Your onboarding has a shelf life

User onboarding for the first several years of Intercom centered on one action: install a snippet of JavaScript.

For our technical user base of developers and small startups, this was simple.

While our onboarding process remained largely consistent for quite some time, our user base was broadening. Many less tech-savvy people were signing up for Intercom, and the work to prepare the JavaScript wasn’t so easy for them.

As integral as UIs and snippets of code are to the onboarding experience, the context informing them – audience, environment, product, packaging and price – is just as crucial, and constantly evolving. In Intercom’s case, our context had changed. It was time for us to change our onboarding, too.

At our recent Inside Intercom event in San Francisco, my colleague Sean McBride and I explain why our growth team, whose top areas of focus include user onboarding, chose to redesign our experience rather than simply optimize it. We examine how context affects onboarding in all industries, why the software industry is often slow to react, and what a dedicated onboarding team can do to close the gap between context and conversion rates.


Stephen O’Brien: Onboarding as a concept is really broad. It encompasses a lot of things: how you get set up with product, how you learn about it, and how you purchase it, too.

Sean McBride: Most people here work on software. But all products in all industries have an onboarding experience. We thought it would be interesting to focus on a few examples from other industries before we dive into software.

Let’s start by looking at automobiles. Here’s a picture of a Ford dealership from 1956. Car dealerships are basically like the onboarding experience for cars. They’re the place that you come to see the different options that are available, to try them out, and to talk to somebody about them. This onboarding experience hasn’t really changed from 1956 to today. It’s a big, black box, or a big, glass box. It’s full of cars. You can talk to salespeople inside, but not a lot has changed about the onboarding experience for cars.

If we look at the broader context that informs the onboarding experience for cars, we can see that lack of change is reflected here too. We have the products that they’re selling. We have the environment in which they sell that product, as well as the audience that they sell to and their attitudes. There’s the packaging, which is the different options that they slice up from products and make available to people to choose from, and the price that they charge.

There’s been incremental changes in these things over that time period, but no massive shifts. Tesla is maybe the first car company that’s actually going to actually shift this formula.

Stephen: Let’s take a look at an industry that has changed a lot in its onboarding, fast food. If we go back to the 1920s, the most popular restaurant chain in the U.S. was Horn & Hardart, a company that none of us have heard much about.

Back in the ’20s, American cities were very different. The population density of a place like the lower east side of Manhattan was four times what it is today. Their onboarding challenge was how to serve a huge number of customers in a relatively small space efficiently, effectively, and well.

They came up this restaurant concept called the Automat, and as the name implies, they automated a lot of stuff in the restaurant. There were no waiters. You’d walk in, go to the back wall where there was an array of windows with different foods inside, choose the food you want, put a dime in, take the food out, and go back to your seat all without any staff involvement. These things were amazingly effective. The biggest one in New York was able to serve 10,000 people per day.

Fast forward 50 years to the 1970s and fast food looks completely different. It looks something like what we have today – places that you now drive to or even drive through, predominantly outside of city centers.

Let’s see what caused this change. Here, there’s been a monumental shift in the environment where people eat and where people live. Suburbia didn’t really exist in the ’20s. Car culture started to dominate.

As the environment changed, as the context changed, the onboarding had to react too. Really interestingly, we’re starting to see something of a regression in that environment. Cities like San Francisco have increasingly high population densities, so it’s no surprise that there’s a startup called Eatsa who has brought back the Automat. Except this time, you can only buy – and this is literally true – quinoa-based healthy foods.

Sean: What about software? Let’s come back to the world of software and look at, “How do the elements of context that affect the onboarding experience affect us in the software industry?” Well, really all the same categories apply. There’s the products that we sell, the environment, the audience, the packaging, and the price. If we look across Intercom’s history as a company, or in fact, the industry as a whole, actually all of these things have changed, are changing, and change quite frequently.

You’d think with so much change going on, onboarding experiences would change a lot as well to keep pace. Actually, from what we’ve seen, that’s not really the case. Onboarding experiences don’t usually change that much. There are many reasons for this, but one that we want to talk about briefly is the decision to optimize versus redesign. For onboarding, in our industry at least, the bias definitely seems toward optimizing even when that might not always be the right decision.

If we look at a simple model for when it would make sense to optimize versus redesign for an onboarding experience, we could look across two axes. The first one is just the change in context, all those elements that we just looked at since the last time you considered your onboarding experience. Have things changed a lot or have they only changed a little bit? The second axis is the performance of your onboarding experience. Is it meeting the expectations and goals that you have?

In the top left, when we have good performance already, and it hasn’t been that long, obviously this is a great time to be optimizing. You’ve got a strong foundation that you can build upon and continue to get incremental improvements in performance. On the other end of the spectrum, if a lot of time has past, there’s a ton of change in context since the last time you examined things, and things aren’t performing well, obviously there’s a case for redesigning – taking a step back and considering the constraints and opportunities in a new way, coming up with a new solution.

It’s the other two quadrants that are interesting to think about a little bit further. In the top right, you’ve got good performance, so it’s probably appealing to just continue to optimize and try to get incremental improvements out of that experience. But the shift in context that’s occurred might actually be masking some new opportunities that could unlock some serious potential in your onboarding experience. In other words, what used to be good performance might actually not be good performance any more.

In the bottom left, if you have something you’ve recently launched, but it’s not performing up to your expectations, it might be appealing to try optimizing incrementally from bad performance to good performance. Just taking a step back and considering things again may get you to good performance a lot more quickly, so perhaps redesigning, also, is more impactful here.

Stephen: Okay, so enough theory. Let’s take a look at a concrete example from here at Intercom. For the longest time, the core of an Intercom installation has been this snippet of JavaScript. You put this on the webpage or whatever app that you have, and you tell us some information about the currently logged in user: their name, their email, when they signed up.

You’ve got to do a little bit of work to prepare this. You’ll typically have to write, use a programming language to dig into your database, take the data out, and dump it on the page to tell us who the user is.

Our onboarding challenge was simply, “Help as many customers as possible to correctly install this JavaScript snippet.” In 2011, Intercom is still in Beta, super early days, long before my time. We were super proud of this awesome little JavaScript snippet, and it was right there on our marketing site. This is clearly telling us that the audience of Intercom at the time were tech savvy, developers, small startups. A year later, we’ve got a real signup flow. We’re no longer in Beta, but it’s still the same snippet of JavaScript.

Fast forward to 2013, one year later, and it’s still there. It’s been spruced up, but it’s exactly the same snippet. One year later, it’s there again. By this point in 2014, this idea has pretty much been optimized to death. There’s now a tab between engineers or business people. If you click on business people, you just get to email your engineer. Then we’re giving you bunch of hints for different languages, we’re telling you to come hang out with us in our install chat room to coach you through this or to watch our super-simple 30-second install video. We’re doing everything possible to try and get you to install this JavaScript.

Sean: In the meantime, we’re also receiving requests from existing Intercom customers about importing CSV data into Intercom. They have additional data that they want to layer on top of the users they already have stored in Intercom. We’re like, “That’s great.” Design and build a CSV importer that people can use to get extra batches into Intercom after they sign up. The breakthrough came when we realized this importer could be useful for people who are signing up for Intercom and are brand new to the product.

That led to this realization of a shift in context that actually needed to affect the onboarding experience in a bigger way. Our audience had changed. We had started out with really tech-savvy people and moved to a broader audience that was a lot less tech savvy, so it was much harder for them to install the JavaScript. Instead we needed to broaden our perspective on what the onboarding problem was and instead focus on how to get users into Intercom.

Stephen: This is where we ended up with our first redesign. This is exactly how it looked, the first thing we shipped after redesigning. For the first time was had a new second possible way to get started with Intercom, import users from CSV. It’s presented here as an equal choice, side by side, with our traditional JavaScript install. How did it work? It worked tremendously well. The three months before we launched this, our conversion rate from emails to signed-up apps hovered around 30%. We launched it in August of 2014, and the three months following, our conversion rate had risen dramatically, 15 percentage points or 50% overall to 45%.

I think even more incredibly, we hadn’t actually cannibalized the standard JavaScript installer. You can see in the same month here that the traditional JS installs are slightly growing too. This green area on the bar chart here, this is a totally new audience who we’d been absolutely neglecting with our previous onboarding. Our context has changed, our audience has changed, so it was time for our onboarding to change too.

Sean: Of course, redesigns have a shelf life too. All that context that changed initially and gave you the opportunity to try something new will continue to change, and so eventually you’ll have to react again. In 2014, where we left off, we’ve just launched this. It’s got good performance. This is the perfect time to optimize. We’ve got a solid foundation that we can build upon and a thesis to use, and so that’s what we did. We just added more ways for users to integrate with their product or import their users into Intercom from additional services that they might be using. Of course, this has diminishing returns over time.

Stephen: Meanwhile, our product team back in Dublin is working on a brand new Intercom product called Acquire, which actually just launched. Acquire is a really interesting product because it’s the first way in Intercom that you can talk with logged-out users – people who aren’t signed into your website or the app. If we go back to that original snippet I showed you guys, Nikola Tesla and everything there, if a user is signed out we don’t need to know their information in Intercom. The snippet in this new case is much simpler. It requires no customization whatsoever. Actually, a profound change has happened to our product. That’s the part of the context for onboarding, so we should be thinking about what ramifications this has.

It provokes this question, “What if all brand new customers start with this new, highly simplified JavaScript snippet?” It works on any page in the entire web. It doesn’t require any configuration. Requires very little technical skill. If you really need the user data, maybe you can add it later. Actually, our reframing from the last time of the onboarding challenge, how to get users into Intercom, might change again. Maybe it’s, “How do we show new users the magic of Intercom as quickly as possible?” We’re working on this behind the scenes, and we’ve got a pretty cool flow that we’re going to be launching soon.

Sean: If you ask me, “What has made the biggest difference at Intercom in our ability to focus proactively and successfully on our onboarding experience? What would it be?” By far, the biggest problem that I’ve encountered is that it’s just really difficult to hold the perspective of both new onboarding customers and existing steady-state customers in your head at the same time. Plus, it’s really hard to prioritize between features that target these two different audiences. By far, in my mind, the most impactful thing we’ve done at Intercom is make the decision to have a team that can explicitly focus on onboarding customers as their top priority while we have the other teams that can focus on other priorities.

This separation and specialization extremely simplifies the prioritization process and makes everybody more focused. Of course, the boundaries between the teams are fuzzy, but they always are. We navigate them, and we continue to move forward and cooperate and work together. At Intercom, the team that focuses on onboarding user perspective is the growth team, which we both work on. We’re here in San Francisco, and we do all those things that we talked about that effect the broader onboarding experience.

I’ll leave you with this: If you haven’t thought about your onboarding in a while, if you don’t have somebody at your company to think about optimizing versus redesigning when one or the other is appropriate, in short, if you don’t have a team to focus on onboarding as their top priority, then maybe you should. We’re excited to talk to you about our experiences building that team right after this. Thanks.

Want to be part of and contribute to the constant evolution of how we onboard our customers? We’re looking for the best Product Managers out there. Come join us :)

Intercom Product Tours