Master Class Toolbox

Course Cheat Sheet

Andrew: This session is on how to grow 200% more traffic using referral campaigns. This session is led by Dan. He runs this site called Fuelzee, a mobile app company that helps you find the right gas station and get the best deal on gas. What we’re going to be talking about today is a system that he learned and implemented at Code School, which is a site that teaches coding online. My name is Andrew Warner, the Founder of Mixergy, and I’ll help facilitate. Dan, thanks for being here and teaching us.Dan: Thank you very much.Andrew: Dan, when you got started with Code School, how many subscribers did they have?Dan: We had just over 4000 subscribers.

Andrew: Four thousand subscribers. Is this what the site used to look like?

Dan: Yeah, that’s what it looked like when I started.

Andrew: Wow! And when you say 4000 subscribers, are they paying subscribers, or paying, I think, $25 a month?

Dan: Yes, it was $25 a month at that time.

Andrew: That’s great, but the grow rate was?

Dan: Well the growth at that time, we were only growing 2% a month, so it wasn’t really fast. It wasn’t enough to really make it a stable business forever.

Andrew: And afterwards where did you get it? After you did what you’re about to teach us?

Dan: We were growing about 15% a month in certain periods, but we basically had a 300% growth in 15 months.

Andrew: Three hundred percent growth in 15 months, so what you’re about to teach us you did deliver on. I see that you created a campaign that got 5500% return on investment, and I’ll talk more about that. All right, let’s talk about double user base, double subscriber, double traffic; everything just went up across the board. It’s not like one metric suddenly shot up.

Dan: Right.

Andrew: Let’s talk about how you did it, and this is the Big Board of topics that we’re going to be talking about. You can see on the bottom of your screen there the word more. That’s because we have even more than this and we’re going to try to zip through as much of it as possible, as quickly as we need to in order to get it all in. So, the first thing that you say is you suggest that we create a free trial. You guys had some hesitation about creating a free trial. What was the hesitation about it?

Dan: When you have educational content people can just take advantage of all of the content if they’d like if they have enough time. At Code School each course was about three hours. If you give somebody too much time in the free trial they can get through all of the content.

Andrew: And then that’s it, and then they say, all right, I got the free trial. I got everything I need, so long, and I don’t actually need to pay. And that was the hesitation. So then what led you to actually create a free trial then?

Dan: One of the things that we paid attention to is we created what we called a delayed free trial, so when you came to the site it wasn’t like sign up today and get 30 days for free. You would come to the site, and of course we have some free content you could play with. But instead of giving that free trial right away, we would wait two days after you registered and then send you the free trial. So in that way if you’re going to buy we let you buy. And then if you decide not to, then two days later we just sweeten the deal to get you to come back.

Andrew: Oh, that’s clever. I see, but how did you figure it out? But I guess we’ll get to how you figured it out two days later. So the flow is, someone comes to the site, and it looks like this. They click on ‘view courses’, and I know things change a lot. Then they would try to access one of the courses and be told that they need to enroll to play. And then after enrolling, that’s where you would get their email address?

Dan: In that flow that one is a little bit different. And some of the site has been rebranded. There was a big rebranding that happened. But when you typically get to a course and you would join the first level, for the first level you would have to give us your email account before you got into that course. We’d say, “Hey, you have to register for an account so we can track your progress.” You would then be able to jump into that course and be able to accomplish the first level. Usually there are anywhere from like eight to fifteen levels per course. Now if you didn’t enroll, which would be the goal anyway, and two days later you weren’t enrolled, we would send you that pre-trial email.

Andrew: And this is what that looked like. You actually called it a hall pass.

Dan: The hall pass, and we tried to keep things as branded as possible through the whole school part of it. Nothing was better than when I was in high school and I got a free hall pass and I could just go do what I wanted to do. So by connecting it to that it made it a lot easier, because people kind of fell into the gimmick and went with it.

Andrew: I’ve got it, so it gave them an opportunity to try the software. After they tried it for a little bit you asked for their email address. You have their email address, so if they don’t buy within two days, then you trigger off an email to them and say, here, try it for free. And the free trial also required a credit card, right?

Dan: No, that pre-trial did not require a credit card at all. It was a two-day pre-trial, and then there was a referral mechanic that was attached to it, and that’s where it really started to take off.

Andrew: All right, let’s get into the referral mechanic, because if we look at the Big Board, that’s the next big thing you suggest; to build a referral component. And how did that work for you?

Dan: It took a lot of testing to create the right referral campaign. And I have to give credit to the guys over at Dropbox. They did a fantastic job. Shawn Ellis over there, I mean, the program they created was amazing. So we looked at how we could give somebody a free trial, but also get them to give someone else a free trial. And we worked a lot on that. And what it came down to was the reciprocal incentive. If there was an incentive for you to take the free trial, you get two days free. But if you got one of your friends to also have a two-day free trial you would get an additional two days. And that reciprocal incentive is the most important part of our referral campaigns now. Because people no longer just want to give stuff to their friends. And they can’t give stuff to their friends and just get something in return. Both people have to get something out of the referral. And there has to be an action, and it takes some time to figure out what those incentives are. But once you get it, I mean, it’s golden.

