Hacker News new | comments | show | ask | jobs | submit login
SendGrid Responds to Amazon's Simple Email Service (sendgrid.com)
64 points by joshfraser 2512 days ago | hide | past | web | favorite | 42 comments

Customer of SendGrid as well and we're shelling out $400/mo with them. Definitely planning on taking a look at SES.

SendGrid is a nice service, but we've had several nagging annoyances:

1) The credentials to send emails are the same as to login and change plan settings. Seriously? Impossible to lock down the account on a large team from a security perspective. Makes their subuser management feature totally pointless.

2) Our dedicated IP got blocked by an RBL and we had to go to them about it. If I'm paying them for the pain of deal with email crap, I expect them to monitor their own logs and detect when an IP is blocked and come to me, not the other way around. And it wasn't just our IP... the whole class B our IP was on was blocked.

3) Their default behavior for bounces is once it fails, they never even attempt delivery to that address for future messages. They accept the messages our app sends, but just drop them. We had a DNS issue for a few days cause daily email alerts to fail for several big customers and this behavior bit us hard. Wasn't well documented. Come to find out we had a lot that were being blocked.

4) Website is horrible. "Recent Activity" section says last bounce was 8 days ago, but when you drill down to the "Email Activity", there were several as recent as a half hour ago. Clicking the buttons to clear all entries on the email lists don't do anything. Site is just slow and feels overly bloat. Almost feels like a template.

1) We're aware the credentials thing - it's a problem, and we're working to resolve. Here is a community discussion with more details: http://send.gd/CLf

2) We do our best to monitor blacklists and proactively notify customers as soon as we detect their placement on an RBL. We also recognize that some RBL information may get past us in certain instances. To address this, we're increasing support staff by 66% next Monday. Finally, it should be noted that some RBL's are legit, while others are, frankly, scams and exist for the wrong reasons.

3) Some things we offer to address this issue: set your own policy for bounce removal and remove them as often as you like; disable bounces so that we don't track bounces at all; get bounce notifications - as soon as the bounce occurs, you can remove it over our Web API.

4) This example could be account specific (please contact support@sendgrid.com so we can look into your account and work to debug). However, it could be more general - we realize our website is far from perfect and are working daily to make improvements. We invite any and all bug/feedback/feature reports so we can investigate, address, and deploy the fixes and enhancements that are most important to users.

My company sends nearly 3MM messages a month through SG. We're largely happy, but I concur on each one of these pain points.

Boy, those are serious issues, enough to keep me from seriously investigating SendGrid.

This isn't really a response. Only two lines relate to this at all: "We’ve heard the news, read the blogs and monitored the tweets. And as the thoughts, reactions, and sentiments roll in, we’re plugging away at what we do on a daily basis: delivering on the future of email."

This is just a post saying what they do, how they do it, and why they're good. It has nothing to do with Amazon SES at all.

    > This is just a post saying what they do, how they do it, and why 
    > they're good. It has nothing to do with Amazon SES at all.
The implied point I got from it is that all those things they do and are good at is what Amazon doesn't do. Somewhat like how Apple doesn't do a point-by-point comparison of the iPhone vs. Android -- they just let the iPhone features speak for themselves.

You're dead on with that interpretation :)

Disclaimer: happy customer of both SendGrid and MailChimp.

The risk of one of the Borg companies doing this to you at any time is yet another reason why you should not attempt to compete on price, and relatedly why you should probably not seek the business of equisitvely price sensitive customers.

To be fair, the price isn't just slightly lower - it is nearly an order of magnitude lower. Sending 500k messages a month with Amazon is $50, and with SendGrid it's $400.

If you send that level of email a month (which is easy to do if you send a weekly newsletter to a few hundred thousand users) you're now talking about saving yourself a few grand a year with Amazon.

Given the "switch" can be as simple as changing some SMTP settings, this is a big deal. I don't think the people who jump ship are necessarily hyper-price-sensitive customers. Amazon is just turning this service into a commodity.

Amazon SES doesn't have an SMTP interface.

You can, however, configure an MTA to deliver through SES by delegating delivery to a Perl script (e.g. http://docs.amazonwebservices.com/ses/latest/DeveloperGuide/...).

While the pricing isn't comparable, neither is the service, assuming that they're being completely honest in their post.

