Product Management · Idea → GTM · 2025.9 – 2025.12
04

FALL-PREVENTION
BELT

Five ideas screened down to one product. AI tools used to push the concept into a demo good enough to pitch on stage. My first end-to-end run from idea to go-to-market.

Product ManagementIdea Validation4P AnalysisAI Concept DevSenior Care
PolyU · Product Management
01 · Brief

From idea
to go-to-market.

The brief: take a blank idea through to a market-ready plan, using the methodology of a product manager. Three months, full chain.

This course doesn't reward "good PPT obedience." It rewards applying the method to a real problem: scanning the segment, converging via idea screening, then replacing expensive product photography & prototypes with AI tooling so the idea looks like it already exists when we pitch.

We landed on a fall-prevention belt — a wearable hip airbag that releases within 0.3 seconds of detected imbalance and cushions the impact on the floor. Falls are the single biggest cause of disability and lost autonomy for the 65+ population. The need is urgent; the market is growing fast.

The work split into two tracks: (1) Idea Validation & Market Analysis — five ideas scored on market growth × pain urgency; (2) AI-Enabled Concept Development — AI image & video generation for product visuals and usage scenarios, plus a complete 4P plan.

Duration
2025.9 – 2025.12 · 3 months
Idea Selected
Fall-Prevention Belt
Target Segment
65+ seniors living independently at home
Tooling
AI Image · AI Video · 4P Framework
Output
Full pitch deck + GTM plan
02 · Idea Screening

5 ideas →
1 product.

Two axes: market growth × pain urgency. The simplest and the hardest call in the whole course — overshoot and you get vanity, undershoot and you miss the real problem.

SELECTED
IDEA · 01
Smart Fall-Prevention Belt
IDEA · 02
Fire Alert for Deaf
IDEA · 03
Smart Planter
IDEA · 04
Smart Bike Lock
IDEA · 05
Screen-Free Soothing Device
03 · Why this

Three urgency signals
pushed it to the front.

Senior falls aren't a "be careful" problem. A single fall is often the moment a person stops being independent. This product isn't solving inconvenience — it's solving the loss of autonomy.

// 01 · Frequency

One in three seniors falls each year

1 / 3

WHO data: about 1/3 of people 65+ fall each year, and 50% of those who fall will fall again within twelve months.

// 02 · Severity

Falls are the #1 cause of disability in 65+

#1

One-year mortality after a hip fracture is roughly 20–30% — one of the most common ways an older adult loses the ability to live alone.

// 03 · Underserved

The market is missing active protection

~∅

Most existing products detect or alert (passive). Almost no waist-mounted product can actively protect within 0.3s. That's the real gap.

04 · Business Analysis

The "Go / No-Go"
gate.

Before touching a hardware spec, we ran the deep-dive. One question — is this product worth building? Three pillars had to clear at the same time.

// PILLAR 01

Validating Feasibility

Can the technology actually deliver the 0.3-second protection promise — reliably, at the user's hip, in everyday motion?

// PILLAR 02

Market Potential

Is the addressable market large enough to support a real product line — not a single-shot pilot?

// PILLAR 03

Profitability

Can we price this at premium and still hold a margin that funds the next iteration?

// THE GO / NO-GO GATE "This is the deep-dive stage before major investment. We answer one thing: is this product worth building?"
04.1 · Technical Feasibility

Sensor data + AI,
protected by gas.

Two stacks have to clear at the same time: predict the fall early enough, and deploy the airbag fast enough.

// 01 · AI-POWERED FALL PREDICTION & ALERT

Sensor data + deep learning models

Stack
Inertial sensors (IMU)Track movement in real time at the user's waist.
Convolutional Neural Network (CNN)Distinguishes between normal daily activities and genuine fall events.
SHAP (SHapley Additive exPlanations)Makes the AI's decisions interpretable — important for medical-adjacent claims.
// 02 · INSTANT AIRBAG PROTECTION

Miniaturized gas inflator technology

Stack
Proven MechanismMiniaturized PGG (pyrotechnic gas generator) — borrowed from automotive airbag tech.
Mechanical ReliabilityHighly dependable, near-instantaneous inflation in milliseconds — crucial for effective protection.
04.2 · Market Potential

A 20-million-person
addressable market.