Andrew: And here’s what that looked like. You can see that in the upper right, but I’ll read it just in case. It says, “Your hall pass ends in two days. Earn an extra two days when someone activates from your URL.” And then you give the URL, and we’ll discuss why you put Twitter and Facebook there in a moment. But that’s the process, and you said it takes a while to figure out what the incentive is. What are some of the incentives you tested that didn’t work?

Dan: We tried to figure out how many days of the trial we needed to give people, so one day give them a one-day trial and see how far they’d gone. Or give them a two-day trial and see how far they’d gone, and even all the way up to seven days inside that trial. We even tested out how many days that we’d have to limit the trial for, so you could only get a maximum of 30 days for free. And again, you could just get as many days as you wanted, and there were a couple really smart people who had amazing Twitter followings. The next thing we knew they had like 3000 free days, [inaudible 00:07:01] campaign. So we had to put the kibosh on those. But there were a lot of tests to see how many courses and how many levels that people actually get the advantage to take. How many of those people actually purchased, how many people were we able to capture to come back after the fact? We really found out that two days was the perfect amount, because it was just enough to get you knee deep in the multiple courses, but it wasn’t enough to get you to complete one single course. Because people tried to go at [inaudible 00:07:29] and take a little bit of that, a little bit of this. But by the time they got into it they were too far into the course to be like, “Oh, I can’t just leave this now. I have to finish it. I’ll pay $25 and finish the course.”

Andrew: That’s fantastic. And they get an opportunity to test out the site before they get to fall in love with it like that. All right, you give them the two days. Then you give them an incentive to tell their friends by offering them two more days, terrific. That’s where we are right now, but that’s not where you ended. You also say we should personalize the media sharing options, and that goes back to what I said before. I happened to see on the previous screen shot Twitter and Facebook. But you guys did something behind the scenes to decide whether Twitter and Facebook were the right media options. What did you do?

Dan: Well, there’s a company called AddThis.com. And what they allow you to do is when AddThis has a JavaScript on a site it will read the other cookies that are there and then display whatever social network is your preference. For an example, that screen shot had Facebook and Twitter primarily because I was the one taking the screen shot. I was either on Facebook or Twitter. Sometimes it would show LinkedIn, sometimes it would show Gmail. But what we really noticed was that when you’re working with developers, they are inherently introverts. They prefer their own networks. A lot of them like Hacker News, or Reddit, or [inaudible 00:08:48]. If you’re able to personalize that share and that button that they see as their network, they are more likely to share. That’s because not all developers are going to share everything on Facebook, so we had to tailor that experience to each individual person, and AddThis made it really easy.

Andrew: They are reading cookies to see which of sites your users are visiting the most and that’s what they end up showing.

Dan: Exactly, and it was awesome. We incorporated that and it worked out great.

Andrew: And it looks really simple and elegant on the site. And from what I see here in my notes, you guys doubled your social shares just by doing that.

Dan: Yes, we even expanded that. We learned in the referral campaign how successful it was. We moved that same program into almost all of the sharing that was on the site, so it was a tremendous help for the campaign.

Andrew: And what I notice going through the site is that after finishing one stage, I get a badge and then there is a button that says, “Share this badge,” and the button for me wasn’t actually a button, but was a Twitter button and a Facebook button. That’s because that’s where I spend most of my time. And if I happen to spend most of my time on Reddit and not on Facebook, then I might have seen a Reddit share button.

Dan: Yes.

Andrew: Okay, and on to the Big Board. The next big thing is send an activation email. I wouldn’t have thought that this was such a big deal. But obviously in preparation for this conversation I looked through the notes and I see it was. Why is the activation email such a big deal for growing user base?

Dan: Well, when you have an activation email, primarily in the [inaudible 00:10:17] campaign, it allows you to put specific pieces of information you need the user to accomplish. And what we did in our activation emails, we sent them the referral link. We sent them sharing buttons, as well. So that way in the email you didn’t have to go to another landing page to share. You just click on the Facebook button. It immediately opened up a Facebook sharing request. And your customer’s URL is already in there. And it helped the campaign expand even faster, because the people weren’t logged in. They were just able to share it right from their inbox.

Andrew: I see, right there, so that’s the reason you care so much about bringing this up in our conversation, because it encourages more sharing the way that you have it. I see it right there in the middle of the email. This is the email that you would send out, right in the middle of the screen. It says, “A quick way to share Facebook, Twitter, and email,” or I could just copy the URL.

Dan: Yes, we tried to make it as simple as possible. The email should be your landing page, and that’s what a lot of people really need to look at. Make your email the landing page. There’s no point in them clicking and going to another landing page and have to fill out information there. Do as much of it in the email as you can.

Andrew: What was your title there?

Dan: It was Vice President of Growth.

Andrew: Well you definitely did grow the company a lot. How long were you at the company?

Dan: Seventeen months. I was the first employee there, which was great. And I worked heavily with the founder and we did a lot of cool things with the company.

Andrew: Did they offer equity in the business?

Dan: They did not, because I had a side project they couldn’t agree with.

Andrew: Really?