You raise a great point. I'd really appreciate a little more input on the inboxing side of the deal though. We could say that SendGrid has that going for it, so I'm wondering if there's something we can do about that.

As mentioned elsewhere, Amazon is one of the largest email senders in the world. They probably know a thing or two about deliverability.

While I agree on principle, Amazon is also a giant merchant who should know a thing or two about payments, and their Amazon Flexible Payment System is a horrible beast to work with.

Sometimes the "massive online store" domain knowledge doesn't seem to transfer over to the AWS team.

This is an extract from "Do More Faster" wrote by Isaac Saldana, the CEO of SendGrid:

"When I offered dozens of companies SendGrid for $100 per month, they all said yes. I raised the price to $300 per month, and they all said yes. $500?—yes. Today, we are working with hundreds of companies, including well-known ones like Foursquare, Gowalla, and GetSatisfaction. When you're selling a solution to a problem and you find that nobody is saying no to your prices, you've found some serious pain. We're building SendGrid to solve a very specific problem that I discovered just by paying attention."

I guess it's time to change...

I'm a customer of sendgrid but I have to admit not a really happy one (the website is shitty, the search is working half of the time, and I have the 20$ plan with heroku which mean that my mails are send about 30min after my request because apparently on shared IP we need to wait...)

What bothers us is that you are "not a really happy [customer]", yet your reasoning is understandable. Your website issue is covered above in our response to @krobertson. With regards to Email Activity Search: we’re working on making this feature more stable (we’ve brought on a new web architect to help as well) - also, we probably should have released it to a private beta group for testing first. We will be taking a new beta testing approach going forward (planning it now for our new newsletter app). For Heroku: we're close to having a fresh API integration with them, which will be a huge improvement. Future, similar integrations with other hosting services will be much better with a seamless integration process for them and their users. It’s a top priority item for us.

This was a good response, if a little arrogant. But they have a right to be arrogant. SendGrid is an amazing story. The execution has been exciting to watch from afar. A great example. I have been a customer since they launched.

My issue is the pricing model, where you pay for 50k (or whatever) emails "per month" but don't get them. If I send 45k one month, and then 55k the next, I have to pay more for those extra 5k emails. I will pay a premium for a premium service, but this is irritating. Irritating enough that I have been monitoring the space waiting for a viable competitor. And with this policy, they are certainly not earning any loyalty on my part. I would leave in a heartbeat.

Have you checked out our "pay for what you use" model at http://postmarkapp.com?

Postmark do not allow bulk email, such as newsletters, and hence are not a comparable product.

There's a reason for that - ISPs throttle delivery based on message type - bulk vs. transactional. We work very hard to make sure that transactional emails have NO reason to fail to be delivered, or even delayed. By allowing bulk sending, your message is likely to get through eventually - but it might end up behind a several hundred thousand/million message bulk email queue - meaning that your customers' "forgot password" email or account verification message could take hours to show up. For our customers, that's simply not acceptable.

We've got 5+ years of experience with our bulk email service, Newsberry, to back this up. Not only are our IP reputations carefully monitored and maintained, but we keep bulk and transactional as far away from each other as we can because they're treated SO differently by ISPs.

I've been sending both transactional and bulk email using SendGrid. If I understand you correctly the only downside to that approach is that transactional email may be delayed because it is coming from an IP that also sends bulk email?

I was aware of Newsberry, but I still think that it is not comparable to SendGrid.

Even when we concentrate on bulk email, on one hand there is Newsberry, MailChimp, Campaign Monitor, Constant Contact, etc. And on the other are SendGrid, AuthSMTP, and now Amazon. The former provide an all inclusive solution: WYSIWG editing, templates, subscriber management, analytics, etc. while the latter are an order of magnitude cheaper, but offer a bare-bones solution that doesn't have much in the form of niceties and UI.

SendGrid is somewhere in the middle actually (and so is their pricing). They have analytics (and subscriber management if you wish) and even a crude UI where you can define email campaigns and newsletter templates (although clearly not their core business.) These feature put them above and beyond AuthSMTP which is basically an SMTP server. If Amazon has per campaign analytics (from my very brief look they do), then SendGrid should really be shacking in their boots, because Amazon has just undercut them big time without compromising on core competencies.

For sure. Definitely aware of it.

