· 6 min read
SES deliverability for newsletters: the problem nobody tells you about
Amazon SES is cheap but its deliverability for newsletters is mediocre at best. Here is why, what actually goes wrong, and what to use instead when you have 20k subscribers and no revenue.
Amazon SES costs a dollar per ten thousand emails. At that price every newsletter operator discovers it, and at that price most of them stay — even after noticing that open rates are suspiciously low and the important readers never reply. The problem is not that SES cannot deliver mail. It can. The problem is that SES delivers your mail from the same infrastructure that delivers millions of other people's mail, most of which is transactional, and inbox providers know this.
What deliverability actually means
Deliverability is not "did the server accept the message." SMTP acceptance rates for SES are near 100% because the mail was accepted — into a spam folder, a promotions tab, or a quarantine the reader never checks. The number that matters is inbox placement: of the messages accepted, how many landed in the primary inbox where a human reads them?
There is no API for this. You cannot query Gmail and ask "where did you put my email?" The only honest signals are:
- Open rate. Not a measurement of interest — a measurement of visibility. A 15% open rate on a list that opted in yesterday is not reader apathy; it is mail that was not seen.
- Reply rate. Replies are the strongest positive signal to inbox providers. Mail that generates replies trains the filter to show future mail from you.
- Complaint rate. The percentage of recipients who click "report spam." SES publishes yours; keep it under 0.1%.
- Bounce rate. Hard bounces (invalid addresses) above 2% damage your sender reputation faster than anything else.
Why SES is mediocre for newsletters
SES is an infrastructure product. It gives you an SMTP endpoint and a sending domain. It does not give you:
- A dedicated IP with a warmed reputation. On SES's shared IPs, your reputation is pooled with everyone else on the same IP. One bad sender tanks the whole pool.
- Automatic list hygiene. SES does not suppress unengaged subscribers, verify addresses before you send, or decay your list for you.
- Feedback loops with inbox providers. Some ESPs negotiate direct feedback loops with Gmail and Outlook. SES does not.
- Send-time optimisation. Sending 20,000 emails at once from a cold domain is how you trigger rate limiting and throttling.
You can fix all of these on SES. Dedicated IPs cost $24.95/month each and require careful warming. You can build list hygiene yourself. You can implement DKIM, DMARC, and SPF (which you should do regardless). But "can" and "will" are different words, and the whole point of a newsletter is writing, not operating mail infrastructure.
The actual cost comparison for 20k subscribers, 4x/month
80,000 emails a month. Here is what each provider charges, what you get for it, and whether the deliverability is meaningfully better than SES.
| Provider | Monthly cost | Dedicated IP | List hygiene | Send-time optimisation | Deliverability reputation |
|---|---|---|---|---|---|
| Amazon SES | ~$8 | $24.95/ea extra | DIY | DIY | Shared pool — mediocre |
| Resend | $80 (Pro) | Shared | Basic | No | Good — built on SES but manages reputation |
| Loops | $99 (Starter) | Shared | Yes | Yes | Very good — newsletter-first |
| Buttondown | $29 (Standard) | Shared | Yes | No | Good — uses SES but curates senders |
| Beehiiv | Free (up to 2.5k) / $42+ | Shared | Yes | Yes | Very good |
| ConvertKit (Kit) | $50 (Creator, 1k+) | Shared | Yes | Yes | Very good — newsletter specialist |
| Mailchimp | $45 (Essentials) | $$$$ | Yes | Yes | Good on shared, varies |
| Postmark | $15 (10k/mo) | Dedicated included | Yes | No | Excellent — but transactional-first |
Prices as of September 2026. Check each provider's pricing page — they change quarterly.
The honest recommendation
If your newsletter makes no money and you send plaintext to 20k people four times a month, here is the decision tree:
Stay on SES if you have already set up DKIM/DMARC/SPF, your open rates are above 40%, you enjoy the infrastructure work, and you do not mind building your own unsubscribe handling.
Move to Buttondown if you want the cheapest real improvement. $29/month, handles compliance, clean editor, and the developer who runs it is reachable. It uses SES under the hood but manages sender reputation at the account level, which is the thing you cannot do yourself on raw SES.
Move to Loops if you want the best deliverability for a personal newsletter at a reasonable price. $99/month is expensive for something that makes nothing, but inbox placement is noticeably better than SES and the product is designed for exactly this use case.
Move to Beehiiv if you want to eventually monetise the newsletter. Free tier up to 2,500 subscribers, paid tiers are competitive, and the growth tools (referral programs, recommendations) are built in.
Do not move to Mailchimp. The free tier is gone, the pricing is opaque past the first tier, and the deliverability on shared infrastructure is not better than what you have. It is the default because it was first, not because it is best.
What actually improves deliverability
Regardless of provider:
- Authenticate your domain. DKIM, SPF, and DMARC. Every provider makes you do this; if yours does not, leave.
- Clean your list. Remove anyone who has not opened in 90 days. A smaller, engaged list delivers better than a large, dead one.
- Send consistently. Same day, same time, every week. Inbox providers learn patterns.
- Write plaintext or near-plaintext. Heavy HTML, lots of images, and tracking pixels are spam signals. A plaintext email from a person looks like a person wrote it.
- Ask for replies. Put a question in the email. Replies train the filter.
- Use a real from address that you check. Bounce-backs and replies should reach you.
The cheapest path from "SES and not sure anyone is reading" to "confident they are" is: set up DKIM/DMARC on your current domain, export your list, import it into Buttondown or Loops, send the next issue from there, and compare open rates. If they double, the answer is clear. If they do not, the problem was never deliverability — it was the writing.
The tool below
We built a deliverability comparison tool that shows the table above in an interactive format, with links to each provider's pricing page and deliverability documentation. No affiliate links, no tracking, no account required.