Dan: But we won’t cover that today.

Andrew: I thought because they were self-funded they didn’t do equity for their team.

Dan: They do have equity, but they don’t do shares. They offer real straight up percentage points, and very interesting stuff.

Andrew: And on to the Big Board. Next you recommend that we send a ‘you’ve got a referral’ email, and I think this is what it looks like. I’m so glad you have these screen shots. You must have taken them before you left.

Dan: Yes, I actually did multiple presentations on this whole campaign while I was still there, so we shared a lot of this while we were still there. But the ‘congratulations, you’ve got more referrals’ was always a really good help, because it reminded people and congratulated them that they’d accomplished something. One of your friends joined and you earned four more days. But it that same message again, here are quick ways to share this again. You’ve already had success, so let’s continue this process, and continue to grow the business.

Andrew: All right, let’s go back to the Big Board. We really are moving quickly through these. Next is, “Send an up-sell email when the free trial ends.” This is how you take someone from just a trial member to a customer, and I’m going to show the email that you sent out.

Dan: When the campaign is over, obviously you want to get them to enroll, and you have to make it was enticing as possible. The call to action on this, of course, is enroll. We did do some testing with enroll, and in this email and with enroll I did perform the best. But you have to make sure you’re trying to convert that lead at the end of the day, because if that person just finishes their trial and goes on, and you don’t reach out to them, they might never come back.

Andrew: And you do something even more than that, but this actually is a huge help, and I’m looking at my notes again. You got 666 new subscriptions in nine months from that part of the process.

Dan: Yes, that was a huge help. When we first started there were not all the email integration, and there was not all that stuff. We had to continuously bulk things onto it. But it was a good help. Six hundred and sixty-six times a lifetime value of $150 adds up.

Andrew: And on to the Big Board, “Send a follow-up email with a promotion.” Now this is something that if someone didn’t become a customer, well what would you send to them, before I show it before I describe it?

Dan: If you did become a customer we wanted to send a way to sweeten the deal. The first offer was to sign up for $25. What we then did was say sign up for $9 for your first month. Let’s figure out a way to see if we could incentivize them to get them in, because once they sign up for the $9 first month, of course the next month is $25. And by that time they’ve already tried more content. We have their credit card on file, and we could still grow the business through that method.

Andrew: This is what that email looked like. I say looked like, because we’ve changed so much online, at any given moment things change. There it is, “Enroll now for $9 and get full access to all of the Code School courses; access code TV Screen cast, with no contract or commitments, and the ability to vote for future Code School courses, etc.” This was just for the first month. You’re just trying to get them to sign up and then afterwards they are regular members like everyone else.

Dan: And as you can tell on this one compared to the last email, this one has a lot more visual cue in trying to explain the value that they are going to get for that $9. And the email worked really well. We were able to capture additional people, and it definitely helped out. You got a lot for $9, but after that first 30 days you went right up to $25.

Andrew: This is what the previous email looked like. This is the one that use send when their hall pass, two-day free trial ends. The interesting thing about that is that yes, you do say Code School enrollment is just $25 a month. But ever completed course earns you $5 Code School cash. “Play often and enjoy payments of only $20 a month.” How did that work?

Dan: If you completed a course, any course on the platform, you got $5 off your next month, so there are a lot of people who actually only pay $20 a month. Since I left, or right before I left, I had that program killed, so that’s a little bit outdated. There is no longer $5 off.

Andrew: Why did you kill it?

Dan: It wasn’t helpful. People didn’t know that they weren’t getting the credits half the time. People didn’t understand, and it just ended up throwing off internal metrics. And when you’re trying to run business metrics, and you’re like wait a second, they paid $25 a month for three months, and then they paid $20, and then they paid $25. It just wasn’t helpful, and it wasn’t a sticky reason why people came and did the contents. They weren’t eager to save $5. They were more interested in doing the course. Probably about three months before I left, Eric Allen and one of the other gentlemen here who helped out the company, we agreed we just had to get rid of it. When you look at the business argument, if we get rid of this next year we’re going to make $550,000 more. So it was okay, let’s just get rid of this, because nobody cares that we have it. We’re just losing money for no reason, and we did eventually kill it.

Andrew: And to the Big Board, “Promote free trial to the right target market.” Did you ever promote to the wrong target market, and what happened?

Dan: Yes, you can’t promote free trials to Russia. We learned that lesson. They didn’t take advantage of all of it, and we had to make sure that we focused on which markets would use it correctly, and who would convert. So what we would do is dynamically change the homepage and offer free trials to certain people, and not offer it to other people out of the gate just because we weren’t going to earn money off them, and we wanted to get the profits over with. You had to be very conscious who you give your free trials to, because if you’re giving away money, and basically that’s what a free trial is, you want to make sure you’re going to give it to the people who are actually going to convert.

Andrew: What would you use to figure out that someone was in the wrong market? And by market, we mean specifically where in the world they are.

Dan: Yes, the region. We actually used Optimizely to do that. Based upon whatever region you were in, we would say if you’re based in this country, show this. If you’re based in that country, show that. In that way instead of using Optimizely as an AB testing tool, which everybody else does, we used it to dynamically change our website based upon certain attributes. If you came from Russia it would say get this free trial immediately so that we can move on. And then, of course if you were from Brazil, you would never have seen a free trial. So it made it very easy and quick to do.

