Short-form video platforms are engineered for engagement, not user wellbeing. I led research on a TikTok extension that uses Just-In-Time Adaptive Intervention (JITAI) to interrupt passive scrolling — without removing user autonomy. Across 22 literature sources, 5 interviews, 2 focus groups, and 16 think-aloud sessions across 2 iteration rounds, only sentiment (VADER) showed statistically significant improvement (VADER: +0.092, extremely large effect size) and avatar emotional investment trended toward significance (p<0.1, large effect size). SUS, avatar personalization, and usability questions trended positive but did not reach significance with this sample size. Honest read: the concept is psychologically grounded and validated by two behavioral-science experts; the measurement design needs a larger N to detect the effects we suspect are real.
22
Paper literature review
27
Participants across interviews, focus groups, & 2 testing rounds
2
Expert clinical reviews
+.092
VADER sentiment · extremely large effect, p < .05
The Problem
Brain rot is a design problem.
Short-form content platforms are engineered for maximum engagement, not user well-being. As a result, users frequently experience a loss of control over their time, attention, and intentions. Project Rewire is a behavior-change intervention that introduces intentional friction, real-time awareness, and adaptive feedback loops to help users regain control without removing autonomy. Unlike traditional solutions (screen time limits, app blockers), Rewire focuses on aligning user intention with actual behavior, not simply reducing usage.
Passive, automatic scrolling
Attention fragmentation
Reduced critical thinking
Cognitive fatigue
Emotional overstimulation
Weakened social connection
Design Question
How might we redesign traditional gamification elements to foster more meaningful engagement on social media, and reduce the cognitive effects of brain rot?
01 — Research
Five methods. One clear picture.
22
Literature sources
7
Competitors analyzed
5
1:1 Interviews
2
Focus groups
2
Synthesis methods
UX Research Artifacts
Three research deliverables synthesized from a 22 article literature review, competitor analysis, and interviews/focus groups with a total of 11 participants.
ArtifactCurrent-State + Future-State Journey Map
Think-Aloud + Interviews
Current State
With Rewire
ScopePassive scroll session, evening context
01 — User Journey Map
What it feels like to lose control.
A dual-state map contrasting the current unintentional scroll loop (red) with the intervention-supported journey enabled by Rewire (purple). Grounded in think-aloud session data and behavioral self-reports. The emotional arc is the diagnostic tool — the gap between the two lines is the design opportunity.
Sara L.
Primary Persona · "The Aware Addict"
Age22 · College senior
Usage~2.5 hrs/day TikTok
TriedScreen Time, deleted + re-downloaded
GoalUse TikTok intentionally, for learning & connection, not default boredom fill
DriveOwl (Smart archetype/Learn Drive): seeks novelty and knowledge
"I know it's bad for me but I literally can't stop. I open it without even thinking."
Post-session: intended vs. actual time, behavioral trend. Session becomes a learning loop, not a waste.
ArtifactThematic Affinity Diagram
Interviews + Focus Groups + Literature
Participants11 (5 interviews + 2 focus groups)
Notes coded~90 observations → 5 clusters
02 — Affinity Diagram
Five clusters. One system shaped by behavioral compulsions.
Raw observations from interviews, focus groups, and behavioral science literature were clustered into five thematic groups. The consistent thread: users are not failing; the platforms are using compulsive behavior patterns. Every cluster points back to a systemic design problem, not a willpower deficit. Five underlying requirements emerged.
5
Underlying Requirements
Relief from boredom or stress
7 observations · Mental decompression
"When I need a good laugh or I'm looking for inspiration,""[it gives me] self-expression and creativity.""laughter, inspiration, escape""I scroll when I'm waiting for literally anything.""Comedy brings me joy in the day.""I've opened it 5 times while trying to stop."Habit loop: cue → routine → reward fires without conscious mediation
Sense of Community & Belonging
5 observations · Meaningful Connection failure
"I appreciate the newer influencers speaking about their own experiences. I feel better even though I don't actually know them.""Finding solutions in common hobby groups [Subreddits] is helpful social media to me.""I feel informed... trying to keep up on news""...shared art boards or styles, I really grow from feedback but constructive comments [versus mean comments with no guidance] are hard to get.""I like finding people that are going through the same things as me."Users find a sense of connection with others, even those they don't know, through shared experiences or emotions.
