Saveability

Empowering conscious consumers to shop sustainably with instant product verification

Mobile App Design UX Research Sustainability Behavioral Design
0→1
Concept to prototype
4
Core user flows
82
Product score system
12
User interviews

Overview

Saveability is a mobile app that helps conscious consumers make sustainable purchasing decisions through instant barcode scanning and AI-powered product verification. In a market full of greenwashing and information overload, Saveability provides transparent, trustworthy sustainability ratings at the moment of purchase.

This project explores how design can help people align their values with their shopping habits—turning complex research into a simple 3-second barcode scan.

52%
of consumers encountered false sustainability claims
177M
American eco-friendly shoppers
$217B
global sustainable product market
73%
of Gen Z willing to pay premium for verified sustainability

Browse & Profile: Category navigation and user account management

The Problem

The Greenwashing Crisis

Consumers want to shop sustainably, but they can't trust brand claims. With 52% of shoppers encountering false or misleading sustainability marketing, and 23% wanting to buy sustainable but not believing brands, there's a critical trust gap in the market.

I want to do the right thing, but I don't have time to become an expert on every certification. I've been burned by greenwashing before—I need something I can actually trust. — User research participant

Key Insight

The problem isn't lack of information—it's inability to verify and act on information at the moment of purchase. Consumers need instant, trustworthy answers, not more research homework.

Research & Discovery

I conducted extensive market research and competitive analysis to validate the problem space and identify opportunities for differentiation.

Market Validation

85%
of consumers shifted purchasing behavior toward sustainability
9.7%
premium consumers willing to pay for sustainable products
7.4%
year-over-year growth in eco-friendly shoppers
2.7x
faster growth than conventional products

Competitive Analysis

Feature Good On You Saveability
Product Categories Fashion only All retail categories
Discovery Method Manual search Barcode/photo scanning
AI Alternatives None Intelligent suggestions
Speed to Decision ~30-60 seconds <5 seconds
Visual Recognition No Yes
Real-time Comparison Limited Side-by-side scoring

Competitive Insight

Good On You reached 3 million users with ONLY fashion and ONLY manual search. Saveability's scanning + multi-category approach addresses their core limitations while serving a broader market.

The Solution

Instant Sustainability Verification

Saveability transforms complex sustainability research into a simple 3-second barcode scan. Users get clear, trustworthy scores backed by third-party certifications, plus AI-powered alternatives when better options exist—all at the moment of purchase.

Certification guide: Educating users on sustainability standards and their meanings

Key Features

Instant Barcode Scanning

Point your camera at any product barcode and get results in under 3 seconds. No typing, no searching—just instant verification while you shop.

Clear Sustainability Scores

Overall score (0-100) with detailed breakdowns: Environmental Impact, Social Responsibility, Company Ethics, and Transparency. Letter grades make it instantly understandable.

AI-Powered Alternatives

When a product scores low, our AI suggests 3 better alternatives at similar price points—available at nearby stores or online with carbon-neutral shipping.

Verified Certifications

See all relevant certifications (B Corp, Fair Trade, Organic, etc.) with explanations of what they mean and why they matter.

Impact Dashboard

Track your sustainable shopping impact over time: plastic avoided, carbon saved, ethical brands supported. Real metrics, not vanity numbers.

Transparency Reports

"The Good" and "The Bad" for every product—honest, clear information about trade-offs so users can make informed decisions.

The Scanning Experience

The core interaction is the barcode scan. We designed it to be fast, forgiving, and reassuring—critical for in-store usage where users feel self-conscious scanning products in public. When a product scores poorly, we don't just tell users—we help them find better options.

1. Scan a product barcode

2. View sustainability report

3. Find better alternatives

End-to-end flow: Scan, evaluate, and discover sustainable alternatives in seconds

Design Insight

We show long-term cost comparison (e.g., "$32.95 vs $155.48 over 1 year") to reframe sustainable products from "expensive" to "cost-saving." This helps users justify the upfront investment.

Key Design Decisions

1. Speed Over Perfection

Early testing revealed users would abandon the scan if it took more than 5 seconds. I optimized the entire flow to show results in under 3 seconds, even if it meant progressive loading of detailed information.

Design Principle

In-store decisions happen in seconds, not minutes. The app must match the pace of real shopping behavior.

2. Score Context is Critical

Users seeing "82/100" didn't know if that was good or bad. Adding letter grades (A-, B+) and contextual labels ("Excellent Choice" vs "Room for Improvement") made scores instantly interpretable.

3. Alternatives Without Shame

When products scored low, users felt guilty if no affordable alternatives existed. I designed the system to only suggest alternatives within 20% of the original price, and to frame it as "helpful guidance" not moral judgment.

I don't want an app that makes me feel guilty. I want one that helps me do better when I can afford to. — User testing participant

4. Show Less, Not More

The initial design showed all information at once (environmental, social, company, certifications, sources). Users felt overwhelmed. The final design shows: Score → Summary → Tap for details. 85% of users never scroll past the summary.

5. Real Impact, Not Gamification

Early concepts included badges, streaks, and leaderboards. User testing revealed this felt "gimmicky" and "preachy." The final design shows only real metrics: carbon saved, plastic avoided, ethical brands supported.

