Convincing potential users to sign up for your product isn’t easy. But what happens next is far more important.
The latest batch of billion-dollar companies are built on making their users successful, and that means providing great onboarding. At Traction Conference, an event all about how to keep and grow customers and revenue at scale, I explained how to build onboarding based on your customers’ goals, and why when your product improves, your onboarding must improve with it. You can watch the talk below.
Prefer a written account? What follows is a lightly edited transcript of the presentation.
I’ve been working with my co-founders since 2007-2008. The nature of traction itself has changed quite a bit in that period. Way back when we were trying to grow our first company, which was called Exceptional, a Ruby on Rails error tracker, all we ever stressed about was the funnel. The idea of getting more things into the funnel was all we cared about. If we did enough of that, we’d get rich, right?
Through a constant series of failure and disappointment, we realized there is no one true way. There’s actually a lot of ways and they all add up together to matter. We gave up chasing these mythical, “grow your website with this one weird trick” things. There were lots of tricks and they weren’t that weird. They just looked like hard work.
Over time as an industry we realized there is no magical one way to become a billionaire overnight. You do a lot of work and hope people will eventually hear of you. Then you get them to your website. Our next challenge was, what the hell happens when they land on our website? Is this the right website? What could we do about that? Could we somehow get them to definitely sign up? We developed all these tips and tricks that would help us optimize what happens when they land. Most specifically, the industry came up with this beautiful idea of AB testing. This was around 2010. It became standard fare to find the best possible way you could lay your page out, which was a positive thing.
It is fair to say that we learned a lot of good stuff. We also learned a lot of stuff we could’ve worked out otherwise. Big calls to action outperform small ones. Calls to action in the right location outperform ones in the wrong location. Calls to action that are visible outperform those that are not visible. You’d like to think we learned these things once, but as an industry, every three weeks it’s like, “We increased our conversions by 40% by making our call to action visible on the home screen.” I’m like, “Good job, guys. We’ll get there.”
As we’ve tried to tackle traction at each step, we’ve found ourselves chasing these silver bullets. We’ve spent a lot of time searching for one of these. To paraphrase Ben Horowitz, it’s actually a shitload of these that we need. Real effective things, not mythical weird tricks.
The changing nature of distribution
If there are no shortcuts in the top of the funnel, it just looks like hard work. There’s no one certain way to design a landing page, it just looks like great design, preferably with a good product to back it up. Distribution has changed in nature because of these realizations. To bring us up to early 2014, you might be familiar with Product Hunt. You just go there every day and there’s more to check out. It is effectively the app store for the web, but it’s weirdly even more addictive. As a result, distribution got easier and easier and easier.
We found ourselves in a world where a product like Yo, which let you send yos to people, gets a million users in four days. That is insanity. Meerkat had two million users. Secret had tens of millions of users. Whisper has 20 million users.
What you realize is today there are dozens of ways to get your first hundred, your first thousand, your first 10 thousand, or even your first million users, so much so that Benedict Evans tweeted:
A cool new messaging app getting to 1m users is the new normal. Keeping them, and getting to 100m, is the question.
— Benedict Evans (@BenedictEvans) March 22, 2015
That’s where we’ve found ourselves, that it’s not just about getting everyone to hit your website. There’s actually something deeper going on. That whole idea is this concept of retention, and it’s the next frontier. We’ve gone from how the hell can we get people to hear about us, to how the hell can we get them to sign up when they land on our page, to how the hell can we get them to stick around? That is the next funnel – the second funnel.
If you want hype, bullshit, and press, you can hack the hell out of the first funnel. If you want to build a billion-dollar business, you need to solve the second one. To give you a simple example, which of these two would you rather have: a million sign ups in four days, or a million active users after sixteen months? The first business is Yo, and Yo is gone. The second business is Slack.
The question then becomes, how do we get from sign-up to success? Intercom has grown like a weed since I’ve been involved, which is August 15, 2011, our incorporation date. It’s been interesting to think about what we’ve done right and what we’ve done wrong. I don’t look at Intercom and think that we nailed marketing or that we nailed comms or media relationships or virality. There are some things we got right, and it hasn’t necessarily been about filling top of funnel.
We realized early that you can have all the tactics in the world but if they hit your homepage and your homepage is bad, it’ll go to shit. If users sign up and you don’t have them doing anything useful after that, it’ll go to shit. If they go through onboarding and your onboarding sucks, it’ll go to shit. If you get it all right, they’ll become a happy customer, at which point your new obsession becomes keeping them. If you don’t keep them, they’ll either leave or go to another customer which means, again, it’ll go to shit. If there’s one point you should take from this, it’s that you have to work very, very hard to stop it all from going to shit.
Onboarding is, in a sense, the new conversion. It’s the new thing we need to think about. The question you have to ask is, what makes for a successful onboarding? I have so many businesses who will tell me things like our cost per lead is $10. Would you consider spending two cents to increase the chance that lead becomes a customer? They’re like, “That sounds expensive.” You’re literally pouring money into a leaky bucket otherwise. Onboarding is actually the challenge of getting people from thinking, “what the hell is this product?” to, “holy shit, this is quite a product.”
The definition of successful onboarding
When most people think about onboarding, they think it’s the bit where we try to get the email. There’s actually a lot going on with onboarding. Is it the idea of trying it out? Is it buying and paying? Is it being set up for success, or is it progressively expanding over time and learning about new changes? There is a lot to it.
Onboarding is the new conversion.
