Sendy - Send Newsletters via Amazon SES | Hacker News Reader

Sendy - Send Newsletters via Amazon SES

sendy.co

67 pointsdwynings60 comments

A small point, but your API doesn't appear to use any kind of authentication/authorization:

$handle = fopen("$sendy_url/subscribe/$sendy_email/$sendy_list/".str_replace(' ', '%20', $sendy_name).'/true', "r");`).

In theory, anyone who knows you are using Sendy (I'll assume looking at the headers in the emails sent would reveal that) could subscribe/unsubscribe any email address and possibly wreak havoc. You should secure that somehow: either an authorization scheme or a secret or SOMETHING.

If this is true, it's like a 'fun mode' for an API.

Other than spammers, who is concerned about sending email newsletters cheaper? Your service looks like a great tool, but I'd encourage you to ask your early users why they like using you. I'd guess that "It's cheaper than mailchimp" won't be the answer you hear.

+ You don't really want to tout yourself as the low-cost-option. You have nowhere to go but down.

Paying mail chimp $150 a month to manage sending to a list of 25k users or more is a cost many folk would prefer not to have. I doubt this is any sort of take over the world project for the author, but it's certainly a useful project that I think many folk would be happy to compensate him for.

Sending to 1,000,000 subscribers costs $55,000+ per year on MailChimp and other premium services. A cheaper option with the same deliverability would be extremely attractive to anyone who doesn't like setting their money on fire!

Exactly this... People seem to jump on the bandwagon of SaaS services and don't really realize that, as a business, you should be trying to minimize expenses and maximize profit. Granted there is flexibility here in the products you use -- ie, the feature/cost tradeoff needs to make sense, but the fact remains, every $1 of money the company saves helps the company last longer, increases its chance at a future, and benefits both the employees and investors (if there are any). People seem to get sidetracked in our industry a lot of times on this startup mantra that makes people feel like their company is a failure if it doesn't buy everyone an Aeron, have a quirky office in SOMA, and utilize all the hot startup tech services that exist.

Your biggest expense as a company is employees. Every purchasing decision that won't keep you alive for another month is a decision that isn't worth thinking about.

"I could switch from A to B and save $100 a month". Each human being in your startup costs $5000 - $10,000 a month in salary excl. benefits. Saving $150 a month on mailchimp is chump change.

Definitely true... I was speaking to the example of spending $50k/yr on mailchimp with 1,000,000 subscribers. If you're only dropping $150, then it's not as big of a deal.

"who is concerned about [doing something] cheaper"??

Really? That's your question? And by "cheaper" it is meant "99.0%-99.5% cheaper"?

I'm skeptical that the main value of the product to end users is that it's cheaper. Do you own an iPhone? Did you buy it because it's the cheapest way to have a phone in your pocket?

On hacker news, people seem to care about pinching pennies on marketing emails, but in the real world where people have real businesses that make real money, the cheapest is not going to win the fight.

> On hacker news, people seem to care about pinching pennies on marketing emails, but in the real world where people have real businesses that make real money, the cheapest is not going to win the fight.

Oh, tell that to envato[1]. Most companies don't have deep pockets and unlimited budgets.

[1] http://graphicriver.net/ http://photodune.net/ People seem to buy stuff there even though istcokphoto exists.

You're comparing newsletter delivery to iPhones? I didn't say price is the only parameter. I was responding to the inane comment that people don't care about cost.

Which real world do you live in? Real businesses spend a good percentage of their efforts in cutting costs. And if they have a large consumer base they push newsletters too, then yes, as an agency, I can tell you that a lot of clients will make the switch.

There a bunch of open source self-hosted newsletter systems doing well (oh, and they're mostly subpar). I like the direction in which sendy is heading and with the amazon ses edge to it, its a great great idea. If sendy would throw in a few more features like multi-user login and access control, I'll make the switch and will convince a few of my clients too.

