February 1, 2018
LaunchDarkly: (How To) Test In Production
Heavybit member company LaunchDarkly held their Test in Production Meetup in the Heavybit Clubhouse on January 10th. In this session the foc...
I'm gonna talk about growth hacking tactics for developer facing startups and for some quick background, I've been doing business on the web for about 11 years now. The thing that I've been doing is basically solving problems for marketers.
My first consulting company was called ACS and what we did was provide marketing services to other companies and marketers.
Then with Crazy Egg, we built a tool to allow you to see where people are clicking on a webpage. It was for designers as well as marketers. And it helped them understand data better without actually having to look through reporting and numbers -- because as you all know, marketers, they don't have math degrees.
Then with KISSmetrics, it was more about going further and providing easy-to-use analytics tools that also tied to user-level data. It was a lot different than Google Analytics and Omniture at the time. Enough about me, I'm supposed to know a lot about marketing - even though I'm always learning every day.
That's why I'm here to talk to all of you. I actually did an analysis and I looked at all the traffic on the properties and businesses that I've started and owned over the years. It looks like up 'til June of this year from 2003, we've driven about 41 million visits to all of our sites.
I like big numbers. I'm a marketer at heart. Although I love product as well, I'm not a developer, I just fake it. So I thought I'd just pop up a big number for all of you, because it's a big number and it was really interesting to go calculate it.
First, I want to talk about some of the shifts that I've seen since 2008, this is when we actually started KISSmetrics.
Marketing was really traditional and I'll talk a little bit more about what I mean by that. But there were a lot of standard practices, especially with B2B. They were very much the practices that you'd see a lot of companies take on and they are very similar.
New Relic was started in 2008. I'd say it was one of the first SAAS B2B developer-focused tools that you could say were developer focused, were SAAS, and made it easy for customers to buy. Obviously, they've been doing really well.
These are some of their stats from their website. They've had lots of downloads. They run on a lot of sites. They have over 200,000 users and they're going to probably go public very soon.
I also pulled up their stats around their funding. One thing I would say here is that since 2008, they've actually raised a ton of money. They've raised about $200 million dollars. It's likely that if you're starting out today, the prospects aren't great that you'll have this much money available to you.
I'm going to break down all the strategies that they use that I would bucket as traditional marketing.
First, if you type in almost any term that's related to New Relic, they're either in organic search or they're in the paid listings or both.
In this case, I just typed in, "application performance monitoring" which is the category they fell into. They also appeared in the enterprise category from back in the day when you had servers and results like that show up. They're in the ads and also they show up organically in organic search. This is a very typical strategy. Everyone will tell you to do this and there's lots of knowledge about how to do SEM and SEO.
Next, I wouldn't say blogging is traditional but you can see they have a blog. You go to their site, you can find their blog and they do blog a lot of things -- updates to their product and all that stuff.
What I will say is it's mostly product updates and announcements. They had an apology for the name of their conference recently because it conflicted with some other domain. It's basically a lot of press release announcement-style blogging.
Then on their homepage, you'll see this resources list and they have all these different things in their resources - videos, white papers, etc. This is a very traditional. This is the high market to B2B. This is what we know.
Another thing, I'm sure some of you have noticed their partnerships and some of you might even be thinking about partnering with them. They have a ton of partnerships with other businesses that are related or even somewhat related to them, and they have these partner pages. They're actually really great.
They include the logo of New Relic and the other partner and the goal here is that every time you have a new [partnership], you basically get a blast to their user base about your product and this free offer that you are able to provide them. It's like a campaign-driven approach where you do a partnership and there's a mutual e-mail that's sent out. That's how partnerships work.
Last but not the least, it's one of my favorite strategies for a lot of reasons but it's basically banner ads and free stuff. Again, this is pretty traditional.
You put banners out there on the internet. People click them. They come to your site. You paid for those impressions etc.
And there's also the shirt thing, which is what I've seen a lot of companies mimic. The offer is that they'll give you a shirt if you actually install [New Relic] and get it all the way to an install. Now, I've seen a lot of companies mimic this.
I think what's interesting is that people mimic and copy a lot of ideas blindly. I'm pretty sure the way they think about the marketing here is that the shirt is factored into their customer acquisition cost.
