Designing an onboarding experience for a AI product
Relevance AI is a no-code platform to analyze unstructured data
🌈 Designing solutions with principles
Let me share a few key principles that we applied to define and solve the problems below.
How we redesigned onboarding experience
Similar to many other products, our onboarding process is crucial for new users to fully access our product. However, we noticed a significant drop-off in the onboarding funnel. Through drop-off analysis using Amplitude, we discovered that our onboarding completion rate was only 61.2%. Essentially, we were losing one-third of our new users during the onboarding process.
So, what's the problem?
Then, why don't we just remove the onboarding? Why was the onboarding even necessary? This question leads back to the definition of acquisition as we should understand the value of the onboarding in the context of acquisition. Here is a definition of acquisition that I think makes more sense than others. (thank you, Ben Winter)
Acquisition: the number of users who 'interact' with your product
Our team recognized the importance of providing meaningful interactions to acquired users. This is why we considered the onboarding experience essential. Through onboarding, we can deliver the core values of our product and create a more personalized experience based on the data provided during the onboarding process. After conducting data analysis and heuristic reviews, we formed problem assumptions and applied several design principles to address the identified issues.
Through our analysis using Amplitude, we discovered a clear correlation between the time users spent on onboarding and the rate of drop-offs. This finding presented us with a key challenge: to design a faster and easier onboarding experience, reducing the time users needed to complete it.
To address this challenge, we began by focusing on the step that consumed the most time for users. Through extensive rounds of Fullstory reviews and user testing, we identified a significant issue with our previous design (refer to the image below). Users had to scroll through a lengthy dropdown list of categories and spend time understanding the meaning of each option under the "Level" section.
Principle 1. Less is more
Keeping those challenges in mind, we began to explore alternative solutions. We asked ourselves if there was a smarter and faster approach, or if there were design components better suited for a more streamlined multiple selection experience. Our next steps involved conducting design research and engaging in explorative ideation. After several iterations, we arrived at a new design (special thanks to Gage and Oliver for their inspiring contributions!). The key solution ideas focused on enabling users to:
-
Quickly find and select categories from suggested tags.
-
Easily understand and select the appropriate level option.
The impact of implementing this new design (refer to the image below) was truly remarkable. Users spent significantly less time on this step:
-
55 seconds → 18 seconds 🔥
This drastic reduction in time spent represents a significant improvement in the overall onboarding experience.
At Arc.dev, we start our design process by defining specific, measurable, and meaningful success metrics. This helps us drive product growth by focusing on user acquisition and improving the onboarding experience. We track and measure these metrics to evaluate the impact of our design solutions and make data-driven decisions. With the process, to drive product growth, we defined user acquisition and redesigned onboarding experience. Here is how we started off.
To optimize the user experience and increase acquisition, we wanted to gather important information about our users. However, we also understood that asking too many questions during onboarding could lead to higher drop-off rates. Instead of simply removing questions, we explored an alternative approach: asking certain questions at a later stage. Upon reviewing our onboarding process, we discovered that the initial question ("Are you actively looking for a job?") didn't significantly impact the acquired users' experience. Therefore, we decided to ask this question at a more appropriate timing after onboarding, with the expectation of reducing drop-offs in the onboarding funnel.
Principle 3. If not less, maybe later
Previously, we invested significant time in determining the questions to ask during onboarding. However, we realized that we hadn't effectively communicated to users why we were asking those questions and what they could gain from answering them. We hypothesized that this lack of clarity was the primary reason for the significant drop-off rate of 26% after the first question. To address this, we made sure to explicitly convey to users what they could expect by answering the questions during the onboarding process.
Another issue we identified was the concern that users might have about encountering another set of questions even after completing the onboarding process. We wanted to alleviate any worries or doubts that could lead to more drop-offs. To address this, we transitioned from a full-page onboarding experience to a modal-based experience. This approach allowed us to show users what to expect next, visually demonstrating the upcoming steps behind the modal.
Principle 2. Why before what
Our journey towards user acquisition doesn't end here; we will continue to strive for improvement. We can explore additional ways to apply the design principles we have discovered, or even uncover new principles that can enhance our acquisition efforts. Analyzing data from different perspectives can also provide valuable insights. As we move forward, there are still numerous questions to be answered and possibilities to be explored.
What's the next?
The results of our design experiment exceeded our expectations. Upon releasing the new onboarding experience, we observed a significant increase in the onboarding completion rate, soaring from 61% to 90%☝️. This remarkable growth of 150% translated into a substantial boost in user acquisition.
The outcome: 61% to 90%
✨ The objective
While many companies focus on increasing user traffic to attract more users, it's essential to consider the conversion of those visitors into acquired users. Simply driving traffic without ensuring user acquisition is like pouring water into a leaky jar. User traffic doesn't guarantee user acquisition. Acquiring a user means they take meaningful actions on your site, not just a brief visit. Buying promotional ads may increase traffic temporarily, but it's not a sustainable or scalable solution. That's why our focus is on improving the user acquisition rate, ensuring that visitors become valuable users.