We're talking MailChimp/Campaign Monitor VS Amazon SES, if we're comparing MailChimp/Campaign Monitor VS sending emails from you own server as a cheaper alternative, then there is no reason for Sendy to exist.

Yes Amazon SES is way cheaper, but it does not mean it's of lesser quality than other more expensive services. Deliverability rates are comparable and at times, higher when compared against Campaign Monitor in our tests when building Sendy.

We see high potential in SES, such great quality at such low price. The only drawback is there is no UI for newsletter campaigns. So we built it and hopefully the general public who'd like to send newsletters but aren't too technically inclined (or are too busy to figure) can take advantage of Amazon SES through Sendy.

The only reason why it's a one time fee of $40 is because it's self hosted.

Disclaimer: I use MailChimp for Hacker Newsletter and Wayback Letter

I guess it depends what type of newsletter you're sending out, but pricing is hardly the reason I would pick one solution over another when it comes to sending out emails. The features and deliverability that someone like MailChimp provides is really invaluable if your newsletter is an important part of your business.

That being said, I always like the idea of a self-hosted solution as an option in any segment and I'm sure there is a market for something exactly like this.

Mailchimp gets brutally expensive at the higher tiers, especially if you have a large list but only send 1 or 2 newsletters a month.

We switched to sendgrid for our newsletters even though their interface is not as nice. Their pay-per-delivery pricing fit our model much better than the pay-per-subscriber.

This looks great!

If I buy it, are updates included? Do you have a demo we can play with? Can we send multipart emails?

Thanks!

Yes updates are included until the next major version, just like desktop apps. And yes you can send multipart emails. :)

Did you know that Amazon SES is built in Romania ? http://romania.amazon.com

Also on the same street a division of SendGrid works too. Iasi, Romania a little Silicon Valley :)

For any residents of Iasi on this thread, I will be heading there in October to speak at our development center.

As the founder of a new email company, actually sending emails is the baseline and assumed part of any ESP service. Whilst deliverability is absolutely important it is assumed that a professional service has this covered.

The real value I have seen is in managing the world beyond deliverability, especially in composition and content creation, design, subscriber management, better workflows for editors/developers/designs, innovative data-friendly tools, and above all features that drive metrics.

How is Mailchimp $200 / 10,000 emails? With their monthly plans you send unlimited (to a limited number of people in your list).

I imagine it's based on their pay-as-you-go plan.

True, using the pay as you go price is slightly bogus because I'm sure most people sending newsletters would be better off with a monthly plan. But it depends on how frequently you send messages, I guess. The monthly price is still $75/mo (and at 10,001 addresses it's $150/mo).

Why should I use this over Mandrill? http://mandrill.com/

That's transactional. Sendy is bulk and should be compared with MailChimp, not Mandrill.

Mandrill's "mostly RESTful" API isn't remotely RESTful and only knows the POST HTTP verb.

Wow, Mandrill is significantly cheaper than postmarkapp who we use now. Anybody using Mandrill in production?

Going to use it to send a couple of thousand of emails with attachments like, tomorrow. Up to now it works great in my tests.

I noticed a reference to .htaccess in the Get Stared Guide. Can this be configured to run on nginx?

I can't speak for the app, but I've yet to discover a .htaccess config that couldn't be converted to nginx.

Other than rewrite rules, I don't have any experience with this, so it'd be nice to know what's in the .htaccess before buying.

The .htaccess contains rewrite rules for pretty urls.

This is an awesome idea.

I might be wrong but I thought Amazon SES was meant for transactional emails only. I browsed their FAQ and they never talk about email marketing. I don't think they will let you send multiple campaigns with an audience of more than 100.000 contacts.

http://aws.amazon.com/ses/

> Amazon Simple Email Service (Amazon SES) is a highly scalable and cost-effective bulk and transactional email-sending service for businesses and developers.

They'll do bulk. In fact, they will do all of the things normal bulk senders will do... dedicated IPs and such.

They don't have any UI though, which will make it hard for most email marketers.

