Independent concept, Not affiliated with or endorsed by WHOOP

Context and Discovery

Context

As a biohacking enthusiast, I've long understood jet lag as more than a travel annoyance. My research into optimisation communities revealed its true cost: compromised cognitive function in critical business decisions, razor-thin margins lost in athletic competition, and precious vacation days robbed from leisure travellers.

Every year, travellers take 1.5 billion international trips, and over 500 million people use connected wearables to manage their health. Yet these devices offer little help for jet lag. WHOOP's expertise in recovery science, continuous biometric monitoring and existing sleep-timing features position it to fill this gap. I approached this case study as a Product Manager at WHOOP tasked with delivering Jet Coach.

Problem Statement

Millions of business travellers, athletes and health conscious individuals struggle with jet lag that erodes productivity and performance.

Current solutions fail them: standalone apps lack biometric integration and work as inflexible black boxes, while wearable trackers offer superficial advice that doesn't adapt to real world constraints or individual physiology.

About WHOOP

WHOOP Today

WHOOP is a human performance and longevity platform built around continuous wearable monitoring. WHOOP tracks sleep, recovery and strain through advanced sensors. Recent launches like Healthspan, Advanced Labs and Strength Trainer signal WHOOP's evolution beyond athletic performance towards longevity. WHOOP is sold as a membership that includes the strap, in three tiers: One, Peak and Life.

Strategic Lens

WHOOP's foundation in sleep and recovery science makes circadian optimisation a natural extension. The platform already reads when and how well members sleep every night, with features like optimal sleep time and sleep consistency tracking demonstrating circadian algorithm capabilities. WHOOP also notices a time-zone change and offers jet lag tips through WHOOP AI. This positions WHOOP to address jet lag optimisation as a logical next step.

Unlike standalone apps disconnected from real-time biometrics or fitness trackers with basic recommendations, WHOOP can close the loop between guidance and physiological response.

Problem Discovery

Research Methodology

To validate the jet lag problem space and uncover opportunity gaps, I conducted secondary research across user communities, product reviews and competing apps.

  1. 1Market Context Research

    Reviewed business travel reports, sports performance studies and biohacking communities to understand which user groups experience jet lag most acutely.

  2. 2User Complaint Analysis

    Reviewed over 500 user complaints and discussions from Reddit communities, travel forums, YouTube comments and App Store reviews. Cleaning the data left 340 distinct pain points, which were analysed for recurring themes.

  3. 3Competitive Product Evaluation

    Tested multiple jet lag apps to understand their UX flows, setup friction and recommendation logic, going deep on the category leaders: Timeshifter, FlyKitt and Garmin Jet Lag Adviser. Other apps were largely derivative.

Research Constraints

This was an independent study, so it relied on publicly available data and firsthand app analysis. Informal conversations with a few frequent travellers confirmed the same patterns. In a real product scenario, structured user interviews would validate problem severity and identify edge cases not visible in public complaints.

AI scraped, cleaned and grouped the data; I reviewed it.

Research Findings

Analysis of 340 pain points revealed six distinct themes spanning the severity of jet lag and critical failures in existing solutions.

  1. Travellers experience debilitating symptoms like vomiting, dizziness, and severe brain fog that prevent normal functioning. Recovery requires 3 to 7 days minimum, with many needing weeks before feeling normal again.

    “I never felt so out of control of my brain before.”
    Traveller, Sarah Thebarge Blog
  2. Apps tell users to sleep during meal service, get bright light when all cabin curtains are closed, or wake at 3 AM before departure. Instructions ignore operational realities of air travel and cabin social norms that make execution impossible.

    “It wants me to get bright light when everyone is going to demand the curtains be closed.”
    Reddit user
  3. Apps provide time-based instructions without explaining the circadian reasoning behind them. Users don't understand why they're being asked to do counterintuitive things, creating confusion and eroding trust when advice seems impractical.

    “I don't understand the logic of the plan.”
    Google Play Store review
  4. Apps apply “one size fits all” algorithms despite substantial variation in chronotypes, age, medications, and sleep patterns. Wearable devices with biometric data exist but remain disconnected from jet lag recommendations.

    “What would make a difference is adapting to your sleep schedule using an Apple watch.”
    YouTube comment
  5. Flight delays completely invalidate plans with no recovery mechanism, forcing manual regeneration. Apps demand perfect schedule adherence despite work and family obligations, with no feedback loop for real-time adjustments.