Andrew: That’s a simple solution for it. I thought because you guys were a Coding company that you would have coded some great solution, and I’m so glad you didn’t.

Dan: No, I was working at a company that was full of engineers. I started at the company, and everybody was an engineer but me. Trying to get an engineer to do something for a developer or a business guy is not always the easiest thing to do. We would hack together a ton of stuff to make things happen; dynamically changing the homepage based upon referrals for us, you would land on the homepage and it would say, “Hey Facebook user, how’s your day?” And it was only because Optimizely knew that they were a referral from Facebook, and we didn’t even change the content. None of the developers there would have built that because it would have been by manipulation, or it would have been creepy. But I said, “Listen, this is what people love.” So we just hacked it all together.

Andrew: You also would encourage companies to get free trials, because companies were more likely to convert. In fact I’m looking at my notes, and ThoughtWorks had 300 employees sign up.

Dan: Yes, and we used the same hall pass mechanics to do this, which was quite interesting. When we first built the sales process we were kind of shrewd with it. “Hey, Stark, we’ll give ten of your employees a free trial. Just give us ten emails, and we’ll send in the advice, and they’ll each get 30 days to try it out.” And that’s because we’re talking about bigger accounts. All right, that’s sweet. If we get ten people that’s $250 a month, and that’s great. And then randomly one day it hit me. We should just get them to invite their entire company to a free trial, and give everybody 30 days free. What happens now is, if those people don’t buy a corporate account, we just gained access to all their developers who now have an account on Code School, which is great. But when that account ended you also had 600 plus people thought wait a minute, I just lost access to Code School. So it gave you this army of people inside the company who would turn around to their boss and ask, “Where is my Code School account?” And because there were so many people now with the ball, the boss would say, all right, let’s get the account. That’s actually how we got ThoughtWorks to be such a big account. It was because we gave their whole company a free trial.

Andrew: How would you know that it was a company that was coming through? I understand how you can figure out where in the world someone is, but how would you know that it was a company, and therefore you should offer them the hall pass for the whole company?

Dan: A lot of that was that we would notice the email address, if certain emails came through. I got an email of all of the emails that came in every single day, and sometimes I would scan through that. But a lot of these were relationships where we had reached out to them and started talking to them. ThoughtWorks actually reached out to us.

The funniest thing ever is how we got ThoughtWorks as an account. We had a rogue developer from ThoughtWorks who was using Code School all on his own, but he had a ThoughtWorks email. So somebody thought it was a great idea to put ThoughtWorks’ logo on the homepage, and say that ThoughtWorks was using us. Well ThoughtWorks then sent us a letter from inside the company saying cease and desist. “You have to remove our logo, as you are not using this with our approval.” And I talked to the guy and he was cool, and then three months later we had a 300-person account from ThoughtWorks. So somehow we spun that into an account, it was amazing.

Andrew: That’s fantastic, and I’ve heard of companies doing that. Of course, if you see someone from a company who is using your software you want to tout it so the rest of the world knows that that company now has your software inside and they’re using it. But I’ve also heard that some companies don’t like it, because they have all kinds of criteria about who can use their logo, and how to use the logo, and what it looks like.

Dan: Right.

Andrew: Most entrepreneurs will do exactly what you guys did and it’s nice to see that there is a positive outcome that could happen when you do that, even if the company isn’t happy with it.

Dan: Yes, and it worked out great for us. I was quite impressed myself. I couldn’t believe it.

Andrew: Way to spin that, and on to the next point, which has something to do with the flu. This one says, “Test your incentives until you get them right.”

Dan: Yes.

Andrew: And you tested day amounts, and what was the longest that you gave to people as a free trial?

Dan: We did seven days at one point for the initial trial, and it was just too much time. We didn’t get the conversion back. People had enough time to go through the courses, so there was definitely a lot of testing on those incentives and figuring it out. But we also did a lot of testing on how many days from the first time that they signed up would we send them that trial email. And we did it all the way up to 15 days, and we found out once again that the magic number was two, yet I don’t know why. But two days after was always better. But we also tested the day of the week, and we found out that Tuesday, Wednesday, and Thursday were the best days to send the email to get people to start their free trial, and were also the best days to get people to actually buy. If you sent it out on Monday they were too busy. They might have been at meetings. We don’t know the reasons, but it just didn’t convert as well. So we were testing not only how long the trial was, but on what day the email goes out is extremely important.

Andrew: I’m going to follow-up on that in a moment, but I’m curious about the software that you used to do all this. Is this something that you did get the developers to build for you?

Dan: It was primarily built in the MailChimp, and we used MailChimp as a giant user table, so we stored every date in there as well. So every single date that had to do with something was stored in there. A registered user was a date, and not true or false. And then what we were able to do was set up automation rules, so I could then change anything I wanted inside of MailChimp.

Andrew: So all this, two days, seven days, etc., was all within MailChimp. But you needed something within Code School software that says this member’s account works for free for the next seven days, or this member’s account works for two days.