Balance between Engagement & Wellbeing
7 observations · Responsible social connection
"I almost always feel like I could have done something better with my time.""I feel stuck & can't stop""Fatigue and overstimulation: Long sessions cause exhaustion and anxiety.""I feel FOMO if I'm not on it.""I look up and it's dark outside. Where did my day go?""I use it for everything from information to inspiration to just random funny videos."Social media satisfies Drive to Bond & Learn — but in a meaningful, responsible way
Alignment with Personal Identity
4 observations · Resonating with users
"If I don't vibe with it, I just swipe.""I like being able to share my experiences and how I see it [through art].""I follow people that have the same experiences as me.""I really miss how you could put so many 'favs' on MySpace, like top songs and glitter fonts or whatever."The change mechanism needs to feel like it matches users' style or they'll abandon it.
Progress & Recognition
6 observations · Visible Achievements & Results
"Honestly, I just want to win.""I like to goal set in ways that adds, not take away.""Winning... getting to that goal keeps me coming back.""Seeing my pieces come together entices me to keep going.""The satisfaction of completion... I was able to figure it out.""As long as I can see I'm getting somewhere, I'll keep trying."
Research Synthesis: Key Insights
01
The need is real. It's the delivery mechanism that is problematic.
Clusters 1, 2, and 3 reveal that social media satisfies genuine social and cognitive needs: learning, belonging, decompression. The goal is not to eliminate these needs but to fulfill them more intentionally. Rewire's challenge system is a direct design response to this finding.
02
Personalization is key.
Cluster 4 shows a clear pattern: users need something that meets and works with their goals and their mental models.
03
Seeing results keeps the momentum of change.
Cluster 5 reveals users need to visibly see and be reminded of their goals when trying to change to keep themselves on track.
Across seven gamification and engagement platforms, no competitor builds toward intentional behavior change grounded in behavioral science. Most rely on extrinsic reward loops — points, badges, leaderboards — without addressing the psychological root of passive or automatic engagement. Rewire occupies the intersection that others miss: intrinsic motivation, adaptive timing, and autonomy-preserving design.
Platform
Primary Context
Gamification Model
Behavioral Theory Depth
Motivation Type
Personalization
Real-Time Feedback
User Autonomy
Chatter Salesforce
Enterprise social collaboration
Contribution scores
Extrinsic
✗Feed-based, passive
SharePoint Microsoft
Document management & intranet
None native
None
✗No behavioral feedback
Packback Education · AI
Higher-ed discussion & inquiry
Curiosity Score
Intrinsic
✓AI in-line as you write
Bunchball Nitro BI Worldwide
Enterprise gamification
Points, badges, missions
Extrinsic
~Push notifications, missions
Centrical fmr. GamEffective
Frontline employee performance
Adaptive narratives
IntrinsicExtrinsic
✓AI coaching + microlearning
Badgeville Acq. by SAP · Deprecated
Enterprise behavior mgmt.
Behavior Platform
Extrinsic
~Analytics-driven, not live
Kudos Recognition platform
Employee recognition & culture
Peer recognition
Extrinsic
~Recognition feed, pulse surveys
Rewire ↗ Project Rewire · MindOS
Behavioral intervention · TikTok
Adaptive JITAI
Intrinsic
✓Avatar-driven, mid-session
Positioning Map — Intrinsic Motivation Focus x Behavioral Theory Depth
Strategic White Space
Intrinsic + Theory-Grounded + Consumer-Facing
No competitor occupies this quadrant. Enterprise platforms (Centrical, Bunchball) have theory depth but target employees, not consumers. Consumer tools (Chatter, Kudos) rely on extrinsic social loops. Packback comes closest but is locked to education. Rewire is the only product built on JITAI + 4DT + CoI for everyday users.