    “My flight was delayed and cancelled...that threw off all their recommendations.”
    Rick Steves Travel Forum
  6. Flight number requirements create confusion when users just want to enter dates and times. Interface prevents scrolling backward to review missed steps or forward to plan around upcoming advice, adding friction during already-stressful travel.

    “Despite numerous attempts to finish the data entry, it refused to do so.”
    Google Play Store user

User Segmentation

While jet lag affects all travellers, pain severity varies. Three segments emerged from the research: business travellers and athletes facing acute, career defining consequences, and health optimisers, the broader wellness focused audience.

  1. Business Travellers

    Consultants, executives and entrepreneurs who fly across time zones several times a year for high stakes meetings and negotiations where cognitive impairment directly translates to career and business impact.

    “I build relationships for months for deals worth $10M. I cannot show up jet-lagged when everything is on the line. There's no second chance.”

    Core Needs
    1. 1Maximise cognitive function (clarity, memory, alertness) post-arrival
    2. 2Recover quickly after long-haul flights without losing productivity days
    3. 3Passive optimisation that fits within constrained work schedules
    Pain Points
    1. 1Jet lag causes fatigue, brain fog, and reduced mental sharpness for several days
    2. 2One-size-fits-all plans feel detached from individual sleep or stress patterns
    3. 3Rigid app schedules conflict with unpredictable work demands and travel delays
    JTBD

    When I travel internationally for business, I want to perform at my cognitive peak, so that jet lag doesn't compromise critical negotiations or client relationships.

  2. Professional Athletes

    Olympians, professional sports players and competitive athletes whose entire year is structured around international competitions, where margins are razor thin and jet lag can decide the podium.

    “My races are decided by hundredths of a second. Losing three days to a flight could cost me the season.”

    Core Needs
    1. 1Have absolute certainty that their preparation will deliver when it counts
    2. 2Trust the science behind any recommendation affecting their performance
    3. 3Gain every possible competitive advantage within legal bounds
    Pain Points
    1. 1Jet lag disrupts sleep, reaction time, and HRV recovery
    2. 2Apps ignore training loads and body feedback
    3. 3Generic recommendations don't account for performance periodisation
    JTBD

    When I compete internationally, I want my body performing at its peak, so that jet lag doesn't cost me podium placement after months of training.

  3. Health Optimisers

    Fitness enthusiasts, biohackers and wellness focused travellers who care deeply about how their body feels. They travel a few times a year for leisure, family trips or personal adventures, and don't want jet lag to steal days from their vacation or disrupt their routines.

    “I track my health daily and optimise constantly. I plan trips months ahead. I'm not wasting three vacation days to jet lag when I have the data to prevent it.”

    Core Needs
    1. 1Feel in control of their body, even when crossing time zones
    2. 2Validate optimisation efforts with quantifiable, data-driven proof
    3. 3Understand the “why” behind every recommendation
    Pain Points
    1. 1Fragmented data across devices and apps
    2. 2Generic recommendations don't account for individual variability
    3. 3No transparency or ability to validate recommendations against personal data
    JTBD

    When I travel for leisure, I want to feel my best from day one, so that jet lag doesn't waste precious time off or disrupt the wellness routine I've built.

Market Opportunity

WHOOP's pivot to longevity expands its focus beyond athletic performance. Jet lag directly undermines this vision by disrupting sleep, recovery and strain patterns during travel. Business professionals, athletes and health optimisers make up WHOOP's core base and travel frequently. Current solutions are fragmented, creating room for differentiation through integrated biometric tracking.

Market Sizing

TAM · Global Problem

1.5 BillionInternational Trips/Year

Insight: ~450 million journeys a year are long-haul, the flights most likely to cross 3+ time zones and cause significant circadian disruption.

SAM · Tech Enabled

150 MillionUsers

Insight: The intersection of the 500M+ global wearable users and international travellers who actively track biometrics.

SOM · High Intent

10–15 MillionFrequent Flyers

Insight: Wearable users taking 3+ long-haul trips/year. This segment aligns perfectly with WHOOP's premium positioning and willingness-to-pay.

Sized top-down: TAM from public travel data, SAM and SOM narrowed by assumption.