I would assume that if someone installs, they're very likely to get X money on average from them. The cost of acquiring the customer, has the t-shirt is factored in and it provides some sort of conversion lift to an integrated customer. I'm sure it makes customers happy and there's some branding effect. But at the end of the day, it's very likely the shirt factored into their economics and that's how they think about it. Anyway, if you're thinking of that strategy, please factor in your cost or it might cost you an arm and a leg for no reason.
Marketing was a silo. That's basically, the thesis that I have for what it looked like in 2008.
You had growth, which I noticed is something we talk about a lot more today. Usually investors, executives in the company and board members were the ones worried about growth. They were thinking about how they're going to grow this business and how [the company] is gonna increase its revenue etc.
And then you obviously have product. You have people building the actual product. A lot of the growth was pushed from these product initiatives. There's this road map and we think it's gonna make us more money.
It still happens today, but that flow was very traditional.
What happened was a product was built usually by engineers and the product team, and then you throw it over the wall to marketing.
Maybe you had a product marketer if you're lucky and then marketing would take it over and hopefully, they'd sell it. And there wasn't as much back and forth. Everything was divided up. This is how traditional marketing looked. This is even how teams inside the companies evolved. That's why you have all these different departments and things like that.
Now today, I would say that growth is actually what you do. It's not just something that investors, executives and board members are thinking about. What that really looks like is merging products and marketing. That would be how I'd form an equation.
A lot of developers like equations so I threw one out there. And I'm gonna dig into what that means.
When I talk about this in its simplistic form, I get a lot of questions and so, I linked to my friend, Brian Balfour's chart about this. He basically says that marketing used to be about acquisition and awareness.
Now, growth is really all about all these activities all the way to retention and the full stack-- it's all the layers of your funnel. There's a lot more to read about in that same post. He writes about "growth versus marketing" and he talks about how growth is focused. But the key point here (in my mind) is actually about growth rate.
Growth is really focused on growth rate. It's not necessarily about things like having a thousand leads a month and what's going to make us money. It's more about how we're gonna grow the leads at 10% every month.
That's really how we grow the business. This is a quick example of what growth versus marketing looks like and last is "growth versus product".
Product is all about creating value -- it's about using your product to create value for your customers. Marketing is marketing.
In essense, growth is really about getting the largest amount of people to actually experience your product's value.
And then it's about getting as many users as possible and getting them often. I would even say getting users "as fast as possible" to experience the products. That's my thesis around growth versus traditional marketing.
Today, if you're a company and you exist and you're trying to grow and try at marketing, you have to think about it from a product perspective. Not just from a traditional tactical perspective.
There's really a very simple reason for this and the simplest reason is that for all the channels that you would constitute as traditional marketing -- whether it's SEO, SEM, banner ads and giveaways -- more marketers are using those channels. Thus, the [channels] are much more expensive and harder to leverage to actually grow your business. Now, you've got to think a little more uniquely about yourself and build something with a growth mindset.
I always get the question: how do I figure out what the core value of my product is? How do I actually figure out how to improve my product?
There's again, another handy thing. There's a survey tool that we created a long time ago at KISSmetrics and the tool still exists if you just send it to your user base.
What you're really looking for is if you have product market fit. It's another thing a lot of people talk about because they say you can't grow your business until you actually have product market fit.
What "product market fit" really means is that you have enough people that believe that your product is a must-have.
There are other ways that are quantitative to assess this, like retention. But I really like this method because it's very qualitative and it can also help you understand other things about your customer.
For example, if you ask people, "How disappointed would you be if this product no longer existed?"
You're really looking for 40% or more surveyed to say they'd be very disappointed.
I always get this question: Well, I did this survey and only 30% of the people said they'd be very disappointed. But 70% say, they'd be somewhat disappointed. So can I merge the two?
The answer is no. This is actually great because you can go learn. You should segment the responses between the two and find out things like how the "very disappointed" people describe the product and how that differs from how the "somewhat disappointed" people's answers.
Just a quick example: Sometimes there are certain features that people are exposed to that make them more inclined to be very disappointed if your product no longer existed.
What you learn there is that you need to get more people exposed to that feature and you can make product improvements and changes around that.
There's a lot more to that and you can find it at that link in the survey. I recommend all of you at least ask that question of your users.