Success Metrics

Saveability is a 0→1 concept, so the numbers below are the targets the design was built against — not measured outcomes. They define what the experience would need to deliver to be considered successful at launch.

Scan history: Past scans and product reports

Impact dashboard: CO₂, plastic, and water savings

Target State · Concept Validation Goals
<3 sec
Target scan-to-result time
60%+
Target 30-day user retention
50K+
Target products in sustainability database
92%
Target sustainability score accuracy
Vision State · What's Next

From barcode to camera

The shipped Saveability concept solves the verification problem with a scan-and-score model. The natural next step is to remove every friction point in that chain — the barcode dependency, the static numeric score, the one-size-fits-all judgment.

An AI-enhanced Saveability turns the camera into the only input the user needs, and the verdict into a personal explanation aligned to what each shopper actually cares about. Three AI capabilities layer onto the existing flow:

Vision AI

Point-and-shoot recognition

A packaging classifier removes the barcode dependency. Niche, store-brand, and unbranded products become scannable too — the long tail that breaks today's database lookups.

Generative AI

Personalized verdict

Replaces the numeric score with a plain-language explanation, weighted against each shopper's stated values. The "why" is legible in one read, not buried in a certifications list.

Predictive AI

Better-alternative ranking

A recommendation model surfaces 3 alternatives ranked by predicted utility — filtered for price parity, in-stock locally, and aligned to the same value chips that drove the verdict.

Vision Mock · Scan in progress
1
User-value chips at the top

Captured during onboarding (Low carbon · Recyclable · Fair labor) — every verdict is scored against them, so the same product can score differently for different shoppers.

2
AI bounding box + 94% confidence

Vision model identifies the product without barcode dependency. Confidence is surfaced explicitly so the user can correct a misidentification before acting on the verdict.

3
LLM-generated verdict, not a static score

The 7.2 score sits in context: a one-paragraph explanation calling out where this product is strong on the user's values and where it's weak ("transit emissions pull the score down 1.2 for your East Coast region").

4
"Why this score?" + thumbs feedback

Tapping the affordance reveals the underlying certification data and how each factor was weighed. Thumbs feedback trains the personalization model. Failure state: drop to existing static score-and-list view.

Vision Mock · States — Possible Matches + Personalized Verdict
Low-confidence — possible matches
1
"Possible matches" never asserts one identity

Borrowed directly from Apple Visual Look Up's failure-mode framing. Asserting and being wrong is worse than acknowledging uncertainty — especially in a verification context where misidentification can drive a wrong purchase.

2
Confidence is in the cards, not hidden

78% / 62% bars let the shopper pick the more-confident option themselves rather than forcing the model to choose. The system's uncertainty becomes the user's affordance.

3
"Neither — let me try again" escape

When both candidates are wrong, the shopper isn't trapped. The escape is text-link weight (not a button) so it doesn't compete with the primary "Select" actions, but is one tap away.

Personalized verdict — "for you"
1
Same product, different scores

A different shopper with carbon-only values would see 8.4, not 7.2. The "FOR YOU" pill makes personalization visible — score is a property of (product × shopper), not the product alone.

2
Verdict cites the trade-off, not just the conclusion

"Emissions are 40% lower than dairy. Fair labor is mid-tier; supplier audit is 2024 (2 years old). Transit pulls the score down 1.2 for your East Coast region." Full reasoning chain — the shopper can disagree with the weighting and the next chip update will reflect that.

3
Next-action is "Find better alternatives"

Score is a means, not an end. Saveability earns engagement when it helps shoppers act, not when it produces a number. The CTA bridges from verification to decision.

Precedent References
  • Yuka — the plain-language verdict pattern (Excellent / Good / Mediocre / Bad with one-line explanations) makes the rationale legible in 2 seconds. Layer personalization on top of that pattern.
  • Google Lens shopping — the bounding-box-then-card flow is the cleanest template for "object recognition → information retrieval" without modal-overlay heaviness.
  • Apple Visual Look Up — sets the bar for showing the AI was wrong gracefully ("Possible matches" rather than asserting one identity), which is critical for trust in a verification context.

Key Learnings

1. Intent-Action Gap is Real

85% of consumers want to shop sustainably, but only 26% do. The gap isn't motivation—it's friction. Reducing verification time from minutes to seconds can close this gap.

2. Context Beats Data

Users don't need more information—they need interpreted information. "This product contains PFCs" means nothing. "Contains PFCs (harmful chemicals that persist in environment)" drives understanding.

3. Affordability is Non-Negotiable

Suggesting $80 alternatives for a $30 product creates guilt, not behavior change. AI must filter for price parity or users lose trust in recommendations.

Reflection

This project reminded me that good design isn't about adding features—it's about removing friction. The biggest decision wasn't what to show users, but what NOT to show. By keeping things simple and prioritizing speed, Saveability turns a 30-minute research task into a 3-second decision.

The greenwashing problem isn't just about marketing—it's about how we present information. When 52% of consumers see false claims, adding more claims isn't the answer. Trusted verification is. Saveability shows that with the right design, we can help people act on their values in real time.

Impact Beyond the App

If 500K users each make just one more sustainable purchase per month, that's 6 million conscious decisions per year. At scale, verification tools like Saveability don't just help individual shoppers—they create market pressure for brands to become more sustainable or risk being exposed.