Acquisition Upside

20K–75KNew Members

WHOOP's 3M+ membersup to 2.5% growth

  • If Jet Coach is the deciding reason for 0.2–0.5% of this SOM over 5 years, that is 20K–75K new members.
  • This would represent up to 2.5% growth on WHOOP's 3M+ members. A ceiling, not a forecast: it counts only once a growth test shows Jet Coach brings people in.

Business Impact

Retention

Existing members are where the value can be proved first. Only those who cross time zones will use it, but travel is when Sleep and Recovery drop, and when the band most needs to earn its price. A plan that gets users sleeping on local time sooner gives travellers a reason to renew. None of Oura, Apple Watch or Ultrahuman offers an adaptive plan; Ultrahuman announced a jet lag feature in 2024 but never released it. Measured as renewal lift against a control group.

Acquisition (Upside)

Jet lag reaches high-LTV people who use standalone apps or nothing: business travellers and touring athletes. At $239/year, that upside would be $4.8M to $18M ARR. Needs a growth test.

Tier Lift

Jet Coach sits in Peak and Life, so One members who travel get a reason to upgrade. Small, but it supports the premium price. Needs its own test.

Enterprise

WHOOP Unite launched with 200+ organisational customers. For global teams, Jet Coach is a productivity case that supports contract renewals and expansion. Needs its own test.

On revenue alone, Jet Coach is a modest bet, and its impact is indirect and hard to prove. That's fine: a better experience for members who travel is reason enough. It's built in stages: pilot first, and scale only if travellers adapt faster and renewals hold.

Competitive Landscape

The market is polarised between standalone apps with superior algorithms but no biometric data, and hardware trackers with excellent sensors but basic or no jet lag guidance.

Standalone appsHardware trackers
AppTimeshifterFlyKittGarminOura / Apple / Ultrahuman
Biometric use
None
None
Partial (doesn't use it well)
Passive tracking only
Adaptive planning
No
No
Weak
None
Cost
$10/trip or $25/year
$99+ starter kit
Free with watch
Hardware cost
Strength
Science-backed algorithm
Calendar sync, flight data, clean UX
Integrated hardware
Mass market availability
Weakness
Rigid plan, breaks on delays
Supplement-heavy, expensive
Buggy, generic, hidden in menus
No jet lag plan(Oura only explains travel's effect on Readiness)
Jet lag UX
High friction
Streamlined
Confusing
None

The Adaptation Gap

No current competitor combines biometric personalisation, adaptive re-planning and transparent reasoning in one system. WHOOP's own AI tips are a first step, but there's no plan and nothing that adapts night to night. This leaves a clear “Blue Ocean” gap.

The opportunity

Circadian Science + Continuous Biometrics + Active Adaptation + Transparency

Prioritisation

From the research findings, I refined 34 core problems across 5 themes through consolidation (removing bugs, app-specific issues, overlapping statements). The “Problem Severity” theme is left out: it describes the problem itself, not a gap in existing apps.

RICE with Strategic Multipliers

All 34 were scored using RICE (Reach × Impact × Confidence / Effort). Standard RICE optimises for velocity, naturally prioritising quick wins over strategic moats.

Solution: Applied theme multipliers aligned with the Blue Ocean positioning:

  1. 1.8xSystem Rigidity(adaptive plans)
  2. 1.6xPersonalisation(biometric baselines)
  3. 1.2xTransparency
  4. 1.1xImpractical Advice
  5. 1.0xSetup Complexity

This prevented strategic problems from being deprioritised due to engineering effort.

Tiering: The scope is the 12 highest-scoring problems. The four that exploit gaps competitors can't fill (adaptation + biometrics) are the Strategic Bets; the other eight are Quick Wins.

Tier 0: Strategic Bets(The moat. Exploits gaps competitors can't fill.)
  1. 1No feedback mechanism for actual sleep behaviour
  2. 2No adaptation when user deviates from plan
  3. 3Plans ignore individual chronotype differences
  4. 4Rigid structure assumes perfect execution
Tier 1: Quick Wins(High impact, lower effort. Builds trust.)
  1. 1Time zone display unclear throughout interface
  2. 2Melatonin dosage and timing guidance unclear
  3. 3Recommendations lack contextual explanations
  4. 4Activity instructions lack specificity
  5. 5Sleep recommendations ignore boarding procedures
  6. 6Light recommendations ignore environmental constraints
  7. 7Pre-departure wake times conflict with work
  8. 8No guidance on meal timing for circadian alignment

Melatonin: v1 covers timing only; dose stays with the user and their doctor.MVP Scope: 12 problems (4 Strategic + 8 Quick Wins). The product solution also covers 11 of the other 22; the rest wait for a later phase.