Investors, board members and executives are thinking about growth and how to grow the business. They have a fiduciary responsibility invested in your business and they hope you're trying to grow your business as fast as possible. That's where the pressure can come from. At this point, everyone's working on growth.
It doesn't matter who you are in the company (especially a startup), you probably won't have a job if you can't figure out how to grow the business. It becomes everyone's responsibility, especially because marketing, customer acquisition and getting a lot of people through the door is becoming harder than ever.
I'm going to give a quick case study of how we've been able to achieve this at one of my companies, KISSmetrics.
We are partly developer-focused but we're also marketer-focused first and foremost.
This is our homepage. I just wanted to put it up because there's this cool-looking person on the homepage with a smile and he's looking at the button. We tested it. When he looks at the button, we get higher conversions.
This graph was actually something that happened recently. This is maybe a month or two ago where the last dot on the right ended. Over 12 months these are the visits to KISSmetrics.com. My 40million visitors would be a little bit higher if this hadn't happened.
What we noticed is that we started losing traffic. One of the reasons is actually, ironically, traditional marketing -- SEO.
Google started saying that they didn't like us for some reason. We actually weren't doing anything spammy or anything like that. They just made some updates and thought that certain things we were doing were spammy.
It had to do with people syndicating our content. We have a relatively popular blog so people syndicate our content. Google doesn't like that anymore and so we started losing traffic.So we got pretty sad.
But we actually got more signups than ever before. That's counter intuitive in a lot of ways. The line of the graph has actually been going up overtime. If you overlaid the two, you'll see that there is a little bit of a trend for a few months. But overall, we keep going up. You'll notice in the middle of that, there is a big steep improvement.
The way we did that is actually through conversion rate optimization -- it was essentially growth. The fact that our whole company is focused on growth means that even when we lose traffic, we've still found a way to increase our signups.
That's literally our conversion rate overtime. That really tracks really well with the signup graph which I'm gonna go back to really quickly. You can see that's the signup graph and that's the conversion rate graph. You can even see in those months, we actually didn't improve the conversion rate. That's when we actually had flat signups but we were growing traffic at the time. So, it wasn't so bad. This is how we did it.
I showed you that homepage earlier and it actually has these components and it's really fascinating. I can look at one page and tell you how different teams worked on it. One of our headlines is about Google Analytics. Basically, Google Analytics tells you what's happening. KISSmetrics tells you who's doing it.
One of the things we did is when we did our surveying with Survey.io, we learned that everyone uses Google Analytics and everyone understands it or thinks they do. So one of the best things we thought of which made us cringe a little, we decided we might as well try it. We tried using Google Analytics to help people understand what we can do for them and then overtime, we started doubling down on that.
We added a Google Authentication on the homepage. You could only log in and sign out with Google.
In fact, we labelled the button "log in" because that increased our conversion rate by another 15% or 10% compared to "signup" or even "sign in". We tested that.
We did the URL first, which is actually like a 36% increase compared to not having a URL first and just making them click. And this initiative looks simple. The page is simple but it actually took some effort by engineering to get the Google Auth going because you run into all these questions: Can people log in with Google?
Actually, log in on the sign-in page and if the answer is yes, then there's all these cascading changes that end up needing to happen.
This was an effort between marketing, product and sales because our lead volume is higher in sales but also because of the type of lead we get.
We wanted to make sure it wasn't hurting the quality of the lead. At the end of the day, I showed you results for my own growth hacking or growth initiatives.
There's this cool cartoon that says, "Growth isn't a strategy. It's actually a result." and I think that's pretty interesting. I actually had "growth strategy" on my slides but I think it's really important to note that if your whole team's working on it, it's actually a result.
It's something you have a goal for and you're trying to achieve. It's not just a strategy like, "Oh, our strategy is growth." I'm going to talk more about that more.
First, hopefully, you're literally, trying to attract, retain, delight customers and that's the goal. And your customers are developers obviously... So , where do they hang out?
They actually hang out at Meetups. That's clear even when I did this homepage shot. I think it was a Swift Meetup or something like that. That was right on the homepage. I logged out. I made sure it was not personalized for me. I cleared cookies but yet, there's the Swift Meetup. I believe it's Swift. Anyways, yeah, it is a Swift Meetup.