Dan: The sharing URL that was given to them had its own information stored inside of our user table at Code School.

Andrew: Then that part you guys had to develop?

Dan: Yes, that part we had to develop. Now there is a way to get around this. You can’t use it for member services. There is a company called ShortStack App.com, and they have an internal referral component, which actually works really well. It’s typically used for baseball contests, but we’ve been able to hack it together so that way you can just imbed it on your site and then create a referral campaign, and issue a standardized promo code for a discount. But ShortStack is actually the easiest one that we’ve been able to hack together for more junior marketers.

Andrew: Is this the right site, what I’m seeing here?

Dan: Yes, you have the right site, but you’re just on an inside page. Just click on the ShortStack logo.

Andrew: Okay.

Dan: Maybe it’s just stuck on that screen.

Andrew: I can’t change it.

Dan: They started as a Facebook application company, and then since Facebook got rid of the default landing tab they kind of changed it into being able to embed things on your website and things like that. But if a user signs up for one of their referral campaigns, they get their own referral link, and they can share that. It integrates with MailChimp and a whole lot of other things. I’ve used ShortStack for referral campaigns at my company now, [inaudible 00:25:22], without having to get a developer involved.

Andrew: That’s fantastic. I’m guessing this is the way that the flu spread, right?

Dan: Yes, this is the R value, which is basically how many times it reproduces. And when you look at any type of referral campaign, if you are not hitting a reproduction rate of three, you’re not going to be able to really grow very fast. I had a scientist reach out to me at a recent conference, and I think the actual number is 2.7. We found three, so if a user did not get three people to join, we had no way to exponentially grow. So it has to be a reproduction rate of three people for every originator, and that’s actually scale.

Andrew: And that’s what you reached?

Dan: Yes, we got an R-3, and a 3.2 at one point. But we were able to get it, but it fluctuates, obviously. But until we hit that point, it wasn’t going to double every month.

Andrew: Wow! That’s pretty impressive, that a Code School, that an online school could actually get that kind of virility is really impressive, and especially one that charges. We’re not talking about Khan Academy where people can just go and click around for free, and so they’re more likely to spread it. We’re talking about a real school where people have to pay, and that it’s getting that kind of virility is fantastic.

Dan: It was the focus on quality. That was one of the greatest things about Code School. The courses are amazing, and the product is entirely solvent. But we struck a very niche market that nobody else was serving in that coding world. Everybody was either focused on entry level, or college, and we were in the middle of all of that. We couldn’t teach entry level people, but we were able to teach people who were intermediary and higher. And developers run in very small circles. They are all on Hacker News, they are all on Reddit, and they hang out in groups. You just can’t advertise to them. So if you can get one developer to tell another developer to tell another developer, that’s when you’re actually going to be able to grow. And that’s how GitHub got wildly successful. They created a social community for developers. It was genius, and that’s what we focused on at Code School. How can we create that, too?

Andrew: The thing that strikes me is I think you guys were fantastic before you implemented all of this, but fewer people knew about it. What’s great about these techniques is that they helped spread the word, and so many of them are reproducible. So, let’s go to the Big Board and look for another one. Here’s something that you mentioned before; you want to test your deliver day. What you’re telling me is that you’d wait to see if somebody paid. If they didn’t pay, then you’d offer them two days free. But you wouldn’t send those two days free the same day they joined, or the next day, or necessarily even two days later. You would send it a minimum of two days later, or the following Tuesday, Wednesday, or Thursday.

Dan: Yes.

Andrew: That’s the thing.

Dan: That was a huge thing for us. We looked at the conversion that happened each day and examined the open rates and click rates, and all the way through the conversion. And when we first started it was two days after you’d registered and that you’re were getting this email through [inaudible 00:28:41] subscriber. Then we started looking at different times of the day, and different days of the week. And of course we did send it on Tuesdays, Wednesdays, or Thursday, which were the best days. Saturdays were atrocious, Sundays were atrocious, Mondays were okay, but Tuesday was like hands down the breadwinner. We definitely set it on that, and it only would go out, if my memory is still there, at 6:00 a.m. It always went out at 6:00 a.m. so we could get the international time.

Andrew: At 6:00 a.m. Eastern Standard Time.

Dan: I can’t see that, of course, because it’s really blurry on my screen.

Andrew: Let me zoom it even more, and move it right there.

Dan: Is that Eastern Standard time, and 6:00 a.m. still?

Andrew: No this is 8:00 a.m. Eastern Standard time, Tuesday, Wednesday, or Thursday. I didn’t even know that you could do all this using MailChimp.

Dan: I love MailChimp; it’s great.

Andrew: I thought that only automation software, like Confusion software, or AutoDesk could do it.

Dan: No, and it’s funny.

Andrew: AutoDesk, or OfficeAutoPilot, sorry.

Dan: MailChimp now has an automation suite, and it’s based upon triggers on your website. So if somebody reaches this page, you can send this email out, but if they reach this page, and it’s during these hours, they send that email. They really have done a big overhaul of their auto responders. And we did a lot of testing on this, but the thing for us is the Code School was an international company. We had more user outside of the country than we did necessarily inside the country. We had to find a time that worked well for the U.S., but also catch people before they went to bed when they might have been in Germany. So there was a lot of time testing that had to go on.