Critical Gap 01
Extrinsic rewards don't produce lasting change.
Bunchball, Badgeville, and Kudos are built on points-badges-leaderboards mechanics. Research shows extrinsic reward loops produce short-term spikes but erode intrinsic motivation over time — users disengage once novelty fades. Rewire is designed around intrinsic alignment, not reward dependency.
Critical Gap 02
No platform intervenes at the moment of behavior.
Every competitor operates in scheduled, pre-set, or post-session modes. None use adaptive, moment-of-behavior intervention (JITAI). Rewire's mid-session intervention is a structural differentiator — it interrupts the automatic habit loop at the highest-leverage point.
Chatter and SharePoint are communication infrastructure — they have no behavior change intent. Their "gamification" (contribution scores, badges) is a peripheral feature designed to boost platform adoption, not reshape how users relate to attention, time, or intention.
Rewire's Unique Differentiators
Motivation Model
95% Intrinsic
Rewire is alone in targeting identity-level, values-driven motivation. Every competitor leans on external reward structures. Rewire leans heavily on intrinsic motivation, with extrinsic motivators used only to help users get started.
Intervention Type
Mid-Session JITAI
The only platform that intervenes during active behavior — not before setup or after the fact.
Theory Foundation
JITAI + 4DT + CoI
Three behavioral frameworks integrated into one system. Closest competitor (Centrical) applies one framework, in enterprise contexts only.
User Stance
Non-Coercive
Every prompt is dismissible. No penalties, no shame mechanics, no points withheld. User agency is the non-negotiable design constraint.
02 — Key Insights
Five patterns. All structural.
Research consistently pointed away from user failure and toward platform design. The problem isn't willpower — it's systems working against users' own values.
Insight 01
The Passive Consumption Loop
Users open TikTok for relief, get stuck, and feel worse. Relief → trap → guilt → repeat.
Insight 02
The Dopamine Cycle
Short-term reward → overstimulation → emotional fatigue → scroll again to reset. Users recognize the cycle but can't break it alone.
Insight 03
The Value Disconnect
Users want learning and connection — but default to mindless content. The gap between stated values and behavior is the design opportunity.
Insight 04
Intrinsic Motivation Matters
Gamification fails with superficial rewards. Identity-driven engagement — goals that feel personal — is what actually sustains change.
Insight 05
Social Media as Emotional Tool
Users scroll for belonging, humor, and identity validation. Any intervention must address the emotional need, not just the behavior.
Opportunity
The Design Space
Intervene at the exact moment of behavior. Redirect, don't block. Give users control. This is the JITAI (Just-In-Time) Adaptive Intervention framework in practice.
Theoretical Foundations
Four Drive Theory
Avatar system + personalization
Maps users to a personality-aligned avatar via core drives: Learn, Bond, Acquire, Defend.
Just-In-Time Adaptive Intervention: interrupt at the exact moment it occurs, not before or after.
03 — The Solution
A behavioral system. Not just an app.
Project Rewire detects prolonged scrolling, prompts through a personality-matched avatar, and redirects users into one of three challenge types — all dismissible, all on their terms.
Feature 01
Avatar System
A short quiz matches users to one of four personality-aligned avatars. The avatar acts as a guide, building identity, trust, and intrinsic motivation over time.
Feature 02
User-Controlled Intervention
Users set their own scroll threshold. The avatar appears with a prompt — always dismissible. Autonomy is the non-negotiable core of the system.
Feature 03
Challenge System
Three types mapped to CoI: Learn (trivia), Connect (message a real person), Reflect (journaling). Short, low-friction, and genuinely meaningful.
Feature 04
Rewire Mode
An alternative TikTok feed showing only high-quality, meaningful, socially valuable content. Opt-in — an upgrade, not a restriction.