Github has over 10 million repositories as of the end of 2013. This is where developers hang out, period.
There's a site called GitHub Archive. That site has a whole archive of all GitHub data up to a certain date. There's somebody that analyzed millions of GitHub Commits and they put this up and you can use Google BigQuery. They even did a whole presentation on it and then there's also this thing about GitHub TreeMaps Visualizations.
So, I haven't bored you to death yet I hope but at the end of the day, your developers are on GitHub.
They're active on GitHub. Just like a marketer would figure out what type of social content is popular and go write about it on a blog, I recommend that all of you figure out how to analyze GitHub. Know who's actually committing to what.
This is what this TreeMap actually does. It was setting examples of bootstrap and I found a bunch of folks that were contributing. Some of these big boxes are actually people. Super interesting. I'd go figure out what kind of real questions GitHub can answer for you to help you figure out what to do next. And next could be any of the things that I'm about to share.
If you have any questions about any of this, obviously we have time at the end. This is another thing which is a little, slight, quick tangent. I like ranting about things as you probably notice.
But basically, I really believe today that Rich Content and Freemium are the only two strategies actually just keep on giving.
At KISSmetrics, if we didn't have all that traffic then we wouldn't really be able to get more and more people to sign up overtime as we increased the conversion rate. It's a flywheel, and it keeps working.
I think Freemium is another one. One thing I like to say about that is that we do all this work to go get customers to signup for trials and then by the end of that trial, we want them to go away if it didn't work out.
I think that's just pathetic. That's just not good for the customer. It's not good for your business. If you can find a way to do something for free and let them stay on and provide value, that would be great.
Now, I don't want to be like the professor who recommends this and then all of you try it and screw up. If any of you are considering it, I'm happy to chat about it but it is something you have to think about very carefully. At the same time, it works really, really, really well. And of course, when it comes to rich content, one big lesson that I've learned is people don't really like being marketed to.
They don't like being yelled at or talked to. They actually do love good content and they'll keep coming back for more.
One of the reasons we've been able to scale our KISSmetrics blog (for marketers especially), is that we write really long meaty posts and we make sure that they're high quality.
We have an editor and he really cares about the audience. If I wrote a blog post and he didn't like it, I'd either have to change it drastically or he just wouldn't post it. So, it doesn't matter who's writing it, we make sure the content quality is really high. That's an example.
I wanted to give some Rich Content examples. I'm just gonna get right into them.
GitHub has guides. I think guides are really helpful. I'll explain a little bit more later.
Segment.io has an Analytics Academy. I'm sure some of you have thought of this idea. People put in an e-mail and you basically, drip lessons to them overtime. This can be a really good tactic if you need to do a lot of education in your market and you think there's some way that you can get a lot of e-mails but that's a caveat.
And then Pardot, which is actually a marketing automation tool that we use at KISSmetrics. They have this ROI calculator. I really like if you're talking to prospects and you're doing sales, or if you're trying to understand why people are signing up and not converting. Users might say, "I don't understand what value this drives for me" or "how much it cost?" or "what is the ROI?". They built the ROI calculator for them and you introduced that into the marketing workflow.
Another one, which is one that's really creative is How Much to Make an App. It's from a company called Crew that actually helps you find people to help you build apps and websites. They made this quiz. It's really cool. At the end, it spits out an answer of how much your apps are gonna cost and then it's actually a lead gen for them.
Then my co-founder Neil, he has a site called QuickSprout and we built an analyzer there. You put in a URL and it will analyze all kinds of things specifically around SEO and social media. We've gotten a lot of feedback on it around the fact that people have made changes using the Analyzer and they've actually increased their traffic dramatically. Those are the five tactics but this is a big statement that I wanna make for all of you that are building developer tools that are B2B focused.
The big statement here is that the type of product you have actually influences everything you have to do and the big thing you're trying to create is actually a relationship with your potential customers, your existing customers, or both. Ideally both. So, the type of product you actually have does influence marketing.
But before I get into that, I have this disclaimer. I've shared this with a few people including engineers on our team and some of the team here. The first thing in the caveat that I'm going to type you and break down your startup into one type and think you're stuck there. That's not the case. This can help you think about what product type you are now and how can you think about your marketing a little bit differently.