Andrew: I really wish that it was easier to do that; to say, give it to them wherever they are in the world, and then when they sign up you can figure out where they are.

Dan: MailChimp now has that; they have their time warp feature, which allows you to send at 9:00 a.m. in every single time zone. They now have an IP tagger, which will also store the IP address of where that person is from. We actually use that at [inaudible 00:30:46]. We now use [inaudible 00:30:49] for our geo-targeted campaigns. We stopped asking people where they were located, and now we just know from the IP. So that stuff is primarily built into the solution now. And if you had developers, and had them hacking into a mandrel or transactional system, you can get very nitty-gritty on that.

Andrew: What did you use to tell you whether Monday was more effective than Tuesday? How would you tie the emails that you send back to orders?

Dan: In the beginning stages of course we used MailChimp to track the opening clicks. But we used KISSmetrics and we also used Mixpanel. We used both in tandem to attach the users and what messages they received back to the conversion. And of course we had UTM campaign codes on everything so we knew all of that information. But definitely KISSmetrics was the winner when I was at Code School.

Andrew: But you would at first look at open rates and click rates, and then that tells you that Saturday is not a very hot day. But it’s actually not the most reliable, because Saturday may not be a hot day for open rates, but anyone who is up on Saturday checking their email might be eager to start learning.

Dan: Yeah, but you have to look at the conversion. That’s a huge thing that we paid attention to. A lot of people forget this, but you have to look at the full funnel and figure out what your business metric is that you really need to increase. Because it’s great to say I had a 40% open rate, but it’s even better to say that I made $40,000 today. So we definitely looked at that full funnel, and then we would compare things back and forth, and even try to change it so maybe we got a really high over rate this day, but we got a better conversion from subscribers that day. We would try to figure out why there is that reason. Is Brazil a market that’s changing? Which day is more successful? And all we were able to come out of it with was that people don’t actually work when they go to work. They play on Code School, and that’s what we gained out of that. People don’t work when they go to work; they play on Facebook and Code School.

Andrew: Let’s keep going, because there’s so much more that we want to cover. The next one is to design your programming to be non-developer dependent. That goes back to what we talked about earlier, which is that you wanted to be able to do it without your developers having to help you with every step of the way.

Dan: Yes, and everything that we built we built so that a developer wouldn’t have to do it. Typically with a referral campaign, or something like this, you would build it all in a transactional system. It would all be hard coded. I knew that it would be a bad idea, because if you needed to get one change done it could take two weeks or two months. It just depends on what sprint the developers are in. It wouldn’t be your priority. We build everything we could in something that a marketer could edit. We had MailChimp, we used Optimizely. We used Unbounce for all of our landing pages. But everything we did we tried to make it so that we could edit it on our own.

Andrew: Did you use WordPress?

Dan: We did use WordPress there, but just because of how we built the program. But every other company I now work at we have put our marketing site as WordPress. We separated the application out so that it’s its own product. But WordPress is definitely probably the best and most flexible CMS to use for a marketer.

Andrew: I understand. The part that faces the non-customers is WordPress. It’s easy for marketers to edit, and then the app itself gets a sub-domain like App.yourdomain.com.

Dan: Exactly, and you see that a lot now. It’s a lot more prevalent than it used to be. At Code School when since we couldn’t edit every single page, we just used Optimizely to edit every page. So everything was dynamically changeable, and we could hide sections. That was the great thing about Optimizely. It’s a WYSIWYG or a website builder in essence.

Andrew: I know what you’re talking about actually. It’s so much better to instead of even asking a developer to change this button, to just go into Optimizely and say you’re changing the button yourself. Now this is not even part of an AB test. I’m just going to change the button, because I know the one AB test. And then when the developers are ready they can get to that.

Dan: Yes, they’ll finally change it and then you can turn off the Optimizely test.

Andrew: That’s cool; I didn’t even realize companies with big developers do it. That kind of justifies the whole approach, and this is such a bootleg solution, a good hustler approach. The next big idea is to get other companies involved. So everything we’ve talked about so far is about how you grow more engagement with your current users, and get them to bring more people in the door, but you guys did this. Here is Summer Camp. What was Summer Camp?

Dan: Summer Camp was awesome. It was a nightmare at the end of the day, but we did think about the logistics. We had a lot of people that wanted to get access to developers. A lot of companies that wanted to partner with us, do something, they always are trying to get us to do something. That campaign solved two purposes. One was to get everybody who wanted access to our developers to leave us alone. But two was to reward all of the developers with all this cool stuff that people wanted to give them. We came up with Summer Camp and we always tried to do holiday theme kinds of things and keep with the times, and keep engagement high, because it’s always personalized them. We partnered with 20 different companies. The requirement was that they had to give our developers something special. One company gave us $10,000 in pre-processing for every single person that bought a backpack for Summer Camp. We had another company give us three months free service, so everybody gave us all kinds of things, which was great. But what happened was that all of those 20 companies also promoted Summer Camp. We had the bigger email list. We had the bigger promotion funnel, and they all gave us free stuff. So we charged $100, and you got a Book bag. We shoved all this free stuff in it, and then you got three months free of Code School, and we shipped it to you. Well, you got $1,500 in free stuff. We sold out of 250 bags in 23 hours, so it was way faster, and we were way more successful than we had planned. But we got all these other companies, and there was so much more promotion than we anticipated that it was overwhelming.

