Targ

Adaptive Workout Decision Support App

Role

UX Researcher · Interaction Designer · Prototyper

Industry

Health & Fitness · Human-Computer Interaction · Behavioral Design

Duration

10 Weeks

Designing for Real-World Constraint, Not Ideal Conditions

Most workout systems assume ideal conditions. Equipment is available. Plans are linear. Users behave predictably.

Field research revealed otherwise.

Workouts are shaped by interruption. Equipment becomes unavailable. Users hesitate. Social discomfort influences decisions. Plans are reordered constantly.

This project investigates a deeper systems question:

How might we design a workout architecture that adapts to environmental uncertainty while preserving confidence and flow?

Rather than optimizing tracking, I focused on decision-making under constraint.

From Field Observation to Behavioral Insight

I conducted structured field observations, semi-structured interviews, and a follow-up quantitative survey to understand how users adapt workouts in shared gym environments.

Three consistent behavioral patterns emerged:

Users rarely wait; they substitute immediately
Hesitation spikes when equipment is occupied
Social discomfort significantly influences decisions

To validate these findings at scale, the survey measured:

Frequency of workout reordering
Substitution versus waiting behavior
Emotional response to crowding
Comfort level initiating interaction

Results confirmed that adaptive behavior is not an edge case; it is the norm.

This reframed the problem from exercise planning to real-time decision architecture.

Designing for Social Confidence Under Pressure

The Conversation Scripts interface emerged directly from this tension.

Instead of assuming social ease, the system provides structured micro-support during high-friction moments. Ready-to-use phrasing reduces improvisation stress while preserving autonomy.

The Feeling Stuck state addresses another insight: hesitation is often silent. Users pause, scan, and lose momentum.

If inactivity is detected, the system surfaces contextual actions:

Continue waiting
Switch to a high-match substitute
Skip and return later
Ask how long remains

The goal is not interruption. It is guided continuity.

Replacing Linear Plans with Adaptive Systems

Field data showed that fixed workout order rarely survives real-world conditions.

The Plan, Equipment, and Smart Substitutes flows introduce structural flexibility:

Exercises are clustered rather than rigidly sequenced
Substitutions are ranked by muscle overlap and movement intent
Environmental constraints are integrated into the decision layer

The Smart Substitute Engine calculates similarity using muscle targeting and movement intent, enabling meaningful adaptation rather than arbitrary alternatives.

The Equipment screen embeds this logic directly into Gym Mode, minimizing navigation during active sessions.

The workout plan transforms from a static checklist into an adaptive framework.

Research-Led Design Principles

Every design decision traces directly to observed behavior.

Minimize cognitive load during physical exertion
Surface contextual decisions only when necessary
Support emotional confidence without a patronizing tone
Prioritize clarity and hierarchy over decorative complexity

The visual system reinforces decisiveness through contrast, typographic hierarchy, and explicit state communication.

Interaction design reduces micro-friction rather than adding features.

Quantifying Behavior to Inform System Architecture

The survey phase extended qualitative insights into measurable patterns.

Findings showed:

72% of participants reordered at least once per session
Substitution occurred 3× more frequently than waiting
68% reported discomfort asking to work in
Confidence strongly correlates with training experience

These findings validated the need for:

Adaptive exercise clustering
Ranked substitution logic
Embedded social support layers

The system is grounded in both observed behavior and quantified prevalence.

Designing Within Real-World Constraints

This system was intentionally designed within practical limitations.

No hardware dependency
The solution does not rely on gym-installed sensors or occupancy infrastructure. Adaptation logic is user-driven, ensuring scalability.

Limited real-time availability data
Equipment availability is inferred through interaction patterns rather than live feeds, balancing feasibility and simplicity.

Privacy boundaries
The system avoids intrusive tracking. Hesitation detection is based on session inactivity, not biometric or camera-based monitoring.

Cognitive load during exertion
Users are physically engaged and often fatigued. Interface complexity was intentionally constrained to preserve focus.

Behavioral variability
Adaptation patterns vary across experience levels. The system avoids over-automation while still providing structured support.

Toward Context-Aware Workout Architecture

This project reframes fitness design as contextual systems design.

Instead of asking how to track workouts more efficiently, it asks:

How can a product reduce decision friction in unpredictable environments?

The prototype demonstrates how adaptive logic, emotional awareness, and environmental context can coexist within a focused, minimal interface.

Future directions include:

Longitudinal personalization based on adaptation patterns
Enhanced contextual modeling of crowd impact
Refined decision-support thresholds

At its core, this work explores how research shapes system architecture, not just features.

A version of the concept prototype can be accessed here:

Targ_Prototype