That being said, all the tactics I share can be used across different product types. It's not to silo anybody but I wanted to add this disclaimer so I don't hear that objection.
I made this up. I'm not an engineer. I think I did a good job but basically, there's three types.
There's infrastructure. There's platform and there's tools.
Based on these types, you're gonna want to have a different strategy at least right now, based on where growth and marketing is today.
These are a bunch of examples of infrastructure companies. They allow developers to get apps into production. I would say they allow developers to get apps into production faster and more accurately. And these are a bunch of your [Heavybit] startups and who I think would fall into infrastructure.
We can debate it during the Q&A if you guys would like but this is how I broke it down.
When it comes to infrastructure, this is the way it works. Basically, you need widespread adoption because your margins are somewhat low and so, you need lots and lots of people using your product and getting exposed to it.
I know everybody needs that but for an infrastructure play, this is very important and that means that distribution and speed to deployment are really what's critical to your success for your customers.
You'll see I focused the growth strategy on customer success. It is not about a customer success team or anything necessarily. This is literally, how you make customers successful.
These are the things that I've identified for infrastructure companies and based on that, out of a long list of strategies or tactics, I would say that subdomains and tutorials are the most common but also the most valuable as a result.
Subdomains are basically where someone can come in, signup and then get a subdomain on your domain. That's how they either deploy or add scripts if it's like Cloudfront. Then what happens is that link gets spread and more people see it.
It's a branding play but it also helps with widespread adoption. If you wanted to get a little more aggressive you would add a "powered by Heroku" badge or something like that in addition to a subdomain. I've seen that be really valuable.
On tutorials, I'm just gonna get right into. I think Parse did a great job of the tutorial strategy and they actually also leverage GitHub. I looked at their GitHub page and even though I know they've been acquired by Facebook, these are all old valuable things and I'm sure they could have scaled their marketing with them.
This is Parse's page on GitHub. They had a demo called Anypic. It's basically an Instagram clone that you could build literally with a couple of clicks and some code in probably a few hours or less. They've had 800 favorites. They had more than 800 favorites and 400 forks on it at the time I did this [presentation] and I'm sure it could have gotten a lot higher if they needed to.
Then they had AnyWall, which had something to do with location based mobile app and adding content. So that was pretty popular.
Then they had a Todo list app. That was actually a fork from somebody else, I believe or inspired by something else and then converted over to Parse.
When I think about a growth strategy, it's something that's more repeatable and unique to the business that you're doing. I thought that this was a really good way to give people an example and also increase that speed of deployment.
One of the big things with infrastructure (and especially, with developer product types like this) is that you might get an engineer just tinkering at home on this.
I'm sure all of you if you're engineers have done this with a product and then you end up bringing it in to your organization.
That's a classic strategy and this is one way to accelerate that and get more people exploring your product. That's also why Hackathons have been so popular.
This is the case with Platforms too. Platform is the next one and this category I've defined as "enabling developers to create things faster".
You could see I have Twilio, SendGrid, Stripe, MailGun and Urban Airship as things are actually helping you build things. This is differentt than infrastructure in my mind because it's actually helping you add functionality to a product more than anything else. This is how I've divided it up. I know some of you in the room again, can debate it later.
In this case, it comes from people understanding the use cases. You need to answer to the customer, "What's in it for me? What can you do for me? How are you gonna do that and help me do something faster and cheaper?"
That's really what it boils down to when you have a product like this and I think infrastructure falls in this category too. It's very hard for somebody to come and rip it out and say, "Oh, we're gonna switch to another tool."
I think infrastructure's harder. Platform is a little bit less hard but also something hard for people to rip out. I know people that had been on SendGrid for a long time and only just moved away from it just because it was so ingrained in their processes.
It's the same with Twilio. In this case, the biggest thing they're trying to do is prove how fast someone can use this product. It's not even how easy it is. It's just how fast they can get something live with the product. That way they can show it off and then obviously, integrate. They show how easy is it to use this API.
Then obviously, I talked about libraries earlier. The more libraries you have, the better.
I mentioned analyzing GitHub. You can analyze GitHub to figure out what type of languages people are using for other libraries that are related to yours or even what kind of languages people are using that are building the type of things that you're interested in.
You basically prioritize your libraries and which ones you support based on that.
This is one from Keen and with open source use cases. It's their dashboard.
