Review
Amazon SES review: the best email infrastructure with the worst product experience
We've watched thousands of SES events flow through our platform, so we know exactly where SES shines and where it hurts. The short version: 5/5 infrastructure, 2/5 product experience. Here's the long version — the good, the bad, and the 2am CloudWatch sessions.
Full disclosure
SendOps builds on SES — that's why we know it this well. Our product exists because of the gaps this review describes, which gives us both unusual insight and an obvious interest. The review below is our honest assessment; the last section is where we pitch.
The Verdict
Two products in one: a 5/5 engine, a 2/5 experience
Reviewing Amazon SES honestly means holding two opinions at once. As sending infrastructure — the machinery that accepts your API call and negotiates delivery with Gmail, Outlook, and every mailbox provider on earth — SES is about as good as it gets: absurdly cheap, effectively unbreakable, and capable of world-class deliverability. As a product you sit in front of — the console, the visibility, the onboarding, the help when things go wrong — it trails every managed ESP on the market, and the gap has barely narrowed in years.
That split is not an accident. AWS sells SES as a primitive, like S3 or SQS. The assumption baked into every design decision is that you'll bring your own tooling. Judge it by that standard and it scores brilliantly; judge it as an email product and it scores poorly. Both judgments are correct.
| Dimension | Score | Notes |
|---|---|---|
| Deliverability capability | 5 / 5 | Excellent — if you manage your own reputation. SES is infrastructure, not magic. |
| Price | 5 / 5 | Unbeatable. ~$0.10 per 1,000 emails; roughly $100/month at 1M sends as of July 2026. |
| Reliability & scale | 5 / 5 | AWS-grade. The same infrastructure that sends Amazon's own mail. |
| API & docs | 4 / 5 | Solid and complete, in the verbose AWS house style. SDKs for every language. |
| Console & tooling | 2 / 5 | No real dashboard, no message search, no template workflow. You build or buy all of it. |
| Onboarding | 2 / 5 | Sandbox by default; production access requires a written request AWS can deny. |
| Support | 1 / 5 | None without a paid AWS support plan. Sending problems are yours to debug. |
| Marketing features | 1 / 5 | Essentially none. No campaigns, audiences, automation, or analytics. |
Scores reflect our experience operating on SES as of July 2026, weighed against managed ESPs like SendGrid, Postmark, and Mailgun.
Where It's Excellent
What SES is genuinely great at
Deliverability capability. SES can land mail in the inbox as reliably as any premium ESP — the shared IP pools are well-policed, dedicated IPs are available when you outgrow them, and mailbox providers treat well-run SES senders exactly like well-run anyone-else. The crucial word is capability. SES gives you the pipes and the reputation ceiling; it does not manage your list hygiene, your engagement, or your authentication for you. Senders who do that work get excellent placement. Senders who don't get exactly what they've earned — and then blame SES. Our SES deliverability guide covers the work in detail.
Reliability and scale. This is AWS. SES has absorbed every traffic spike we've thrown at it and every one we've watched customers throw at it. There is no volume tier where SES becomes the bottleneck — it's the same machinery Amazon uses for its own mail. Multi-region availability, predictable latency, and the kind of uptime record that makes status-page checking feel quaint.
Price. Around $0.10 per 1,000 emails as of July 2026, with no platform fee, no contact-count pricing, and no annual contract. Nothing else on the market is within an order of magnitude at volume — more on this in the pricing section below.
API and documentation. The v2 API is complete and consistent, SDKs exist for every mainstream language, and IAM gives you permission granularity no ESP API key model can match. The docs are classic AWS: everything is documented, nothing is concise. You will find the answer; you will read four pages to get it. We call that 4/5 — verbose, but trustworthy.
Where It Hurts
What SES is genuinely bad at
Everything below is a product gap, not an infrastructure flaw — but if you're the person running email, the gaps are what you live with day to day.
No real console
The SES console shows account-level reputation metrics and a verified-identities list, and that's roughly it. There's no sending dashboard worth the name, no per-recipient history, no message search, and no template editing workflow. As of July 2026, "what happened to the email we sent Jane on Tuesday?" is a question the console simply cannot answer.
Visibility is DIY
Every real SES operator we know has a version of the same story: it's 2am, a customer says they never got their receipt, and you're stitching together CloudWatch metrics, SNS notifications, and application logs to reconstruct one message's fate. SES emits every event you need — deliveries, bounces, complaints, opens — but you build the pipeline that catches, stores, and searches them.
Onboarding friction
Every account starts in a sandbox that can only send to verified addresses. Escaping requires a written production-access request that AWS reviews — and sometimes denies with a boilerplate response and no explanation of what to fix. We've seen legitimate senders bounce off that wall twice before getting through. Our SES setup guide covers how to write a request that passes the first time.
No support, no marketing features
Without a paid AWS support plan, there is no one to ask when sending breaks — the "did that email send?" support ticket doesn't exist because there's no one to file it with. And SES ships zero marketing features: no audiences, no campaigns, no automation, no analytics. It sends what your code tells it to send, full stop.
None of this is hidden — AWS is upfront that SES is a building block. But teams comparing SES's price to an ESP's price often don't price in the weeks of event-pipeline work and the permanent on-call burden that come with the discount.
