Intro
Product Overview
Research & Findings
Problem Statements
Key Decisions
Design Solutions
The Outcomes
The Learnings

Employees wanted healthier food. The menu gave them no way to find it.

No nutrition transparency, no goal-driven sorting, no reason to engage with health beyond willpower alone.

I built SmartPrep — a meal ranking system on clinical nutrition logic, shipped end-to-end in 1 month.

SmartPrep — personalized meal ranking in Cubbi

Product
Overview

Responsive Web Mobile (React Native) 1 Month UX / UI Design Visual Design Design System GTM

SmartPrep is Cubbi's personalized way of browsing meals. It helps employees quickly find food that fits their health goals without changing how they normally order. Setup is optional and takes about a minute: answer 4 questions, or enter your own macros if you already track nutrition.

Using a clinically validated formula (health goal, BMR, TDEE), SmartPrep generates a personalized lunch macro target and sorts the menu by best fit. For companies it's a lightweight wellness tool. For restaurants it means a higher-value, health-motivated customer. For users the right choice just feels obvious.

Research &
Findings

User Archetypes
The Champion (Admin) The Team Manager The Employee
Role Office manager, EA, or HR coordinator managing Cubbi for their organization. Primary driver of adoption and benefit configuration. Department head or team lead who occasionally coordinates group catering. Not responsible for individual employee ordering. Individual contributor ordering their own lunch through Cubbi. High mobile fluency. Primary SmartPrep user.
Goals Ensure the meal benefit runs smoothly. Monitor team usage and adoption. Reduce support requests from employees. Place team orders efficiently. Not have to manage individual dietary preferences manually. Eat healthier without extra planning effort. Understand how today's meal fits their weekly nutrition goals. Reorder meals they loved.
Challenges No visibility into per-employee nutrition data. Hard to know whether SmartPrep is being used or is driving better choices. No visibility into what team members prefer or have ordered previously. No way to see nutrition at a team level. No nutrition data visible at ordering. No memory of what they ordered before. No cumulative tracking — each meal was evaluated in isolation.
Skills Medium-high technical comfort. Comfortable with operational platforms and dashboards. Medium technical comfort. Needs scoped, task-focused views. Does not need billing or full benefit visibility. High mobile fluency, low patience for complexity. Expects one-tap decisions and clear feedback.
Research Method
Product and Operational Context

Cubbi works with a curated vendor list with known, stable nutrition data — the enabling constraint.

Observed Ordering Patterns

Repetitive orders, no nutritional visibility at point of decision.

Competitive Analysis

Analysis of nutrition-tracking and meal planning apps for design direction.

Key Findings
FindingSourceImpact
Cubbi's curated vendor model enables precise nutrition data — unlike open marketplacesProduct contextMade SmartPrep feasible where competitors couldn't do it
Employees order the same meals repeatedly without visibility into cumulative nutritional intakeObserved behaviorJustified a nutrition tracking layer, not just per-meal labeling
Employees want to eat healthy but don't have the time or context to planProduct assumptionDrove the meal planning / goal-matching angle
User Journey Map

Archetype: The Employee  ·  Scenario: Meet nutrition goals while ordering through Cubbi

Stage
01Trigger
02Vendor Browsing
03Meal Selection
04Goal Tracking
05Reordering
User Action
Employee wants to eat healthier or meet a specific nutrition goal (calories, protein, etc.)
Employee browses available vendors for the week
Employee compares meals across vendors
Employee wants to know if today's meal fits their weekly targets
Employee reorders a meal they liked and that fit their goals
Emotion
Motivated but uncertain
Unsure
Time-consuming
Disconnected
Efficient
Pain Point (Before)
No nutrition tool existed. Employees made purely manual choices based on habit — no data, no goals, no guidance
No filtering or sorting by nutrition criteria — every choice was made on appetite, habit, or guesswork
Nutrition data was inconsistently formatted across vendors — some showed macros, some didn't, some showed per-serving vs. per-meal
No tracking layer — each meal was evaluated in isolation with no cumulative view
No memory of which meals worked — had to manually search again
Design Response
SmartPrep introduces nutrition data and goal-setting directly into the ordering flow
Personalized nutrition recommendations built into the browsing experience based on user-set goals
Cubbi standardizes and presents nutrition data in a consistent format across all vendor meals
SmartPrep tracks progress against nutrition goals across the ordering week, not just per meal
Recommendations factor in past orders and goal performance to surface relevant options faster
Emotional Journey
Cautious / Uncertain
Overwhelmed / Stressed
Neutral / In process
Relieved / Confident

Problem
Statements

Healthy eating without visibility

Employees increasingly prioritize nutrition, but the menu gave them no signal about which meals fit their goals. Every decision required external tools, manual macro lookup, or plain guessing — friction that killed the impulse at the speed lunch decisions actually happen.

﹅﹅

...I have no idea if what I'm ordering is actually healthy — I just pick what looks good...

﹅﹅

...I end up googling macros while I'm trying to pick lunch — by then I've already lost interest...

Discovery not built for health goals

Browsing was organized by vendor and category, not nutritional fit. Users with specific goals — high protein, lower carbs, weight management — had no sorting, no filtering, no path to relevant meals. The menu was the same for everyone regardless of intent.

﹅﹅

...I'm trying to eat high protein but the menu just shows everything mixed together...

﹅﹅

...there's no way to filter by what I actually want — I just scroll and hope for the best...

A data advantage going unused

Most delivery services can't enforce accurate nutritional data from restaurants. Cubbi can — direct vendor relationships meant macro data existed. It just wasn't surfaced in a way that was useful. The competitive moat was sitting idle.

﹅﹅

...I wish I could see calories right on the meal card without having to click in...

﹅﹅

...the nutritional info is buried — I'd use it every day if it was actually visible...