Four Avatars · Four Drive Theory
Owl
Drive to Learn
Knowledge-seeking users who grow through content
Elephant
Drive to Bond
Community-oriented, motivated by connection
Lion
Drive to Acquire
Goal-driven, motivated by achievement
Beaver
Drive to Defend
Values-driven, habits reflect identity
Ethical design stance
Non-coercive. Transparent. User-controlled.
Nothing is blocked. Every prompt is dismissible. Users set their own limits. Project Rewire is a non-coercive behavioral system — not a parental control.
04 - Design Iterations
What changed. And why.
Every iteration was driven by a specific failure observed in testing, not aesthetic preference. Three areas went through the most significant rounds of change: the microcopy, the challenge system, and the avatar onboarding flow.
01
Microcopy
From excessive to just enough
Before
Problem identified in testing
Most participants indicated that they wanted more information about what Rewire does, but the level of detail we included required too much reading.
Insight drove change
After
Design decision
Microcopy and other informational points (i.e. challenge scoring) was written to be easily scannable and fast to digest.
Result
In follow-up think-aloud sessions, reported a better understanding of Rewire and ease of informational text.
02
Challenge System
From just gamification to genuine engagement
Before
Problem identified in testing
Users only had the option to complete challenges, not send them. We were unintentionally preventing engagement opportunities that users wanted.
Insight drove change
After
Design decision
We had actually not made a Make A Challenge ability originally, relying in AI to determine users' level and needs. But users' wanted more control and engagement. Additionally, tap over text is important. Easy choices engagement opportunities created an easier, more seamless experience.
Result
Participants' post-session qualitative feedback shifted from describing challenges as "tasks" to "things I actually wanted to do," a direct signal that the intrinsic motivation framing was working.
03
Avatar Onboarding
From personality test to personal commitment
Before
Problem identified in testing
Participants were skipping their avatars by just clicking next or not identifying with it.
Insight drove change
After
Design decision
We added the ability to name the avatar as soon as it was assigned, giving users a chance to personalize it and prevent them simply clicking next and skipping it.
Result
Avatar alignment scores improved by trending towards significance in follow-up testing. More importantly, users reported their avatars "matched" their own personalities more.
What iteration taught us
Reactance
Any design that feels controlling, even subtly, triggers resistance before the user processes the message. Autonomy signals must be visible immediately.
Cognitive load
More options reads as more effort. In a moment of attempted behavior change, friction is the enemy. Three choices beat ten every time.
Identity language
The difference between "your personality type" and "your guide" is the difference between a label and a relationship. Copy is UX design.
Evaluation
Testing the system.
Current Testing
Avatar alignment — does the matched avatar reflect self-perception?
SUS (System Usability Scale) — baseline usability score
Think-aloud protocol — real-time reasoning during prototype interaction
Future Validation Plan
Stroop test — attention and cognitive control pre/post
Memory recall — effect on working memory
Critical thinking assessments — depth of reasoning over time
Screen time reduction — actual TikTok usage change
Mood tracking — longitudinal self-reported affect
Usability Results
Overall results of usability.
Across two rounds, three comparative questions tracked whether the avatar system felt personal, engaging, and psychologically aligned. The layout below preserves space for your final boxplots while matching the rest of the case-study structure.
Comparative Analysis · Rounds 1 & 2
Overall Results of Usability
Q1
Did customizing the avatar make the experience feel more personal or meaningful to you?
Boxplot
Iteration 1
Iteration 2
Q2
In your opinion, would the experience feel less engaging if the avatar were removed?
Boxplot
Iteration 1
Iteration 2
Q3
How much do you feel that the avatar you received matches your own personality?
Boxplot
Iteration 1
Iteration 2
Boxplot scores on a 1-5 Likert scale comparing Iteration 1 (N=8, ages 25-34) and Iteration 2 (N=8, ages 23-28).
Testing Metrics
SUS and VADER results.
These two comparison views isolate the broader usability signal from the emotional one: the System Usability Scale captures perceived ease of use, while VADER highlights whether the second iteration produced more positive sentiment overall.