China's elderly population is the largest pool we can realistically reach. We sized it conservatively, then layered on the wearable-fall-detection growth curve.

Elderly Population (65+) · China
200M+

National Bureau of Statistics, 2024.

Target High-Risk Segment
20M

Conservatively, the 10% of the 65+ pool who are high-risk and have purchasing power or family support.

// MARKET GROWTH DRIVER
6.2%

The wearable fall-detection & prevention market is projected to grow at a CAGR of 6.2% from 2025 to 2033, driven by the global aging population. (Grand View Research, 2023)

04.3 · Financial Analysis

Premium price,
protective margin.

Pricing strategy: value-based premium. The reference is not a competing belt — it's the cost of not wearing one.

Smart Fall-Prevention Belt
¥1,799
One-time purchase
Hip Fracture Surgery
¥100,000+
Often more — the cost of a single fall

Per-unit economics

Assembly & Manufacturing~¥450
R&D Amortization~¥200
Marketing & Sales~¥100
Other~¥50
Estimated total cost¥800
Gross profit per unit · margin¥1,000 · 55.6%
YearTarget Market (M)PenetrationUnits SoldRevenue (¥M)Profit (¥M)
120.000.04%8,00014.48.0
322.520.10%22,52040.522.5
525.350.19%46,90084.446.9

Year-5 penetration is still under 0.3% of the 20M target — i.e. the forecast is conservative on purpose. Growth at CAGR 6.2%.

05 · Product Development & Prototyping

Driven by real
user experience.

We pulled real reviews and interview voices from the existing fall-belt category. Six recurring complaints drove the iteration from V1 to V2.

What V1 got wrong
— in users' own words.

"It beeps when I just bend over, it's annoying — and I can't hear it clearly anyway." Comments like this told us the comfort, alert and detection layers all needed a rebuild.

01Comfort 02Volume 03Clear Guidance 04Early Warning Sensitivity 05Airbag Thickness & Shape 06Airbag Deployment Speed
// 01 · COMFORT

Lighter, breathable, hypoallergenic.

V1
~800g — noticeable burden. Generic nylon shell, stuffy and non-breathable.
V2
~480g — wearable for full-day use. Skin-friendly silicone + breathable mesh, hypoallergenic.
// 02 · VOLUME

Adjustable alert volume.

V1
Fixed volume — not adaptable to individual hearing or environment.
V2
5 adjustable volume levels via App (60–85 dB) — suits varying hearing abilities and ambient noise.
// 03 · CLEAR GUIDANCE

Dual-modal alert + context-aware AI.

V1
Single 85dB high-frequency beep — harsh, and easy to miss because of age-related hearing loss.
V2
Vibration + mid-frequency sound. Context-aware AI learns the user's habits (e.g. bending over) so it stops crying wolf.
// 04 · EARLY WARNING SENSITIVITY

Acceleration + posture angle.

V1
Single accelerometer — triggered an alert any time a rapid descent was detected.
V2
Acceleration + body posture angle → dual validation. False positives dropped below 5% — far fewer false alarms during sitting or stairs.
// 05 · AIRBAG THICKNESS & SHAPE

3 cm even-wraparound airbag.

V1
Hip-focused thickening shape — extra cushioning on the pelvis, but localized pressure when sitting and less coverage elsewhere.
V2
Tested 2 / 3 / 4 cm and chose 3 cm even-wraparound — distributes pressure uniformly, better overall protection and comfort.
// 06 · AIRBAG DEPLOYMENT SPEED

Pre-pressurized gas cylinder < 80 ms.