We certainly didn't mean to come across as arrogant - just confident. While your frustration on the pricing structure is understandable/reasonable, we don't have immediate plans to make significant changes. If there is anything else we can do to earn your loyalty, we'd be grateful to hear what it is.

SendGrid has about 1 month to lower it's price to at least $0.20 per thousand if they want to keep my company (We send about 3 million emails per month). They have provided me with a good service so far and I hope they are ready to compete on price.

For their sake, I hope they are not planning to compete on price.

Deliverability is the biggest issue. The other things they mention are just nice-to-have. If they can prove Amazon's system has a significantly lower deliverability rate they've got a solid argument to hold on to (for now).

We truly believe that deliverability is where we win - it's been our focus from day one. We've learned a lot about it (and continue learning every day), and we use the lessons to serve customers better. While we're confident in our practices that help with deliverability (namely, dedicated IPs), we can't jump to conclusions - so we'll just have to wait and see how SES customers fair in this area.

A customer of sendgrid as well and we send over a million emails a month using their services. Now we will try to send alerts/non transactional emails through amazon ses to test them out. The price difference is too big to ignore.

I think what we're going to see are companies like SendGrid, AuthSMTP, etc move to providing deliverability and analytics services on top of Amazon email ala Heroku building on top of EC2.

Thats sounds good in theory, but I think the technical details will make it impossible in practice.

Mailchimp launches Simple Transcactional Email service on top of Amazon SES: http://blog.mailchimp.com/mailchimp-launches-transactional-e...

What would make you change your mind?

I'm a very happy SendGrid user, but I simply cannot ignore the huge price difference. I will be giving Amazon a shot.

Before SendGrid, I was using Campaign Monitor, it was great, but outrageously expensive. SendGrid kicked off because it was an order of magnitude cheaper that CM, Mailchimp, etc. It seems that Amazon did to SendGrid what SendGrid did to "1 cent per email" providers such as Mailchimp.

Sorry for missing the point but this all seems like a lot of money for email. 100K/day outgoing is nothing that should tax a dual core CPU w/ SCSI disks and 100Mbps connectivity.

Are we talking about emails that aren't 100% legit? If so then what does SG offer other than an ever changing set of geographically disperse outbound MXs?

The pricing of Amazon SES is certainly very attractive, but it's worth considering why the price is so low. If you are looking for a minimalist service, then it makes sense to go with Amazon SES. I work for JangoSMTP and we often have clients coming to us from other transactional email companies, because they needed more than what these companies offered. Deliverability is a major issue that comes up. You may be able to send out mass quantities of email, but without an experienced email marketing provider you may struggle to reach people's inboxes. Delivery is as much an art as it is a science, and it takes constant focus to stay on top. JangoSMTP offers the choice between sending via API or SMTP. With Amazon, you have to use the API. We also offer open and click tracking, as well as advanced reporting (Google Analytics, logging, etc.). If reporting is important to you, you will want to consider a service like JangoSMTP that is more advanced than Amazon SES.

We believe in our service and the premium value it provides over competing services. We have our sights set on long term success and plan to continue our leadership in the industry. Your comments here are awesome, and we welcome further feedback at any time - on these topics and any others you want to bring up. We appreciate the supporters and detractors, because all perspectives must be taken into consideration in our efforts to continually improve. Contact us directly (email, chat, phone, twitter) or just keep on having your discussions here on Hacker News - we're listening ;)

--Tim Falls, SendGrid

How long before someone starts a start-up that adds open / click tracking to SES for a small fee on top? Given that SES is an order of magnitude cheaper, there is quite a bit of room to play with.

Argh. Logo anchor != Home page

Having played with the SES api they're just not the same thing. Sure ultimately they deliver email but the functionality you get from an app like SendGrid is far higher. SES might make sense for some transactional emails but for anything marketing I don't think I'll be moving away from SendGrid/MailChimp.

I've an account there because of the appsumo deal.

Please, fix your website and write some docs.

"Documentation for this feature is coming soon. Meanwhile, contact us if you need any help."

We recognize that documentation is of utmost importance for a service such as ours. In most cases, a lack of documentation is intentional and reflects that we are working on improving existing docs or writing the original content. All we can say is that this will get better and better over time. Your feedback/suggestions are always appreciated.

Put a fork in it.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact