Onboarding - designing towards the first value

role in the project
Research and Design
4 Months

Building my first onboarding experience

This is about what I learned and applied while improving onboarding for the past two years at ServiceBridge, along with some of the resulting outcomes and metrics. But first, familiarity with some concepts is needed.

The Pirate Metric: AARRR

When it comes to product, this is the metric referred to the most. The acronym stands for acquisition, activation, retention, referral, and revenue (AARRR). It can be viewed as a funnel, a timeline, or whatever an organization prefers. Please google it for a more thorough explanation. You can also watch this video breaking down onboarding at ServiceBridge.

Pirate metric as a funnel.

You can also watch this video breaking down onboarding at ServiceBridge.

When does onboarding happen?

Onboarding typically happens during activation. It may affect other areas too. For example, one could say that a great onboarding experience may also affect retention metrics.

Majority of Onboarding activities happen during activation

Planning to "pay off" six years of product debt

For over half of a decade, our engineering and sales teams gathered requirements from hundreds of our users to build out our feature-rich onboarding product. But the product became so complicated, it took almost an hour to demo and set up onboarding for a new user. This was unnecessarily unscalable as our sales team found itself devoting more time onboarding new users at the expense of selling the product.

Without proper onboarding complex workflow requires a lot of support.

UX Roadmap

The "Northstar" for this engagement was to release self-onboarding in four months. We planned on spending the time fixing the various usability issues; developing a design system, all while designing the self-onboarding process.

A very optimistic Roadmap

Underestimating the effort

During the first heuristic analysis, I found a lot of things that needed improvement. At first, we only prioritized the small quick fixes, those requiring the least amount of effort, but offering a lot of value. However, we kept uncovering dependencies, and the effort became burdensome.

Example from the Heuristic analysis report of multiple search and filter fields with similar icons.
New users struggled at every step of the way from Sign up all the way to subscribing.
Some of the errors were really hard to understand.

What is the right way to do onboarding?

We also found ourselves asking: What is the right way to do onboarding? Should you start with a product tour or a checklist, guided steps or video tutorials, webinars, or build great "zero states"? Well, it’s all of the above.

An effective way of learning about onboarding is from the companies that make the tools for it. AppCues, Pendo, Intercom, and others provide really great tutorials, white papers, and webinars. They have accumulated knowledge based on the many onboarding experiences they helped facilitate for others. One major commonality among them is that there is no one best way of approaching this task. People learn differently and have different needs. Some users will close out every helper tool and just explore. Others will read every support article, watch videos, and read email questions. The bottom line is that you have to maintain access to all resources so that users can find their own path.

Needed features, some we had some needed improvement and some needed to be built.

Should we build or buy?

To build a tooltip manager, checklist, webinars, zero states, support chat, and other learning aids, is quite a big task and will probably consume all a firm’s engineering efforts. I started looking at different solutions that could help us alleviate that task. We chose Pendo for tooltips, checklist, and the guided walkthrough, Fullstory for analytics, and Intercom for customer support and drip campaigns.

We evaluated a few products but ended up subscribing to Pendo, Fullstory and Intercom.

Three things to get started?

We learned we had to find the point of our first frustration. Onboarding doesn't start with the login. It begins much earlier. Maybe the current process sucks, it’s too expensive, or too cumbersome. Perhaps the existing software is just frustrating. Therefore, to determine where the process needs to begin, first, decide what job the product needs to do. Then identify the activation metrics necessary to keep the onboarding process on track.

Where do i get this information?

It’s not rocket science, but it actually is science. A hypothesis is needed. Ask your team, observe and talk to users, and then map the onboarding process backward. My experience found that talking to users took the most effort and created the least value. Scheduling the interviews, taking and synthesizing notes was very time-consuming. The most productive activity for me was talking to our sales and support team. Very quickly I was able to gather lots of crucial information.


We launched the checklist and the walk-through guide based on our hypothesis and starting monitoring how new users were interacting with it. Unsurprisingly, we saw some people dismissing the guides right away, while others would read every word (you could tell because the mouse cursor followed every line of text), and everything in between. We're still constantly adjusting the experience to make it better so that each of our new users is a happy user. As I learn more, I plan to add to this post.