Andrew: They gave you the stuff for free. You sold it to your audience at $100 per?

Dan: Yes.

Andrew: And they didn’t get any cut of it. All they got was your users who bought this trying out their software for a limited time.

Dan: Exactly, and an introduction to their solution. It was sent out in an email and also the Book bag. Now after the fact we didn’t think about the logistics that Amazon goes through every single day. We had to take 20 different things and then shove them into a Book bag and ship them wherever.

Andrew: Why didn’t you do it all digitally?

Dan: We thought about it, but then you wouldn’t have anything go in the Book bag. And the whole point was to have a full bag show up stuff with our logo on it.

Andrew: I see.

Dan: It was actually a mistake, because I had one of my guys in a room for three weeks just stuffing bags, and then putting them in bigger bags with labels on them and sending them out. That part was a little bit of a nightmare. But at the end of the day, to sell out of that in 23 hours, I was magically impressed.

Andrew: That’s $25,000 in less than a day.

Dan: Yes, less than a day.

Andrew: And I have here in my notes that you only lost $10 per student. What does that mean?

Dan: Based on the normal price of $75 for three month, with the book bag it was a total of $102.

Andrew: With the cost of actually buying the book bag?

Dan: Yes, with the cost of buying the book bag, and printing the logo and then shipping it there. Off the normal price that we would typically have, we only lost about $10.

Andrew: What’s the coolest thing that you were able to give away for free?

Dan: I don’t really know what would be the coolest thing that we gave away for free. I think the shirts that we gave away, the Summer Camp shirts, were pretty awesome. The bootcamp shirts were really cool. I don’t know, I can’t remember.

Andrew: Here’s what I see that was included in the list, and I’m not going to be able to get them all, but I see GitHub by C Intercom. I see StickerMule, I see Clickee [SP]. I see Active Campaign, EngineYard, New Relic, Benchmark Email, GoSquared, Geckoboard, just so many different things. And all these guys gave you something free, and it had to be more than just a free month. It had to be something that felt big.

Dan: Yes, it had to be something of value. We really turned down quite a few people. They were like, “Yeah, we’re giving 15 days, free,” and I’d say, “That doesn’t cut it.” And you have to send us something. And that was the hard part for some people; that we said they had to send us something, because it had to go in the bag. So, there were a lot of companies that made postcards and fliers and things like that.

Andrew: Just to be included in this.

Dan: Just to be included in this.

Andrew: Do you think that this would have worked out if somebody is listening and saying, “Hey, you know what, I would like to do this, too, but I can’t send out a book bag. I just want to partner with these companies. They can promote my thing and I will promote them in this bag and give it away to everyone. But it would be a virtual bag.

Dan: For us it was more about putting something in the hands of these people. When you think about it that’s the opposite of working in the Internet. We are so used to everything being digital, so the moment that you put something into real life you’re kind of changing the perception of the value, and that was the goal. You can definitely do digital. Maus has Maus perks, right. You can get these awesome deals from being their customer and get all these other perks, but to me it’s just a wall of stuff I can go take advantage of. To actually get something that’s in my hand, I think is even cooler.

Andrew: This is the email that you guys sent out.

Dan: Yes, the Camp day was a lot of fun. And that sale [inaudible 00:40:18] actually moved on the landing page, which was cool. [inaudible 00:40:21]. But the email went out to around 350,000 people on our side, and then obviously it was promoted all over the place. So it went really, really well.

Andrew: Well, let’s talk about “May the 4th be with you.” I’m bringing up the Big Board here. The next big idea is to run limited time discounts, and you recommend that they be centered around holidays like this one.

Dan: Yes, May the 4th be with you was awesome. We stuck it with developers, and I’d say this in no negative way possible, but they are nerds and they like Star Wars, which is great. We tried to follow that theme and that premise, and it was Star Wars Day on May 4th, so May the 4th be with you. And we offered an exclusive discount, and I think it was only available for 24 hours, or maybe a couple of days. But we did $35,000 in sales that day.

Andrew: Just from that discount, a 30% discount. What did you do when someone says, “Hey, I see this discount is available online, and I’m already a customer and have been paying for a long time? I want that, too.

Dan: We would usually just credit their account. The email didn’t go to current customers, but if someone was to complain about it we would always just credit their account.

Andrew: Okay, the next big point is to find other companies that have an interest in your users. We talked about the book bag, but that’s not the only way that you were able to do that. You guys also did that with New Relic. Do you remember what you did there?

Dan: Oh man, Relic was awesome. The New Relic was a big company, and wants to get more developers on their platform. What we did was we came up with a way for them to buy non-customers of ours for three months free, and they had to pay us full price. So if you deployed on New Relic, and you came through this campaign, they would buy you three months of Code School and send you a unique code. And then they would pay us $75 for you to have the three months for free. We definitely leveraged that campaign to make a lot of money.