Companies want wellness, not programs

HR and benefits teams wanted to support employee health but had no lightweight way to do it. A full wellness program is a major undertaking. SmartPrep could be a tangible talking point — visible, zero-overhead, and employee-controlled.

﹅﹅

...we want to support healthy eating but can't commit to a full wellness program...

﹅﹅

...something tangible we could show as a health benefit would be huge for employee engagement...

Key
Decisions

Keep it optional — never force it

SmartPrep is entirely opt-in. Users browse normally by default. Personalization activates only when they choose to set it up. The design goal was to make the healthy choice feel effortless — not to lecture or gate behavior. We designed for goal-setting (targets) rather than just informational labels. A manual macro override was built in from day one for users who already track nutrition.

Clinical logic, invisible to the user

The BMR/TDEE formula is validated and accurate, but users don't interact with the math. They answer 4 intuitive questions. The system generates a personalized lunch macro target silently. Credibility lives in the outcome, not in explaining the science. We built nutrition tracking on the curated vendor constraint rather than requiring open data, and standardized nutrition data presentation across all vendor meals for consistency.

Three entry points, not one

A single entry point for a new feature gets missed. We designed three: an in-app sheet on first visit after launch, a sort option that redirects to setup if not configured, and an animated comparison chart on meal cards once macros are set. Each meets users at a different moment of intent. Personalized recommendations were built into the browsing experience based on user-set goals.

Own GTM end-to-end — B2C and B2B

I owned the full launch surface: a marketing email hero for B2C users, and a champion deck for HR and benefits buyers. SmartPrep needed to land differently for an employee ordering lunch versus the administrator deciding to keep the Cubbi subscription.

Design
Solutions

4-step questionnaire with manual override

The setup flow had to feel lightweight — completable during a first browse session, not a health app onboarding. We kept it to 4 steps: health goal, sex and age, height and weight, activity level. Advanced users who already track nutrition skip the questionnaire entirely and set macro targets directly.

One question per screen — linear, no cognitive load

Each screen has a single task. No overwhelm from seeing the whole form at once. The questionnaire was completed in under 60 seconds in all usability testing. The flow feels fast because it is.

80/20 split — auto vs. manual

Post-launch data: 80% of users completed via questionnaire, 20% set macros manually. The manual path was the right call — without it SmartPrep would have felt too simplified for the health-conscious users most likely to engage with it deeply.

Step 1 — health goal selection Step 2 — sex and age Step 3 — height and weight Step 4 — activity level
Personalized macro target + menu ranking

After setup, users see their personalized lunch macro target and a preview of how the ranking works — a before/after of the menu sorted by best fit. The comparison table is the key moment: it makes the value tangible before any order is placed.

Show the ranking logic visually

Users needed to trust the recommendation before acting on it. The animated comparison chart — rolling out on the meal card after setup — demonstrates how sorting changes, making the system feel personal rather than generic.

Single-meal target: lunch only

We scoped the recommendation to lunch, not daily intake. Lunch is the moment users interact with Cubbi. A full-day target would have required more onboarding. A single-meal target is immediately actionable with the next order.

Personalized macro target after setup Extra context — how meal ranking works
Three entry points — meet users at their moment

Features fail when users don't know they exist. SmartPrep needed presence without being intrusive — discoverable across different user states without interrupting the normal ordering flow.

(1) In-app sheet on first visit post-launch

Existing users saw a contextual bottom sheet introducing SmartPrep on their first session after release. Skippable, non-blocking, with a single clear action. One shot at awareness — designed to not feel annoying.

(2) Sort option that redirects to setup

Users who tried to sort by "Best match" without a profile were redirected into the setup flow. High-intent moment — they were already trying to use the feature. Completion rate from this entry point was highest post-launch.

(3) Animated comparison chart on meal cards

Once macros are configured, the ranking visualization appears directly on meal cards — showing how each meal fits the user's target. Passive discovery for users who missed the sort option; active reward for users who set up their profile.

In-app sheet — first visit after launch Sort option — redirects to setup if not configured Animated comparison chart on meal cards
GTM — B2C and B2B launch assets

Beyond product design, I owned the full launch surface. B2C messaging focused on personal health empowerment. B2B messaging — for HR buyers and company champions — focused on offering a lightweight, zero-overhead wellness benefit employees would actually use.

Hero image for SmartPrep B2C marketing emails
B2B champion deck — page 1 B2B champion deck — page 2

The
Outcomes

+19% more orders per month

SmartPrep users placed 19% more orders per month and had a 2× ordering rate relative to their share of monthly active users — making them a disproportionately high-value segment, not just a health-conscious niche.

+7pp weekly return rate, weeks 2–6

Once users set up their profile, they had more reason to come back — their menu was already configured for them. Personalization created a stickier product experience.

Frictionless setup won

80% of users completed setup via the 4-step questionnaire; 20% used manual macro input. The guided path was fast enough to finish during a first browse. The manual option retained the power-user segment it was designed for.

8% adoption — highest-value user segment

The feature shows clear signals of driving adoption and retention. The planned next experiment: integrate SmartPrep setup into first-time onboarding — capturing health-motivated users at their highest-intent moment rather than via post-launch discovery.

The
Learnings

Healthy features work only when
logic meets emotion.

The nutritional model matters for credibility, but it's not what drives behavior. Presentation — how meals are shown, ranked, and framed — has more impact on whether people actually pick the healthier option. The science needs to be invisible. The confidence it gives needs to be felt.

Personalization works when
it stays out of the way.

SmartPrep succeeds because it guides choices without lecturing or forcing behavior. The goal was never to teach nutrition — it was to make the healthy decision feel as effortless as the easy one. The best personalization systems are the ones users forget are running.