V1 · Solenoid
Mature, lower cost — but ~150 ms total response. Risk: airbag inflates after the user hits the ground.
V2 · Gas Cylinder
Single-use mini-canister, < 80 ms. Average imbalance-to-fall is 200–300 ms — enough time to deploy before impact.
// PRODUCT OVERVIEW & CORE FEATURES

One belt.
One complete safety loop.

SureStepSafety covers the full timeline of a fall: before (smart early-warning), during (instant airbag), after (SOS auto-alarm via the app to family & emergency), and long-term (continuous health monitoring and trend reports for caregivers).

Fall detection accuracy
99.2%
Airbag deployment
< 0.3s
Continuous monitoring
24 / 7
Weight
≈ 480 g
// IMU + Posture · dual-validation
// SOS auto-alarm via App
06 · Market Testing

Online preannounce,
offline free trial.

Two channels, two different audiences — the decider / purchaser online (adult children with budget authority) and the actual user offline (the senior who wears it).

// ONLINE · PREANNOUNCEMENT

Sure Step Safety website + social.

PlaceSure Step Safety official website + Instagram / Taobao / TikTok.
Target audienceDecider / Purchaser — the adult child who authorizes the budget and makes the actual purchase.
Site sectionsMain features · Function intro · Pricing · FAQs.
Mechanism1-month pre-launch period + discounted pre-order offer.
// OFFLINE · FREE TRIAL

Flagship shop + community trial.

PlaceSure Step Safety flagship shop + community partners.
Target customerUser — the senior who actually wears the product.
Quantity50 free-trial belts + post-trial follow-up survey.
In-storeSpecialist demos · belt function presentation · on-body wear experience.

Market response dashboard — sales monitoring.

A small group of early data and feedback that lets us project trend, sustainability and what to optimize next. Online vs. offline are tracked separately so we can decide whether to keep / scale offline stores.

Online views
2,847
+12.5%
Online sales
75
+8.3%
Online conversion
30%
+2.7%
Offline conversion
80%
+4.8%
Interest level
8.7 / 10
+3.2%
Product feedback we're tracking
  • How is our pricing strategy received?
  • How much do customers trust our technology?
  • How is our product design (comfort, appearance, durability)?
Demographic distribution
  • Who are our actual main users? (50–60: 21% · 60–70: 22% · 70–80: 23% · 80+: 14% · Family / Caregiver: 21%)
  • How do we adjust promotion to target the main users?
07 · My Role

What I owned.

A team of five. I owned the market & financial analysis on the Go / No-Go gate, and led the V1 → V2 iteration logic that turned user complaints into spec changes.

// BUSINESS ANALYSIS

  • Market sizing — built the 200M → 20M high-risk addressable funnel and the CAGR 6.2% growth case
  • Pricing logic — anchored ¥1,799 against the ¥100,000+ cost of hip-fracture surgery rather than competitor belts
  • 5-year forecast — penetration model kept under 0.3% of target market on purpose, so the plan stays defensible
  • Per-unit economics — landed total cost ¥800 → 55.6% gross margin to fund the next iteration

// V1 → V2 ITERATION + MARKET TEST

  • User-voice synthesis — extracted the six recurring complaints from real reviews and turned each into a V2 spec
  • Detection redesign — pushed the team from single-accelerometer to acceleration + posture dual-validation (false positives < 5%)
  • Two-channel test plan — separated decider (online preannouncement) from user (offline 50-belt free trial)
  • Response dashboard — designed the KPI set (online vs offline conversion, interest, demographic split) used to decide whether to scale offline
— TAKEAWAY —

What this
project taught me.

"Product manager" sounds like a collection of frameworks in class. Doing it taught me that the hardest part isn't the 4P — it's convergence. Picking one idea out of five means killing four others that all looked plausible.

Second thing: when budget is zero, AI is a leveler. We had no money for prototyping, no team for video shoots — but AI tooling pulled the visual density of the pitch up to "looks already real." I reused this exact playbook later in my internship.

The third thing — and the most important — who buys decides what the product looks like. The real buyer for this belt is the adult child, not the senior. So it has to look like an insurance product, not a medical device.