Andrew: How?

Dan: When it first started out you would come to the web property for us. You would sign up and then we would send out an email from time to time. But the real way that we grew it is that we started a retargeting campaign. Once again we had a delayed start on it. If you did not become a subscriber within the first 48 hours, we actually shut off hall pass for this campaign, which was a bummer. But you would then start receiving, “Get three months free at Code School,” sponsored by New Relic, and you would see those ads everywhere. We would pay a lot of money because we knew that we had $75 to make back, guaranteed if you deployed. So we had a $25 budget for the retargeting campaign. And people would go and sign up for that and go deploy it with New Relic on their application, and then get a code for three months free at Code School.

Andrew: I see, so you were advertising them, so there was something in it for them? Now there’s a new person going out there and promoting them.

Dan: Exactly, and they were able to get a pretty engine. I worked down to just over $500,000, and it was about nine months that we did that. And I think they actually just ended another one of those campaigns recently. I don’t know if they did it the same way, but I got an email from them, “Deploy New Relic and get three months free at Code School.” I remember at the time we were the most successful campaign New Relic had ever had. And it was just because we leveraged their emails and we leveraged their targeting. We really went out and exploited it.

Andrew: They are basically paying you $75 to go bring in a new customer for them. Did they do much to promote it?

Dan: They sent their email to people who were non-current customers, so it definitely helped us. They did have an email system set up on it. I don’t think they ran any ads, but I do know that they did send emails out on our behalf. It just worked out really well, because it was they knew the cost per acquisition for any given user and what their lifetime value was. So the business venture for both sides worked, and it was really good. The only thing that after the fact that we didn’t really like was it made our churn look pretty high a few months after that, because all those accounts ended. So in one month we had a really high churn and then it went away. We still converted a lot of them. We worked really hard to do that. But when you give somebody something for free they’re not always going to convert at the end of the day.

Andrew: Did you get their credit card information on that one?

Dan: No, I think later on we did set the system where we did get the credit card information. The actual Big-a-thon that we had with that, which is how we track it on the back end; we actually gave everybody this internal membership thing to make it work. And we ended up having 700 members we couldn’t see. We couldn’t tell if they were actually paying subscribers, so there were some interesting things that happened, but nothing with credit cards.

Andrew: And here’s what that looked like. It’s actually pretty cool here, that you guys are having trouble, too, because it seems like it stalled so simply for a company like Code School. There it is; this is what you would send out.

Dan: We started with Geek Cry Day and we actually got some press for it. Once again, we stuck to a lot of the holidays and got a lot of stuff out there. When you attach the things that they can be relatable to, it inclines them to do something. And it was Geek Cry Day, so we actually got some press out of it, too, which was really nice. People signed up and it was for a limited period of time. There were people who went through 15 courses in that weekend. It was just amazing amount of work.

Andrew: The final point that we have is cross promote, but that’s essentially what we’ve just talked about here. Is there another partner that you guys did this kind of thing with where you said you pay us every time we get you a new customer and we’ll email our mailing list?

Dan: Yes, we cross promoted with a lot of companies. The company was very much bootstrap driven, get money. Obviously it’s a business, but we partnered with companies like O’Reilly. They needed a course on R, so they paid us to make a course, and then we gave that course away for free. We had a new free course, and then they had a course that helped get people to buy more books. So we were cross promoting all the time. We did relationships with GitHub. We have a huge relationship with GitHub, Google, O’Reilly books, Pearson, Block.IO, Flatiron Schools. We tried to work with as many people as we could inside the industry, and cross promote their products, and they cross promoted ours. But we never looked at anybody like a competitor. That was one of the things that I really took away when I left Code School. We didn’t have competitors. We looked at everybody as a friend, and just wanted to work with them. We wanted to help out in any way we could.

Andrew: What do we follow up with you? How do we read more about the ideas that you use today to grow your business, Fuelzee, or any other ideas. I love these growth ideas, because they are easier to implement. They just have simple logic behind them.

Dan: I would definitely call me up on Twitter. It’s really simple; @daniel mcgaw. There is no R in my last name. I show a lot of stuff on Twitter and on Facebook and things like that. I typically host on LinkedIn pretty often, as well. I’m pretty crappy at making a blog. I try to do that, but I just can’t keep up with it. I don’t like typing that much.

Andrew: That’s when you go to a mail-in list. With a mail-in list you don’t have to publish as often.

Dan: Exactly. I have a mailing list set up, but I’ve just never actually done it. I’m too focused on my business now and trying to knock that out. But once a month in Orlando I run the Growth Hackers meet-up here, so every month I’m doing something like that. But I typically share all my slides on Slide Share and stuff.

Andrew: This is your site, Fuelzee, and your Twitter account, and I hope we’ll get to see you. Are you speaking at any conferences any time soon?

Dan: I’ll be in, and I can’t remember the country. It starts with an A, but I’ll be there in March. I was supposed to be in Turkey next month, but I decided not to go. I can’t remember what it’s called now. That’s going to drive me nuts.