If you're the kind of founder who reaches for AWS before you reach for a SaaS subscription, cold email outreach looks like a solved problem. Spin up SES, write a sending script, loop through a CSV of leads. A few hours of work, and you've replaced a $97/month tool with something that costs fractions of a cent per email.
That math is correct for the sending cost. It's wrong for almost everything else, and the gap is where most DIY cold email setups quietly die.
What SES actually gives you
SES is a mail transfer agent. It accepts a message and relays it. That's the whole product. It does not:
- Track whether a recipient opened the email or clicked a link
- Know if a reply came back, let alone whether that reply was interested, not interested, or an out-of-office autoresponder
- Warm up a new sending identity for you
- Stop you from burning your domain's reputation by sending too much, too fast, from a brand-new identity
- Rotate sending across multiple inboxes to keep volume per sender low
Every one of those is something a dedicated cold email tool handles by default, and every one of those is something you'll build yourself, badly, around week three, after your open rates mysteriously drop and you can't figure out why.
The part that actually bites: warmup
New sending domains and new SES identities start with effectively zero reputation. Mailbox providers watch for sudden volume from unfamiliar senders, because that's exactly what a spam operation looks like on day one. Send 500 emails from a brand-new SES identity on day one and a meaningful chunk land in spam, permanently damaging that identity's reputation before you've sent a single useful message.
The fix is warmup: gradually increasing daily volume over one to three weeks while engaging in enough back-and-forth email activity that the sending identity looks like a normal human mailbox, not a blast tool. Building this correctly means:
- A scheduler that ramps volume daily on its own curve, not a fixed number
- A pool of reply accounts that can exchange believable, varied messages with your sending identity automatically, since sending the same templated "hey, how's it going" to the same five accounts every day looks exactly as synthetic as it is
- Monitoring to catch deliverability drops early and throttle back before damage compounds
I've seen homegrown versions of this become a bigger side project than the actual cold email campaign they were supposed to support. It's a solvable problem, but solving it properly takes real engineering time, which is worth pricing against your own hourly rate before deciding DIY is cheaper.
Where SES is still the right call
None of this means SES is wrong. Owning your own sending infrastructure, rather than renting space on someone else's shared IP pool, is a genuinely better position for deliverability long term. You're not sharing reputation with strangers, and you're not locked into a vendor's rate limits. The free AWS SES cold email setup guide walks through the actual domain, SPF, DKIM and DMARC configuration if you're setting this up yourself, SES-specific gotchas included.
The honest trade-off: build the sending pipe yourself and you own SES's low cost and full control, but you're also building warmup logic, reply tracking, and reputation monitoring from scratch, or wiring together three or four separate tools to cover what a cold-email-specific platform does out of the box on top of your own SES connection. EmailQo runs on that same own-infrastructure model (Gmail, Outlook, or your own SES), with the warmup, deliverability monitoring, and reply handling already built, if you'd rather spend the engineering time on your actual product. There's a free calculator at that link if you want to run the actual cost comparison for your own volume before deciding.
The real question to ask before building this
Not "can I build this." You can. The question is whether warmup scheduling, reply-pool management, and deliverability monitoring are worth being your side project this quarter, or whether that time is better spent on the product you're actually trying to sell with these emails.
Top comments (0)