Designing the first-time user experience for ARious to help young learners transition into 3D interactive lessons without the cognitive overhead that AR apps typically introduce, by treating the FTUE as the first lesson itself rather than a separate tutorial.
Young learners coming into ARious were excited by the AR concept before they downloaded the app. They had usually heard about it from a parent or seen it demonstrated. The first thirty seconds of using it tended to deliver on the promise: scan the target, watch the 3D content appear, feel some genuine wonder. The first thirty seconds were not the problem.
The problem started around minute three. The technology was new to learners, the device handling was unfamiliar, and the content shifted between 2D guidance and 3D interaction in ways that broke their flow. They would lose focus, the AR target would drift out of view, the lesson would pause, and the learner would feel like they had done something wrong. Several would put the device down and not pick it back up.
This was particularly consequential because parents who set up the app for their children were not always around to help. The product had to work for unattended young learners, not for guided ones. If a learner failed in the first session, they rarely came back, and the parents who had downloaded the app rarely launched it a second time once their child had associated it with frustration rather than delight. The product was losing users in the first session, not over weeks. The brief I took on was to redesign the FTUE so that the first session produced the same delight as the first thirty seconds.
Research focused on two questions: how young learners encounter new digital products, and what cognitive load actually feels like at that age. The literature on child cognitive development was useful here, and so were direct observation sessions with kids aged 8 to 12 using the existing ARious app and several competing AR products.
The observations produced several patterns that adult-focused product design instincts would miss. Young learners are quick to recover from low-cost failures and quick to abandon high-cost ones. A misread gesture that retries silently produces a shrug; a misread gesture that triggers a visible error screen produces a withdrawal. Adult products are usually fine with the error screen because adults can read it; child products cannot rely on reading as the recovery mechanism.
Young learners are also less tolerant of mode switches than adults. Switching from a 2D tutorial into a 3D lesson, then back to a 2D progress screen, then into another 3D lesson, produces context-switching fatigue that adults barely notice but that young learners experience as exhausting. The principle that emerged was to minimise mode switches throughout the first session, which had implications for how the FTUE itself should be structured.
The research also surfaced a counterintuitive finding about tutorialisation: young learners often resist tutorials more than adults do, because tutorials feel like school. Adult products can use tutorials as a respected onboarding pattern; child products cannot, because the tutorial framing itself activates a school-like response that suppresses curiosity.
Instead of a separate tutorial sequence followed by the actual product, I designed the FTUE as the first actual learning lesson with a gentler difficulty curve and embedded teaching moments. Learners practised core AR gestures in the context of doing something they wanted to do, rather than being lectured first and rewarded later.
The cognitive load this approach demands is high for the designer but low for the learner, because every gesture they perform is tied to actual content they care about. Designing this way required carefully sequencing the lesson so that each interaction the learner needed to learn was introduced at a moment where the lesson naturally required that interaction. A pinch-to-scale gesture was introduced when the learner needed to examine something closer. A two-finger rotate was introduced when the learner needed to look at the back of an object. The lesson taught the gesture because the lesson required the gesture.
Failure paths were designed to be near-invisible. Missed gestures retried silently. Lost focus self-corrected without dramatic recovery screens. Mistakes were never highlighted as mistakes. This was a departure from the conventional gamified feedback model that punishes errors with explicit losses, and it worked better for the audience we were designing for. Young learners are quick to recover from invisible failures and quick to withdraw from visible ones.
The emotional arc of the first session was deliberately scaffolded toward a moment of genuine accomplishment near the end. The first lesson built progressively toward a payoff that would feel earned, because the most reliable retention predictor we observed was whether the first session produced a feeling of accomplishment in the learner. A flat first session, no matter how well-designed mechanically, was less effective than a slightly less polished session that produced an earned win.
Make the first failure recoverable, ideally invisible. Young learners are quick to recover from failure when the recovery cost is low and quick to abandon when it is high. I designed the early failure paths in the FTUE to be near invisible. Missed gestures gently retried. Lost focus self-corrected without explicit error states. Mistakes were never highlighted as mistakes. This required prototyping the failure paths as carefully as the success paths, which is rarely done in product design but is essential for child products.
Reject school framing. Tutorial framing activates a school-like response in young learners that suppresses curiosity and engagement. The FTUE was deliberately framed as 'a lesson you're going to do' rather than 'a tutorial that will teach you the app'. This was a copywriting choice as much as a design choice, and getting the tone right took several iterations with actual young learners reading the text aloud.
Earned wins, not given wins. The first session built progressively toward a payoff that would feel earned, because the most reliable retention predictor we observed was whether the first session produced a feeling of accomplishment in the learner. Many child products give wins (you did it, great job, you earned a sticker). Earned wins (you figured this out, which was actually hard) produce stronger retention because the learner believes the win.
Minimise mode switches. Switching from 2D tutorial to 3D lesson to 2D progress screen produces context-switching fatigue. The FTUE kept the learner in the 3D lesson context as much as possible, with progress and guidance integrated into the spatial environment rather than separated into 2D screens.
Lesson-as-tutorial in action
Without quoting specific numbers, the proportion of new users who completed the first lesson improved noticeably in tests after the new FTUE shipped. The qualitative observation that mattered more was that parents started leaving the room sooner during the first session, which was the truest sign that the FTUE was working. Parents only leave when they trust the product to handle the child without intervention.
The pattern of embedding teaching moments inside real content rather than separating tutorial from product became the template for subsequent ARious lessons. The lesson-as-tutorial approach replaced the separate-tutorial approach across the product, which compounded the original FTUE work across all subsequent content the team produced.
The broader principle (FTUE is the first lesson, not a separate experience) has stayed with me across product design work in different domains since. Most products have a tutorial period that exists outside the real product. The best onboarding designs I have seen collapse that distinction, and that lesson came from this project.
The traditional onboarding model is closer to a manual than a moment. It exists outside the product and disappears once the user is 'onboarded'. Designing FTUE as the first lesson collapses that artificial distinction. The user is always onboarding in some sense, and the product should always be teaching in some sense.
I have carried this view into product design generally and it has made my approach to introducing complexity much more confident. Complexity should be introduced in the moment it becomes necessary, not in a pre-emptive tutorial that the user has to remember when they later encounter the complexity in context. The moment is when the user learns. The manual is when the user forgets.