resend is the modern competitor: nicer API, comparable deliverability for newsletters, slightly cheaper at the lower tiers, similar at the higher ones. Postmark is older, more conservative, and has a meaningful inbox-placement edge specifically at Gmail and Outlook. If you’re building a SaaS app and want clean developer ergonomics, Resend wins. If you’re running a newsletter and the only metric you care about is open rate, Postmark wins.
vs Mailgun — Roughly comparable price at low volume, cheaper at high volume, slightly worse deliverability in our testing. Mailgun’s main advantage is they accept high-volume sends without a manual review process Postmark sometimes does on accounts approaching 1M/month. If you’re going from 100K → 1M and don’t want to negotiate, Mailgun is the easier path.
vs Mailchimp / ConvertKit / Beehiiv — Different category. These are mailing platforms (list, editor, analytics, automation, signup forms). Postmark is a sending engine. You wouldn’t compare a hosting provider to Squarespace; same here.
vs SendGrid — SendGrid is fine. The pricing tiers are less predictable above 100K, and the deliverability has slipped in independent benchmarks over the last 3 years. Use Postmark instead unless you’ve got a SendGrid contract you can’t get out of.
Pitfalls operators hit (in the order they hit them)
1. Picking the wrong server type. If you accidentally created a “Transactional” server and start sending newsletters through it, Postmark’s abuse team will flag the account within a week. Create a Broadcast server. If you created the wrong one, delete it (after migration) and start fresh — don’t try to flip the type.
2. No DMARC record. Gmail and Yahoo’s 2024 sender requirements mandate DMARC for any domain sending more than 5,000 emails/day to their users. The minimum acceptable record is v=DMARC1; p=none; rua=mailto:dmarc@yourdomain.com. Without it, Postmark’s deliverability advantage evaporates because Gmail/Yahoo discount your sender reputation. Add the TXT record before you send to a real subscriber.
3. Sending from a free email address. If your From Email is yourname@gmail.com Postmark will refuse to send. The From must be at a domain you’ve verified in Postmark. Most operators trip over this on the first test send and lose 15 minutes wondering why nothing fires.
4. Suppression list surprises. Postmark’s automatic suppressions are aggressive — they’ll suppress anyone who marks one email as spam, anyone whose domain bounces hard, anyone who clicks unsubscribe via Postmark’s unsubscribe handler. This is good for reputation but means you need to be intentional about list re-engagement. The Suppressions tab in Postmark’s console shows the full list; you can manually remove someone if they email you to say they want back on.
5. The 10MB attachment limit. Postmark won’t accept emails over 10MB total (HTML + inline images + attachments). Most newsletters are under 200KB so this never matters — except when someone uploads a 4K-resolution hero image and the rendered email balloons past the limit. Compress images to under 1MB before they go into the email body.
6. Approval delay on first 1M-email month. When you cross approximately 1 million emails in a billing month, Postmark’s account review team will email you to verify you’re a legitimate operator. Reply within 48 hours with proof of the list source (signup form URL, growth pattern). Don’t ignore the email — they’ll throttle the account if you do.
Long-term maintenance
Less than SES, more than a pure SaaS like Beehiiv.
Daily/weekly: nothing. Postmark is set-and-forget for the first 6 months. The activity log emails are a useful occasional read.
Monthly: review the Suppressions tab. Look for patterns — if a corporate domain is on there 50 times, you’ve probably been blocked at the SMTP level by their gateway and need to get IT-whitelisted. Look at bounce-rate trend in the dashboard; if it’s creeping past 3%, you have a list-quality problem worth fixing.
Quarterly: review your DMARC reports (the rua address gets weekly XML aggregates from Gmail/Yahoo/Microsoft). Tools like Postmark’s DMARC monitor (free) parse these into a readable dashboard.
Annually: rotate your Server token. Settings → API Tokens → “Generate new token”, paste the new one into One Two Three Send’s SMTP settings, deactivate the old one a week later.
When Postmark is the wrong choice
If you’re sending fewer than 5,000 emails a month, Postmark’s $15/month minimum is overkill and you should stay on Resend or Mailchimp’s free tier until volume justifies the spend.
If you’re sending more than ~500,000 emails a month and have technical capacity, Amazon SES becomes meaningfully cheaper. The break-even for “the time SES costs me to set up” vs “the dollars Postmark saves me” lands somewhere around the 200K–300K/month volume mark for most operator teams.
If your audience is concentrated in countries where Postmark’s IP pools have weak reputation (a few specific Asian and African corporate-email gateways are the usual culprits), test before committing. Postmark’s free tier of 100 emails/month is enough to verify your top-10 corporate domains accept mail from them before you migrate.
For everyone else: Postmark is the boringly-best transactional and broadcast email provider available to newsletter operators in 2026. Pair it with One Two Three Send and you have an inbox-placement-optimized newsletter stack that takes 8 minutes to set up and runs without intervention for 6 months at a time. That’s what infrastructure is supposed to do.