Tuesday, November 13, 2012
How Sendgrid Makes Sure Those App Emails Go Through, With Jim Franklin
Story by Benjamin F. Kuo
If you're an application developer, sending emails through your application can be a chore. It turns out it's not just a quick line of code to send an email -- it's also a hassle to make sure that those emails get delivered, that your email isn't classified as spam, and handling all sorts of delivery issues. How do you offload all those headaches? Sendgrid (www.sendgrid.com) has stepped into that void, providing application programming interfaces (APIs) and all the infrastructure required to handle that email delivery. We spoke with Jim Franklin, the firm's CEO, to hear more about how the company went from an accelerator startup at TechStars just a few years ago, to now more than 100 employees spread across offices in Boulder; Colorado, Anaheim, California; Frankfurt; London, and elsewhere, to tackle the problems of transactional email. (Editor's note: we're crossposting a version of this interview between our Rocky Mountain site, techrockies.com, and our Southern California site, socalTECH.com).
What is Sendgrid?
Jim Franklin: Sendgrid is a cloud-based email service. We move transactional mail between web applications and end users. The best way to describe this is by example. If you think of Pinterest, one of our customers, every time content is pinned or re-pinned, an email is generated to let users know that is happening. Every time something is tagged in a photo, the application generates emails, which goes to the user of that app. Sendgrid started in Summer of 2009 at TechStars in Boulder, and we've now grown dramatically to 100 employees and 60,000 customers. We now send over six billion emails per month on behalf of those customers.
Explain why someone would use your service, rather than set up your own email servers?
Jim Franklin: Web developers are good at building web applications. However, every web application needs and email component. It starts when you register for a web application, the app send you a confirmation email, which is transactional email. It can be deceptively easy to set up your own, SMTP server, to send email on behalf of that app. The problem is when the things go wrong. One in five emails can go missing, and that's complicated by the way the Internet hands off emails from web applications to the end user. The problem we solve is where an application generates email to a human, where lots of algorithms are looking for spam and what to block before it gets to the inbox. What we do, is we handle the art and science of getting emails from an application to the inbox.
It looks like you were able to sign up lots of customers very quickly, what was your secret in gaining such rapid adoption?
Jim Franklin: TechStars was at the exact time and place to launch this kind of company. Our users are web develops, and technical people, who are the companies who are applying to TechStars, and who closely follow TechStars as a community. David Cohen is a very active blogger and spokesperson, as is Brad Feld. They've talked about this in their book, Do More Faster, where they talk about a variety of web services, including Sendgrid. This is one of those services that when you create a new company, it makes sense to outsource things. You want to use services like Gmail, Dropbox, and services like that, and Sendgrid is one of those. It doesn't pay to spent time or money to reinvent certain things in a startup. We offer a freemium model, where it's free to start, and you can easily point your application at the service. We take the email from there, and make sure it gets to the inbox. You only pay for us after you grow. If your startup doesn't go anywhere, you don't pay anything, unless you grow to become the next Pinterest, and you're sending hundreds of millions of emails a month, and we provide high reliability, scaling, and very low cost.
When do costs kick in for a startup?
Jim Franklin: It's driven by volume. We offer up to 40,000 emails like free if you're hosted with partners like Rackspace. That's quite a lot of mail, and you don't have to spend any money at all. For $79, you get 100,000 emails, and you're able to control your own IP address. Controlling your IP address is a big deal. For our free or lower tier plans, you have a shared IP, where you might be on a cluster with a hundred other companies. The analogy is like a credit score. When you first graduate from college, you have no credit, which is bad credit. The same is with a startup. When you start you rown web site, you have an IP address which no one has ever heard of, which gives you a bad sender score. That makes it very hard for you to get your emails to an inbox. Receivers such as Gmail, AOL, and Yahoo don't trust that number. So, if you're hosting on Amazon on a shared cluster and sending email, that reputation comes from the group of companies using that cluster. Instead of having your credit score with your social security number, that's like having it tied to your street address, and everyone on your block shares the reputation. If your neighbor's wife has a spending problem, that would mean you couldn't use your credit card, because they're linked. That why we offer a price point where you own your own IP address, and you can control your own destiny. That pricing scales up with volume, and it's pennies per thousands of emails sent, a very low cost.
What was the magic about scaling this up so quickly?
Jim Franklin: Our three founders, Isaac, Tim, and Jose, are all in their late 20's. All they had ever done was build web applications. Every one noticed that all of those applications had an email component, and all of their friend's apps had this email component. But, no one liked dealing with the email part of this. Among the cool kids, they wanted to build apps, but no one likes to deal with the email. It has a low status, and is considered a "necessary evil" in the web computing world. They said, we kind of like this, and why not build this once, build it right, and put it in the cloud. That was the magic formula. They focused on where other people were not focusing, and in an area that was not sexy. It's not geolocation, gamified, mobile apps. There are lots of areas that get lots of attention, but when you go into the middleware of cloud computing, there are lots of technical challenges, but there are not many people wanting to jump in and do that. It's not something a 26 year old developer usually wants to think about. They'd rather have one company like us to take care of that, where we can develop those relationships with receivers, so that they don't have to develop that relationship, because that's not your business.
What's the next big goal for the company?
Jim Franklin: We want to keep growing at scale, reliability, and move up from the stack to the application layer. We just launched a newsletter application built on our infrastructure, which is very inexpensive. Since we control the infrastructure, it's one fifth the cost of a MailChimp or ConstantContact for a newsletter service. We think developers can use this to grow their business. If you're just three guys in a garage, and have started to hire nontechnical people, that newsletter can be used by someone in marketing. Where our service previously required someone who can code, this new service lets someone without coding skills use our service.