Would be nice to get a sense of how this handles large volumes. I would be wary to try this for anything serious as it might not be set up properly for large volumes 100k, 500k etc. Maybe speaking to this concern would help business.

Very cool. I've been waiting for someone to add a layer on SES for sending newsletters instead of transactional emails.

The only bummer with SES for serious newsletter folks is that you can't get a dedicated IP address.

Some argue that dedicated IPs aren't all that important: http://blog.postmarkapp.com/post/14127210172/the-false-promi...

Interesting link. I agree that perhaps the value is sometimes overstated and I agree that we are moving towards domain-based reputation, but I've personally been bitten by other customers getting my shared IP on a blacklist.

While it's true that ISPs and spam filters will take the reputation of neighboring IPs into account, surely having a dedicated IP insulates you from the actions of others better than sharing the exact same address as them.

I agree with the bottom line, though: use an ESP that has a low tolerance for shenanigans.

Presumably, systems like SES have direct lines to the major ISPs, and religiously monitor the blacklists. I'd imagine any affected IPs get pulled out of the pool within a minute of going on any of the major lists.

If the provider is responsive to spam, you've probably got more benefit from their IP block being considered very safe than you'd get from a dedicated IP, and having a dedicated IP doesn't protect you from your ISP's entire range getting flagged if someone else in it goes on a rampage.

Disclosure: I used to work at Postmark.

ISPs are way more interested in maintaining relationships with ESPs that keep junk from getting to them in the first place than the ones that "quickly fix once the damage is done". It's one case where "asking for forgiveness" isn't preferred over the more careful approach.

SES is more like outsourcing sendmail on your own server so you don't carry the burden of maintaining/securing it, except that you don't have control over it when something goes wrong and you have the same lack of customer support if you don't know how to fix it. If you value your messages getting to the inbox, it's the least desirable option. For sending low-value emails, though, go nuts!

I'm not sure what this means. SES focuses on deliverability.

If you're looking at spending $40 on an email solution, then chances are good that a dedicated IP address isn't something that you need.

Hey guys! Great to find a discussion here.

Lots of people has been requesting a demo, we'll definitely put this up.

And yes, free updates are included up to the next major version (just like desktop apps).

I like the looks of this.

Any chance of a read-only demo to poke around in?

I second that, I like the product and would love a demo.

I'd like to see some actual high-volume tests done with SES, anyone got any links to something like this? I've Googled — nothing.

Very cool. Would have been supercool if it could have all been implemented in JavaScript and hosted on S3 (with Dynamo for the list storage), though. Having to host a server makes it less appealing for me.

How do you propose you'd track e-mail opens in such a setup?

Use an image pixel on s3 + s3 access logs.

Can't reply below- but you can use Elastic MapReduce or an EC2 instance I can boot up when I want to send mail. The point I'm getting at, is if I need to keep another LAMP setup running 24/7 ($10/mo), I might as well just save myself the headache and spend the extra $10/mo to use MailChimp.

OK, and what goes through the logs to make the pretty graphs?

Either way you'd need an EC2 instance doing some work. No way to do it with just S3 and Dynamo.

Yes you can, you can access Dynamo directly from the client. Javascript goes through the logs to make pretty graphs, it's fairly simple. That this app is in php is kind of a disappointment.

It's not that I don't want an instance doing some work, it's just that I don't want it running 24/7 (expensive) or having my opens/clicks depend on it. Much rather only have to pay for an hour of CPU time and do the processing the background.

Why would you need a new instance for it?

Host it on phpfog.com for free. (I'm not sure if it's possible to host PHP on heroku yet.)

Actually, it is possible to host PHP on Heroku. For some reason, it is not particularly well advertised (unlike the Python/Node support) but if you push your PHP files they should do the right thing: http://stackoverflow.com/questions/7947499/deploy-php-to-her.... I believe this is what they use for their automatic Facebook integration (see https://github.com/heroku/facebook-template-php).

This would make an ideal demo app for Elastic Beanstalk.

Reply on news.ycombinator.com