Comparative Analysis · Rounds 1 & 2
Results of Usability Testing
✕
SUS Score
Boxplot
Iteration 1
Iteration 2
✓
VADER
Boxplot
Iteration 1
Iteration 2
Comparing Iteration 1 (N=8, ages 25-34) and Iteration 2 (N=8, ages 23-28). ✕ no statistically significant change·✓ improved
Qualitative Findings
Think-aloud findings across two rounds.
These interview cards translate the two qualitative sandbox files into the case-study’s existing structure: round-level takeaways, quoted wins, and observed points of friction, all within the same content width as the rest of the page.
Qualitative Findings
Results of Usability Testing — Round 1
Qualitative
Method
Think Aloud
Wins
"
Okay, I'm loving these stats [insights] and that you can view them really easily.
— Participant #1
"
Oh I'm an Owl! How cute! I love it!
— Participant #3
"
I like the app, though. You know, like I can see why somebody will probably want to use this.
— Participant #9
Challenges
"
I'm confused what to do here. What is XP?
— Participant #4
"
what am I doing? I think it's cool. But is it like, is it [Project Rewire]?
— Participant #1
"
I like the badges but I'm not going to sit here and read each one.
— Participant #9
Quotes captured verbatim during think-aloud sessions · Round 1 · N = 8 · Ages 25–34
Qualitative Findings
Results of Usability Testing — Round 2
Qualitative
Method
Think Aloud
Wins
"
The everything…the whole challenge flow was very easy.
— Participant #18
"
It [the avatar assignment] is very true about myself.
— Participant #12
"
Look, this is fun. I've literally been wanting something like this”
— Participant #17
Challenges
"
I feel like this is intrusive and I don't want to read this much
— Participant #11
"
Mm-hmm. Like just being like. I'm your owl, I'm THE owl. I'm something like, hello, my name is blah blah blah. [Give me more personalized, more in-depth or more personal to ME individually]
— Participant #17
Quotes captured verbatim during think-aloud sessions · Round 2 · N = 8 · Ages 25–34
Results Summary
What changed across rounds.
This synthesis compresses the full evaluation set into one view: three avatar-related questions, SUS, and VADER. The clearest movement came from sentiment, where VADER was the only statistically significant measure and showed the strongest overall effect. Q3 trended towards significance with a p<0.1 and large effect size.
MindOS · HCDE 591 · Project Rewire
Summary of results
Q #1not sig.
—
no mean change
var. -0.15
effect size
none
Q #2not sig.
+0.75
mean change
var. -0.44
effect size
medium-large
Q #3nearly sig.
+0.63
mean change
var. -0.43
effect size
large
SUSnot sig.
+7.19
mean change
var. -7.24
effect size
medium
VADERsig. ✓
+0.092
mean change
var. -0.031
effect size
extremely large
Q #1
Q #2
Q #3
SUS
VADER
statistically significant
approaching significance
not significant
bars show normalized mean change relative to each measure's scale range
Honest Results Framing
What the null results taught us.
Four of our five measures didn't reach statistical significance; however, three of five showed positive trends with medium to large effect sizes. That's not a failed study — it's data that sharpened what to do next.
The pessimistic read
The intervention isn't strong enough, or the avatar system doesn't meaningfully move usability perceptions.
The methodological read
N=8 per round is too small to detect the medium-to-large effect sizes the variance reduction suggests are present. Four of five measures showed reduced variance round-over-round (a real signal that the design is converging on something), and one (VADER) cleared significance with an extremely large effect.
My honest take
The methodological read is the right one. The next study needs ~30 participants per round, longer exposure, and behavioral outcomes — not just self-report — to actually answer the question.
Expert Review
Two experts. One clear direction.
Behavioral psychology feedback helped validate the concept while sharpening where Project Rewire needs the most care: execution, ethical clarity, and long-term behavior change support.