Given WHOOP's premium positioning, this scope makes the V1 release more than an MVP: a 'Minimum Lovable Product' that does what standalone apps can't.

Design Solution

The Prototype

  • Live: flight lookup, the planning engine §2.2 and the AI Coach (GPT-6 Luna).
  • Stored: WHOOP sleep history, pulled once through WHOOP's API.
  • Simulated: nightly recalculation, progress and adherence, labelled as sample data.
  • Built with: Next.js and a Python planning engine.

Product Vision

Build a jet lag coach that adapts to how the body actually responds. It reads when and how well the user slept every night, and when reality deviates from the plan, it recalculates the protocol and explains every change.

The Core Bet: Adaptation beats perfection. A plan that adjusts to 60% adherence beats a rigid protocol that demands 100%.

Why This Approach

This case study has centred on one strategic moat: the adaptation engine. It is defensible because standalone apps have no sensors of their own, and passive trackers lack the algorithmic depth to turn data into guidance.

Everything else in this solution supports the engine. Chronotype personalisation seeds the baseline. Transparent reasoning builds trust. But the engine is the product.

Core Design Principles

Dynamic Adaptation

Plans recalculate when reality deviates (missed sleep, poor sleep, low recovery).

Zero Friction

Automate data capture from sensors, APIs and user context. The plan runs on sleep alone.

Transparent Reasoning

Every recommendation explains why using user-specific context.

1Understand the User

Before building a plan, Jet Coach needs to understand the user, using the data WHOOP already holds. That data is its unfair advantage over competitors who ask, “Are you a morning person?” and hope for an accurate answer.

1.1The Sensors

WHOOP tracks continuously using an accelerometer, a gyroscope, an optical heart rate sensor (PPG: heart rate, HRV, respiratory rate, SpO2) and a skin temperature sensor.

Three core metrics

Sleep

Efficiency, debt, stages (light, deep, REM, awake), stress and consistency.

Recovery

HRV, RHR, respiratory rate, sleep, skin temperature and blood oxygen.

Strain

Cardiovascular load from heart rate zones and activity time.

Profile DataAge, gender (captured at setup).

1.2The Chronotype

Using historical data, Jet Coach calculates the user's natural chronotype (early, balanced, late or irregular), their usual sleep window and duration, and how consistent their sleep is. This becomes the starting point, not population averages or self-reported guesses. It reads the last 14 nights, so two weekends are included. Below 7, Jet Coach asks for a few more nights before building a plan.

2Build the Plan

Using the user's baseline and trip details, Jet Coach builds a plan and estimates the chance of settling into the destination sleep schedule by the end of the plan.

Adding the trip: A flight number and date are enough, or the user can scan a boarding pass, upload a ticket or sync from email. If the flight changes, the user edits the trip and the plan rebuilds.

Derived from Flight Details
  • Time zone difference (hours)
  • Direction (eastward vs westward)
  • Flight type (daytime vs red-eye)
  • Flight duration, layovers and connections
  • Lead time (days available before departure)

Every plan is one way: It covers one journey to one destination, connecting flights included. A return flight gets its own plan, because the body clock starts somewhere different on the way back, and nobody knows yet where the outbound will leave it. One trip is active at a time.

2.1Plan Setup

Setup asks for three key decisions only the user can make.

The destination schedule: The bedtime and wake time they want after landing, pre-filled from their usual sleep in destination time. Every plan aims at it.

The plan intensity: Light, Balanced or Aggressive, each with its day count. Aggressive pushes the clock harder over fewer days; Light goes gentler over more. But the body only shifts so fast: where a short plan can't reach full adaptation in time, Jet Coach says so. It recommends the option with the best trade between days asked and adaptation reached.

The adaptation split: Start shifting before they fly, or wait until they land. Early works better, because they arrive part-way adapted, but it costs them while work and family still expect home time. That trade is theirs: Jet Coach suggests a split, and they move it.

Also at setup: Recommended tools for seeking and avoiding light, an opt-in for melatonin timing, explained before they choose, and a safety screen ruling out pregnancy, sleep disorders, and sleep or time-critical medication.

2.2The Algorithm

Light is the body clock's strongest cue, and when it lands decides which way the clock moves. Published chronobiology research maps this as the Phase Response Curve (PRC): the same light can shift the clock earlier (an advance) or later (a delay).