The biggest thing I find myself beating into the heads of either startups who I chat with or new folks who start at Intercom is this idea that if you’re doing things right, you’re going to ship a lot. If you’re doing things really right, you’re going to ship all the damn time, which means your product is constantly in flux. It is nothing short of a moving target.
Every time your product changes, your definition of success for a customer has to change too, which means that every time you write anything substantial on your product, you need to reconsider what a successful user looks like when they land on your page. What are the things you want them to do, in what order, and how?
Signing up for your app is literally the only thing that everyone does. You can tell me, “We’re going to nail our new calendar feature or export to Dropbox.” I’m like, “Maybe if you’re doing things really well, 30-40% of your customer base will use that. Have you looked at this thing that literally everyone does?”
Most product owners sign up for the product once. This is the problem. When they signed up, they were literally granting themselves access to the product in a database somewhere. It’s not the same thing as the funnel that you’re asking and paying for your customers to go through. If you don’t continuously reimagine what it’s like to land on and start using your product, all this stuff goes out of date. Your product tour is out of date, your help docs are out of date, your welcome email is out of date.
All that stuff will go wrong, which means that a neglected onboarding is actually net negative over time. You’re literally pointing people to yesterday’s ideas of your product, not today’s, which means you’re literally pitching them on the wrong shit when they decide to use your product. Meanwhile you have your product team working on some awesome stuff that users are never going to bump into.
Never stop signing up for your own product. It should be someone’s job to sign up for your product every week and make sure that everything still works well and makes sense.
The current generation of billion-dollar businesses are obsessed with onboarding. Slack built this incredible Slackbot, which does two things in parallel: it talks to you and teaches you how to use their product while collecting all the information, and then it suggest things that you should do next, and it works.
You can argue Stripe’s entire product is simply a much better onboarding for payments. They made it incredibly easy to start getting paid and did that through onboarding. They’re famous for a thing called Collison install. In the early days, if you bumped into John or Patrick Collison, they would literally say, “You want to use Stripe? Give me your laptop,” and they’d actually do it for you. That was how serious they took onboarding.
The current generation of billion-dollar businesses are obsessed with onboarding.
Now, you can’t onboard everyone. It’s a mistake to try because you’ll end up bending your product around everyone. You can only onboard people who have a need, a desire, and a capability. No two of these three are enough.
If they need it and desire it but they don’t have the authorization for purchase or they don’t have the skills to install or whatever, then you won’t be able to bring them on board. Similarly, if they want it and are capable of doing it but they don’t need it, it’s not going to work. That’s like me going to look at Oculus Rift. I really want this thing and I’m capable of buying it, but I can’t justify it right now. Lastly, if you need it and you’re capable of doing it but you don’t want it, it’s a hard onboarding. This is like me going for a tooth extraction. Yes, I need it. Yes, I’m capable of booking one, but I don’t want it.
When to optimize and when to redesign
The first wave of onboarding, going back five years, was simply pointing out the UI that we just built, because we need to explain it. That was onboarding 1.0. Onboarding 2.0 was like LinkedIn’s style – let’s give people progression meters so they know how far they’ve gone.
Modern onboarding requires you to understand your user’s definition of success and break down the barriers to get them there. This is why tips and tweaks won’t get you there. You can’t always optimize your way to better onboarding. Optimization can very quickly become a black hole where you sink all your team’s time and get literally very little in return. You need to know when to reset.
Your product changes either a little or a lot, and your onboarding either works or it doesn’t. If your product doesn’t change a lot and your onboarding is good, that’s a great time for optimization. If a lot has changed and it doesn’t work, that’s a great time for redesigning the whole thing. If things have changed a lot and you’re working, you should probably go back and reimagine it anyway, and if your onboarding isn’t working, you should fix it. There’s only one great case for optimization, which is we’re doing the right thing and we need to make it better.
We’ve been working at this forever, and it got more fancy over time. When 2014 arrived, we had engineers, business people, videos, our chat room, etc. We had everything covered and optimized. At the same time in our product land, customers are telling us, “I actually just want to upload users from a database.”
Simply doing that gave us a 40% kick in our conversion rate, which is huge. Most curiously, we found it didn’t cannibalize any of our existing stuff. This was net new customers. It wasn’t like all of our old people were going through this route. It was a whole new type of people we were getting. We had taken people who need it and want it but can’t make progress, and we let them make progress. Now they need it, want it, and they can use it. That was huge.
A relentless focus on high impact work.
The work your growth team or your onboarding team does can be high impact or low impact. It can either be a lot of work or a little work.
Take a look at the four quadrants above. Most people are good at staying out of the lower right. That’s a lot of work and does nothing. The top left tends to be quick wins, but there’s very few of those. The top right is your strategy.
The hardest thing you’ll ever do is get your team to stay out of the lower left. That work is easy but does nothing. When it does nothing, you say, “But it was easy.” And because it’s easy you say, “Well, we don’t expect it’ll do much.” You have this circular logic where you spin your wheels and you produce nothing. Only focus on high-impact work in onboarding, because there are so few occasions when it makes sense to tweak.
Billion-dollar businesses of this generation will be built on great onboarding. It’s not about everyone hearing about you; it’s about everyone who does hear about you using you successfully. Great onboarding is built on the understanding of your customer’s goals. Do they need it, do they want it, and are they capable of progressing?
Your onboarding has to improve every time your product improves. Don’t snack. Don’t focus on the low impact, low effort work. Focus on impact.