What they do is they allow you to build a dashboard using Keen's API and whatever else you want and make it really easy to visualize something. They're totally a platform play.
They don't have to be doing this, but this is the kind of thing that I'd be doing in terms of marketing.
It's basically demonstrating the use cases of your product, adding some value, and making it easy for people to integrate.
The next example that was really exciting to me was when you start seeing blog posts like this from SendGrid. This is SendGrid's blog where they talk about how SendGrid event data plus Keen equals awesome. It's also a repeatable strategy.
It is a little bit similar to the partner strategy in New Relic but it's actually way more product based. It shows how much value you get when you connect these two companies. It's very much a product play and less of just a marketing play.
Here's one I noticed on Intercom when we were installing it on one of our projects.They provide communication with users and a whole platform to do that.
They made it so that if you install it and you're using Rails with this RubyGem, it is just amazing. All the stuff is automatically in there, all the tracking for the users is in there and you literally are just installing it and it's done.
It's like magic and I was really impressed by the ease of deployment. And nobody has even copied this yet. So that's another example for you guys.
Really this is probably the most traditional SAAS product category where you're just trying to help people do their jobs. In this case, you're trying to help developers do their jobs.
And these are a bunch of the companies that I would say fall in that category from Heavybit.
The key is that you're actually solving a problem that's habitual and what I mean by "job" is they're buying your products so that they can make their lives easier. You're actually doing a job for them, helping them do that job.
In New Relic's case, it has to do the job of monitoring. It was really a pain in the ass before they came along but at the end of the day, it's a habitual job. It's something you need all the time and the product experience is literally the bottom line to all of this in terms of the tool.
For me, product experience starts with onboarding and basically ease-of-use and that's the way I think about it.
I have this thing called Click, Click, Awesome. It's what you should ask yourself if you're building a tool. How do you make the product experience feel so easy that at the end of the day, the customer feels like it was a click, click, awesome experience?
For Amazon, it's called a "wow" experience. Sometimes you can't go that far when it's a boring tool. But click, click, awesome is what I would aim for.
In that vein, something you can try on onboarding is to let people signup with GitHub. Let people sign in with Google. I gave you the example from KISSmetrics.
On one of my company's products -- HelloBar, we did the URL first onboarding there too. There's a really important purpose with these three strategies or anything like this.
Think about what kind of authentication or other info can you ask for that's a small ask.
It's not just their e-mail because usually, that's a bigger ask. You wanna do that later on or you wanna do it as part of an authentication if you can pull it off.
Basically, how can you make the product experience on the next screen 10 times better by using the authentication or whatever info you're acquiring (like a URL) on the homepage?
The simplest thing with Hello Bar is if you put in a URL, you're seeing your site. We noticed that had an easy double digit 20% - 30% increase in conversions with people giving us their e-mail. They were able to see their site and imagine what this thing was gonna look like on their website.
With Google Authentication you can do the same. With GitHub, you need to know how can you make the experience better because they authenticated with GitHub, that's the key.
A friend of mine, Samuel wrote this book, User Onboarding. He's gonna come speak at some point at Heavybit. But the book, it's literally, the definitive resource for onboarding.
He's got a website called, Useronboard.com where he shares a lot of free teardowns of other products and then he's also got this whole book about it. He's gonna come speak. It's great.
He's also got this great graphic that really always makes me laugh and makes everyone laugh so I thought I'd steal it and put it in here. But basically, the idea is that you aren't making the flower. You're actually helping people be like fireball Mario.
For all of you that played Super Mario Brothers, this is awesome so it's probably my favorite image about products from a very long time. That's the key. People are buying better versions of themselves.
I like TLDR's so if you didn't pay attention at all, this is the key.
The key here is that you're just trying to make it as easy as possible for people to understand and use your product.
I debated here whether should I put "use" first and then "understand" last? And when I thought about it, I chose understand first. "Understand" is the onboarding experience, understand is even before that.
Do they even understand what your product does? It has to do with the naming of the product. It has to do with how you market it, how you get it out there, etc.
I chose "understand" and then "use" because I think understanding your product tends to come before they even use your product even if it's word-of-mouth and someone is explaining it.