The turning point is Tmin, when core body temperature bottoms out, roughly 2–3 hours before natural wake. Light after it moves the clock earlier; light before it, later. That is why the plan also says when to keep it dim. Jet Coach estimates Tmin from sleep and wake timing, not a thermometer.

When to seek light, by direction of travel
DirectionRuleIn the plan
Eastward (advance)Seek light after TminLight on waking
Westward (delay)Seek light before TminLight in the evening

Tmin falls during sleep, so light goes in the waking hours either side of it.

How the plan is chosen: Rather than one plan, the engine builds three, at different paces and lengths, and tests each against a published model of the human body clock (Hannay, Booth and Forger, 2019, Journal of Biological Rhythms). It then recommends one §2.1.

How fast it moves: Pace is personal. It scales with the trip and the user's baseline: sleep timing and consistency, chronotype, HRV, recovery and strain, plus melatonin if opted in. Less data means a slower pace. This is a conservative planning pace, not a measured limit.

Honest position: The science is published; this implementation is a prototype. None of it is clinically validated, and no published rule turns baseline signals into a personal shift limit. Production thresholds would be set by an Applied Research Scientist.

2.3The Protocol

The plan times six cues. Light, sleep and melatonin move the body clock, light most of all. Caffeine, meals and workouts support the shift.

Moves the body clock
  • Seek / Avoid LightWhen to get bright light, and when to keep it dim
  • SleepWhen to go to bed and wake, moved a little each day
  • MelatoninOpt-in30 minutes before bed, or 3 hours before to shift earlier
Supports the shift
  • CaffeineA morning window, and a cut-off 6 hours before bed
  • MealA broad eating window, ending 3 hours before bed
  • WorkoutA window in the first half of the day, nothing hard near bed

In the cabin, light therapy glasses, blue-light blocking glasses or a sleep mask stand in; a cue the cabin makes impossible is skipped, and the plan rebuilds from the next night's sleep §3.1.

Each cue is a block on a day-by-day timeline, alongside flights and layovers. Tap one to see its window, why it matters, what to do, and what to do if it's missed.

3Adapt Every Night

The plan is only the start. This part keeps it alive: each night, Jet Coach reads how the user slept and rebuilds the next day from where the body clock is estimated to be.

3 numbers on the dashboard, updated every night

Body Clock Position

Where the clock is now. The plan is rebuilt from it.

§3.2

Confidence Score

The chance of adapting by the end of the plan. Set when the plan is built, re-checked every night.

§3.3

Adherence

What the user did. It never changes the plan.

§3.4

3.1Protocol Recalculation

Every night, WHOOP reads two things: when the user slept and woke, and whether that sleep went well. Together they update where the body clock is estimated to be, and tomorrow's cues move with it. The second also sets the pace.

  • Slept well

    The clock is likely following. The pace holds.

  • Slept badly

    Slow to settle, broken or cut short. The clock may be lagging, so the pace eases.

  • Signs of illness or substance use

    A bad night with another explanation, so it isn't read as a lagging clock. If recovery drops, the plan still eases (Guardrails).

WHOOP looks for signs of illness and substance use in overnight changes in RHR, HRV, respiratory rate and skin temperature. None is specific, so a flag raises doubt, not a diagnosis. If the user logged the night in the Journal, the log wins; nothing has to be logged.

The plan follows the clock, not the schedule: It never plans from where it assumed the clock would be, because an out-of-date estimate can push the clock the wrong way. Where the estimate is uncertain, timing stays conservative. Each changed block says why: “Slept later than planned, so today's light moves later to match.”

Guardrails
  • Recovery comes first: Where recovery is low, the plan eases, whatever the night showed.
  • A night with no data still costs a day: If the strap is off or flat, the deadline still moves one day closer and the plan carries on from the last estimate.

3.2Body Clock Position

The body clock is the internal 24-hour rhythm that decides when someone feels sleepy and alert. Jet Coach tracks it through one marker, Tmin, the time core body temperature bottoms out §2.2, and shows how far it has moved from home time towards destination time, against where the plan expects it to be by today. The time difference sets the distance.

The estimate comes from sleep timing. As the user's sleep moves onto the new schedule, the estimate moves with it. Sleep quality backs it up: falling asleep quickly and sleeping through at the new times suggests the clock is keeping up.