Dr. Siefert — Clinical Psychology & Behavioral Science
Dr. End — Research / Behavioral Psychology
Both experts validated the overall direction of Project Rewire. While they raised important cautions around execution, moderation, and behavior change limits, the broader takeaway was encouraging: the concept is psychologically grounded, strategically promising, and worth developing further through careful iteration.
Both experts agreed
✓
The problem is real and worth solving
Both experts treated the core issue as legitimate rather than manufactured. The concept responds to an actual behavioral and cognitive challenge users already experience.
✓
Friction matters more than persuasion
Prompts work best when they interrupt automatic behavior. The value is not in lecturing the user, but in breaking the loop long enough for reflection to happen.
✓
The avatar system has strong potential
Emotional investment in the avatar was seen as one of the most promising parts of the concept. It adds care and identity in a way other features do not.
✓
Execution and testing will determine success
Both experts framed validation as a starting point, not a finish line. Real user behavior and iteration will matter more than theoretical promise alone.
Key cautions raised
!
Behavior change will not be universal
The strongest impact is likely to come from users who already recognize the problem and want help changing it. This is not a universal cure-all.
!
Direct messaging needs safeguards
Direct messages can create the richest form of authentic connection, but also carry the most interpersonal risk. Moderation and structure are important.
!
Rewards should scaffold, not dominate
Extrinsic rewards are useful here, but they should support the deeper intrinsic motivation and goal of reflection, habit interruption, and self-awareness rather than becoming the whole experience.
!
The framing should stay ethically transparent
The concept was not seen as manipulative, but users should understand what prompts do, why they appear, and how much control they have over them.
What I'd do next
If I were taking this forward, I'd position Rewire as a friction-based scaffold for users who already recognize the problem and want help — not as a universal intervention. The next study would scale N to 30+ per condition, lengthen the testing window to 4–6 weeks, and replace single-session SUS with longitudinal behavior measures (Stroop, memory recall, actual screen-time data). The two behavioral-science experts validated the conceptual foundation; the work ahead is measurement rigor.
Expert validation
Dr. Siefert (Clinical Psychology) & Dr. End (Behavioral Psychology)
Two clinical experts reviewed Rewire across 14 questions on authenticity, harm potential, behavioral mechanism, and ethical framing. The consensus:
The problem is real and worth solving both
Friction matters more than persuasion both
The avatar system is the strongest emotional hook both
Low psychological risk; harm-reduction framing is appropriate both
Caution: behavior change will not be universal — the strongest effect is for users who already want help
Caution: extrinsic rewards should scaffold, not dominate, the intrinsic core
Full expert Q&A
14-question expert review
Final Design
The screens.
The final high-fidelity screens below show the core product flows: personalization, intervention prompts, challenges, insights, and reflection.
Hyper personalization — Avatar design
Personalization
Avatar customization
Intervention Prompt
Threshold Prompt
Avatar-driven scroll intervention
Challenge Selection
Challenge Selection
Learn / Connect / Reflect
Learn Challenge
Learn
Trivia + cognitive engagement
Insights
Insights
Overview of user's behavior for quick viewing
Reflect Journal
Reflect
Self-awareness journaling
Reflection
What this project taught me.
Key Contributions
Led research synthesis — interviews, focus groups, affinity mapping
Connected Four Drive Theory to concrete design decisions in avatar personalization, and CoI to challenge design decisions.
Drove the non-coercive ethics framework as a core principle
Wrote UX microcopy for avatar prompts, challenges, and onboarding
Created "Friend-Authored Challenges": using prosocial peer pressure to encourage engagement and connection
Ongoing Challenges
The hardest tension was building something that changes behavior without being controlling or policing. Every added point of friction risked making users feel controlled; every reduction risked having no impact.
The solution: make user autonomy the non-negotiable core, and let everything else be a trade-off. The next step is proving that meaningful engagement actually improves cognitive outcomes, not just satisfaction scores.
This project reinforced my belief that the most important UX work happens at the intersection of research, behavioral science, and ethics. Designing for attention isn't a usability problem. It's a question of what kind of relationship a product should have with its users.