I would say that honestly, the right way to do it would be to look at your data. Ideally, you're looking at a funnel or some metrics and trying to figure out where's the biggest lever and where's the biggest opportunity.
You might look at your business and say lots of people are signing up but not enough of them are activating. So, sign up may not be the first thing you work on. Instead you work on what can you do around the onboarding experience to activate more people and the things you've got more control over.
If you don't wanna bother engineers, I'd probably start testing the drip e-mails you send to get people activated. I'd probably start making them smarter. It's just an example but it always starts with data in my mind.
I am sort of a data analytics geek but if you start with data first at least you'll know you're working on the most important problem.
In my mind, if you're a very small startup, it's very likely that you don't have resources to go hire an editor. You could either make [content production] a shared thing amongst all the people in the company or you could work on building a definitive content hub or a blog for your category.
I'd probably have shared resources on the team and I'd probably have somebody on the team start getting really good at writing content for your audience. You should understand your audience better than anyone else. That being said, if you need to find someone, there are a lot of people writing content out there.
You either want someone who's writing or critiquing content really well. In our case, the person who's our editor, he edits content all day. He rarely writes any content but I would probably suggest finding someone who's actually already writing content that's the kind of content you wanna be writing. You can see if you could hire them even if on a part-time basis just to help you edit your content or help audit it.
It's likely if someone's writing good content out there, they're actually pretty good at editing as well. They care about the quality and understand the audience.
There's a lot of great content on habit forming. Thankfully, I don't have to cover it all right now. The truth is I probably think of it in much more basic ways than that.
The first part of habit forming in my mind is the communication channels that I have that I can use with my customers or my users.
E-mail is obviously the classic one. I'd be trying to test things around SMS even or even instant message. Once there's API's for some of these mobile messaging tools, I'd be thinking about that.
Another great one is Chrome extensions and the communication channels on Chrome extensions. If you can get people to install that stuff or give you their e-mail or even their phone number, that's a great communication channel.
I think habit forming requires a trigger. My favorite content on the topic is by this professor at Stanford called BJ Fogg. He talks about triggers. First, it's communication channels. If you're building a mobile app and you're relying just on push notifications to get people back to the app, it's likely you're not gonna be able to scale the business. I think what you have to rely on is somebody triggered based on their first experience.
This also applies to applications on the web. Based on the first experience, you literally incent them to understand when they need to come back or when they need to use your product. I think that's the bottom line.
If I were building a mobile app today, I would build it without notifications and basically, see how you can tweak the experience to get a higher retention rate. The notifications are just an accelerant on top of that and you test them on the next phase.
There's obviously all kinds of caveats to that like if your system's really notification oriented. But at the end of the day, I really like to think about the first experience. I know I've mentioned it a lot in the different categories of the talk. But you need to know how you can make them come back just because of what you did to them when they first came in.
Communication channels are also key.
Yeah, so you're asking about GitHub and what I would call the communication channels within GitHub. Facebook also has those.
I think there's a whole talk on that and I don't think anyone's given it yet. It's not out there yet.
You guys can all probably understand this way better than I can. I'm not in GitHub all day but basically, if somebody commits to something, what are the communications that go out to other people?
That's why I think there's a whole talk on this but I would literally map what happens when I do something on GitHub and how it touches other people. And then using that, you could probably figure all kinds of stuff out.
For example, if you fork somebody's code, do they get a notification? Because if they do then you might wanna fork more code if or projects.
One tactic I've heard, which I really love is actually editing the documentation for code or the read me file because nobody does that.What if you were the person that just went in to the read me file and then edited all the things you wanted to influence or integrate with while improving all the documentation?
I think people would pay attention and that's still a commit happening when you do that because of the way GitHub works.
These are all the kinds of things I'd be thinking of.
One example, which is unrelated but is really exciting to me is every time there's an update to a file on DropBox, there's a little notification that comes up in everyone's computer that has access to it. DropBox is another underserved opportunity but I think a lot of that can happen with GitHub as well.
We talk a lot about Google Analytics and we help people with it. The strategy there is we did a lot of surveying and a lot of analysis and what we learned is that everyone that signs up for KISSmetrics is a Google Analytics user.
Helping them use Google Analytics better (especially considering it's free and it's backed by Google) only helps us.
One, it helps us get them earlier before they're ready to add us onto their tool set, but it also, lets them know that "these people will help me."