3.3Confidence Score

Confidence Score is the estimated chance the user is adapted by the date they chose. It is a forecast, not progress: Body Clock Position §3.2 shows how far the clock has moved; this shows whether it will get there in time.

Adapted means two consecutive nights meeting the timing and sleep criteria in Success Metrics.

  • At setup: set from the required shift and its direction (eastward is harder), the days available and the user's baseline sleep. It assumes realistic execution, not perfect.
  • Each night: updates after re-planning, from the estimated clock position and the days left. A night off plan only costs points if it leaves the clock further from the target.
  • Add a Day: offered when the score falls below 50%, the point where missing the date becomes likelier than making it. It offers the shortest extension that raises the score, one day or more, and previews the result: “Adding one day may raise your chance to 78% by Friday, 6 March.” The pilot tunes the 50% line. Only the user can extend the plan.

3.4Adherence

Adherence is a daily score of how closely the user followed the plan. It measures effort, not outcome.

  • 75% sleep: How closely bed and wake times matched the plan. The strap measures it, so nothing is logged.
  • 25% logged tasks: The other cues that move the clock: light sought, light avoided and, if opted in, melatonin. One tap each in the Journal.

Light moves the clock most §2.3, but the strap cannot see it. So most of the score is measured, which keeps it honest without becoming a chore. 80% is optimal, 100% is the goal. Sleep tops out at 75 and tasks at 25, so reaching 80 takes both.

Adherence never changes the plan or the Confidence Score: Both follow the body clock §3.1, which already reflects whatever the user did.

3.5Smart Nudges

A notification is sent only when it changes what the user does next.

  • Reminder

    Before a window opens. Can be logged with a tap.

    “Seek Light in 15 min”

  • Update

    When the plan changes, usually once each morning.

    “Slept late. Today's light moves to 10am.”

  • Reminders cover only the time-critical cues: light, sleep, melatonin and the caffeine cut-off.
  • Reminders are stored on the phone, so they work in flight mode and across time zones.

Principle: Guide, don't nag. Explain changes. No judgement.

3.6On the WHOOP Home Screen

A Jet Coach user still opens WHOOP every morning for the three rings: Sleep, Recovery and Strain. The plan has to live alongside that ritual, not compete with it.

One widget: It shows the body clock position, the timeline with what to do now and next, and a short AI insight. Actions can be logged right there §3.4, and tapping it opens the plan. If the widget answers the question and the user never goes further, that is the product working: the daily burden stays under two minutes (Success Metrics).

The scores stay honest: Shifting sleep on purpose often lowers Sleep and Recovery. Jet Coach does not adjust them: a bad night is still a bad night, and that is what makes WHOOP trustworthy. The AI insight adds context, “Jet Coach is shifting your sleep this week,” but never says a drop is expected. Recovery also falls for other reasons, such as illness, and calling every drop normal would teach users to ignore the number that protects them.

3.7The AI Coach

An extension of WHOOP AI that knows the user's plan. It writes the short insight on the widget and dashboard, and answers questions in chat: “Why is my light window so early?”

It explains the plan; it never changes it: The plan comes from the engine §2.2. The AI cannot move a window, add a day or give medical advice, and every number it quotes is checked against the plan, so it cannot invent one.

Validation & Evals

The engine does two kinds of work. It follows rules, which can be checked before anyone flies. And it makes estimates about a body, which can only be checked against what actually happens to people. Product sets the bar for each and decides what happens when it is missed; Applied Research owns biological validation.

The rules: Every plan must be safe, doable and push the clock the right way: no caffeine or melatonin at the wrong time, no sleep during boarding, no light on the wrong side of the estimated Tmin §2.2. Before each release, a set of east, west and multi-leg test trips runs through the engine, and Scenario Pass Rate must be 100%: one broken rule stops the release. The same checks run on every live plan.

The estimates: There are three, each built on the one before, and each bar is set before the data arrives. Where the body clock estimate is uncertain, timing stays conservative §3.1. Until the pilot checks it, the Confidence Score is a starting assumption, not a proven number.

The engine asksIf it gets it wrongHow to tell
  • Was a bad night caused by the clock, or by illness or substance use? §3.1

    It thinks the clock is behind when it isn't, or on track when it's behind

    Before launch, test it on past nights WHOOP members have already logged in the Journal

  • Where is the body clock today? §3.2

    Light meant to move the clock one way moves it the other

    Alongside the pilot, compare it with a lab test of the body clock on volunteers

  • Will this user adapt in time? §3.3

    The Confidence Score says 80% and only half make it, or says 40% and most do

    In the pilot, of everyone shown 80%, about 80% should adapt by the date they were shown

Once live: Four things are watched: plans that fail to build, plans that aren't updated by morning, plans that undo yesterday's change, and nights with no usable data.