Pricing Reality
The pricing reality check
This is the number that makes everyone tolerate the rough edges. As of July 2026, SES charges about $0.10 per 1,000 emails. At one million emails a month, that's roughly $100. Comparable volume at managed ESPs typically runs $650–$1,200+ per month, and marketing-oriented platforms that price per contact can go far higher.
| Monthly volume | Amazon SES (approx.) | Typical managed ESP |
|---|---|---|
| 100,000 emails | ~$10 | ~$100–$250 |
| 1,000,000 emails | ~$100 | ~$650–$1,200+ |
| 10,000,000 emails | ~$1,000 | Custom contracts, usually $3,000+ |
Approximate US-region figures as of July 2026; ESP figures are typical published tiers for transactional-focused providers. See our full Amazon SES pricing breakdown for attachments, dedicated IPs, and the fine print.
The honest caveat: the sticker price isn't the whole cost. If you build your own event pipeline, dashboard, and suppression handling, you're paying in engineering time instead — and that bill can dwarf the savings at low volume. The economics genuinely favor SES when volume is meaningful or when you layer existing tooling on top rather than building from scratch.
Who It's For
Who should (and shouldn't) use SES
SES is a great fit if you…
- Send meaningful volume and the cost gap matters to your margins
- Already run on AWS and are comfortable with IAM, SNS, and the console
- Want to own your sending reputation instead of renting a vendor's
- Have (or can buy) tooling for events, suppression, and monitoring
Pick a managed ESP instead if you…
- Want zero AWS involvement — no IAM, no sandbox request, no console
- Need vendor support SLAs and a human to call when delivery breaks
- Send low volume, where the savings never repay the setup effort
- Want deliverability managed for you, not just made possible
If you land in the second column, that's a fine outcome — managed ESPs exist for good reasons, and paying for support and polish is a legitimate trade. Our Amazon SES alternatives comparison walks through the strongest options and where each one wins.
Fixing the Gap
Fixing the tooling gap
Almost everyone who commits to SES ends up solving the tooling gap one of three ways:
Build it yourself
Event destinations into Kinesis or SNS, a datastore, a dashboard, suppression logic, alerting. Full control, and the most common choice at companies with platform teams — but it's real, ongoing engineering work that has nothing to do with your product.
Self-host a frontend
Tools like Sendy put a campaign UI on top of SES for a one-time license fee. Cheap and proven for newsletters, but you're now hosting, patching, and scaling a PHP app — and transactional visibility is still largely on you.
Key-paste SaaS
Hosted platforms where you paste in SES credentials and get a UI. Convenient, but scrutinize what the keys can do and where your mail actually flows — some route through the vendor's account, which quietly gives back the reputation ownership you chose SES for.
Our approach — this is the pitch
SendOps is a fourth option: a platform layer that runs against your own SES account. You keep SES's pricing, your quota, and your reputation; SendOps adds the missing product — message search, delivery dashboards, suppression management, templates, broadcasts, and alerts. We built it because we kept living the 2am CloudWatch story ourselves. There's a free plan if you want to judge it against the alternatives above.
FAQ
Frequently asked questions
Is Amazon SES good?
As sending infrastructure, yes — arguably the best available. SES offers AWS-grade reliability, excellent deliverability potential, and pricing around $0.10 per 1,000 emails as of July 2026. As a product, it's rough: there's no real dashboard, no message search, no template workflow, and no support without a paid AWS plan. It's a 5/5 engine wrapped in a 2/5 experience.
Is Amazon SES worth it?
For most teams sending real volume, yes. At 1 million emails per month, SES costs roughly $100 versus $650–$1,200+ at managed ESPs as of July 2026 — savings that compound every month. The trade is that you take on work an ESP would do for you: reputation monitoring, bounce handling, event tracking, and tooling. If you can absorb that work (or layer a tool on top), SES is worth it. If you want a vendor to own deliverability end to end, it isn't.
What are Amazon SES's biggest weaknesses?
The console and tooling. SES has no meaningful sending dashboard, no way to search for an individual message, no template editing workflow, and no built-in suppression insight beyond the basics. Answering 'did that email send?' means wiring event destinations to CloudWatch, Kinesis, or SNS yourself. Add sandbox onboarding friction and no support without a paid AWS support plan, and the weaknesses are all product experience — not infrastructure.
Is Amazon SES good for beginners?
Not really. Every account starts in a sandbox limited to verified recipients, and getting production access means writing a request AWS can deny. After that, you're expected to configure DNS authentication, event destinations, and bounce handling on your own, with no support unless you pay for an AWS support plan. Beginners who want to send email today are better served by a managed ESP or by putting a management layer on top of SES.
Is Amazon SES good for email marketing?
Not by itself. SES is a raw sending API — it has essentially no marketing features: no audience management, no campaign builder, no visual templates, no automation, no analytics. Teams that want SES pricing for marketing email typically pair it with a layer on top, such as self-hosted tools like Sendy or a managed platform like SendOps that runs against your own SES account.
Get Started
Keep the 5/5 engine. Fix the 2/5 experience.
Connect your own SES account and SendOps adds the dashboard, message search, and alerts that AWS never built — free to start.