One of the biggest opportunities in analytics and data today is that big customers might say, "Google doesn't help me. They don't help me get things set up. They don't have regular check-ins with me."
The whole idea, if you already think about it more psychologically, is that can we get people used to thinking of us as a trusted resource. If we're talking about Google Analytics, it's free. We don't expect anyone to switch from them to us. We actually expect them to use us in addition to Google Analytics. We don't tell them to cut it off or anything. If we can teach them how to use it, we believe that they'll trust us more. And that's been working.
It's "embracing the gorillas" and it's actually what we call it inside the company.
It's not even thinking like a marketer. It's actually thinking about growth - is probably the better way to frame it.
Yeah, the people that I know that are in growth teams actually taught me a lot of this stuff around how to think about it. Basically, they were saying, "Look, you have a startup, you're trying to grow it and if you're not actually trying things to grow it. If some mental block, emotion or feeling is getting in the way of you doing the right thing to grow -- you're just doing it wrong."
You're not doing your duty to basically grow the business.
When I hear somebody having a somewhat irrational assumption about what's gonna work with their customers or their visitors, that's when I feel like, " oh, that's developer thinking." But thinking more along the lines of growth, means that if you have an idea like an assumption about GitHub authentication -- like people won't sign in with GitHub because they don't trust us to give us that data, you can just test it.
I would just test it. Try GitHub authentication versus what you're doing today, which is normal e-mail and password and see if more people do it. Now, if more people don't do it, then go learn why they didn't do it. On that GitHub authentication landing page, do a little pop-up, which I know is another issue but it's temporary. Ask them, "Why haven't you signed up?" and make them tell you, make them validate your assumption.
Make them tell you, "Oh, I don't trust you." or "I don't know what you're gonna do with my data." Then what you would do is you would test copy right below the button and say what you're gonna do with the data.
To me, it's just about testing and not about anybody's opinion. Let's get some facts, and let's test it and see what happens.
We can decide whether we feel good about doing it or not, but if you don't test ityou're not gonna know whether it's gonna work for you or not. You're also not gonna know how people feel about it. You're just making assumptions and in our company, we'll test something, learn something and then someone else will mention it again and we'll look at that test. We'll be like, "Oh, we did it wrong," or, "Something's changed, "let's go try it again."
Instead of making all these assumptions about your visitors or anything like that, go find out. That would be the more direct way to say it and it's really a growth mindset --we need to grow. We need to figure out how to do that. We shouldn't be above or beneath any tactic. We should try it and learn about it.
Obviously, there's ethics and morality, but that's different than just making some assumptions or even thinking you know developers just because you're a developer. You don't necessarily know.
Okay, the question is, do I think that every developer product should have a free plan or be freemium? I don't think anything. I think the right answer is it depends.
In my experience, the businesses that grow the fastest are ones that have a free plan and I should be more accurate about that -- it's the businesses that grow fastest with the least amount of effort put into marketing and sales.
I know you guys are allergic to marketing and sales. If that's the case then free helps you do what you do best, which is make a good product, build stuff, and focus on those free users turning into paying customers.
Again, the caveat I gave before still stands. It's like some professor telling you, "I think this is great". But the nuances of your business mean it all depends.
For example, if you're selling to enterprise customers, I might build more free tools like the ROI calculator or HubSpot's Grader or something similar instead of actually building a free product in your tool.
If you don't believe there's enough developers out there that can utilize your product -- if there's not hundreds of thousands to millions -- then you might not want a free plan.
If the market is limited to 10,000 then I'd probably tell you, go deliberately target those 10,000 and lose your allergy to sales. Find the Benedryl for that or something.
At the end of the day though, it's just math. At the end of the day, I know all of you like math probably and it's binary actually. If you're compounding free users all the time, then all those free users have potential every month to turn into paying customers. You don't have to keep reacquiring that batch and on top of that, you're probably reacquiring new users too.
For me, the businesses that I've seen grow the fastest tend to be ones with free products. There's a lot of caveats. It is very difficult and the people that analyze pricing give me a hard time when I say freemium is what I would do but personally. But I wouldn't start a business today unless I could have a free plan on it. It just makes things easier. It makes the focus on product also the emphasis, not just sales and marketing.