  1. The engine is not generative AI: the same inputs always give the same plan. The coach is, so it gets its own evals.

    • Before launch: test it on real user questions and tune its instructions until every answer matches the plan and gives no medical advice.
    • After launch: set up LLM observability. Every conversation is logged, a sample is reviewed, and every mistake becomes a new test question.

These checks explain a failure; Success Metrics decide whether Jet Coach works.

Success Metrics

Jet Coach is judged on three questions, each with its own number.

  • Does it work?

    Days to Adapt (North Star)Pilot, monthly

  • Do people use it?

    Trip Activation RateWeekly

  • Does it pay?

    Renewal Rate LiftAs memberships renew

No single number answers all three. Days to Adapt is the North Star because the other two only matter if it moves.

North StarDays to Adapt

Median days from arrival until a user is sleeping normally at their destination.

Adapted means two nights in a row that pass three checks. The first night is the one counted; the second rules out exhaustion after the flight.

  • Timingthe middle of their sleep is within 60 minutes
    of the middle of the schedule they chose at setup.
  • Durationno more than an hour shorter than usual
    (or than their chosen window, if shorter).
  • Qualitysleep efficiency no lower than their usual range.

Read with it

  • Adaptation Rate: % of trips that adapt within a set number of days. The number is set before the pilot.
  • Miss Reason: for trips that don't adapt, which check failed (timing, duration or quality) and by how much.
  • Adherence: the average daily adherence score §3.4, split by sleep, light and melatonin. It shows whether people followed the plan.

Small print

  • “Usual” means their sleep in the 14 nights before the plan started.
  • A trip that ends before the user adapts isn't dropped. Flying home after three nights still not adapted counts as “more than three days”.
  • The 60 minutes and the one hour are starting values, to be tested in the pilot.
  • A wrist strap can't tell why someone slept late, and two good nights don't prove the body clock has fully moved. So the number is always compared, never read alone.
  • Left out on purpose: HRV, because it is not specific to jet lag; the engine's own body clock estimate, because it cannot grade itself; surveys, because asking adds effort and the answers could be biased.

Metrics by stage

An eligible trip is a journey that crosses three or more time zones.

  1. Adoption
    L1

    Trip Activation Rate: % of eligible trips by Peak and Life members that start a Jet Coach plan

    • L2

      Prompt Tap Rate: % of Jet Coach prompts tapped

    • L2

      Setup Completion Rate: % of started setups finished, with drop-off by screen

  2. Engagement
    L1

    Active Plan Days: % of plan days with at least one action: a task logged, the plan opened or the coach asked

    • L2

      Task Logging Rate: % of planned light and melatonin tasks logged

  3. Retention
    L1

    Renewal Rate Lift: how much more often travellers with Jet Coach renew than the control group

    • L2

      Repeat Trip Rate: % of users who use Jet Coach again on their next eligible trip

  4. Referral
    L1

    Result Share Rate: % of adapted trips whose result card is shared

  5. Revenue
    L1

    New Member Lift: how many more people join WHOOP when its marketing shows Jet Coach than when it doesn't, from a separate growth test

    • L2

      Tier Upgrade Rate: % of One members who upgrade from the locked Jet Coach entry point

Renewal and new members are indirect, because Jet Coach isn't sold on its own. Both come from tests, read last, and are never used to steer the product. Tier upgrades are counted directly. Unite counts once it has its own test.

Every rate is shown with its count, and read by direction, size of the time shift, baseline sleep consistency, and first versus repeat trip.

Measurement plan

Before the pilot: The baseline comes from WHOOP's trip history: how long members take to adapt today, against their usual sleep in destination time.

The pilot: Before full release, Jet Coach is tested with members who are invited to early access and have an eligible trip coming up. Each picks the sleep schedule they want at their destination. They are then split at random into three groups:

  • Jet Coachthe plan, re-planned every night.
  • Fixed planthe same plan with re-planning switched off.
  • Control groupWHOOP as it is today, with no plan.

All three signed up the same way, so the only difference is what they were given.

How the result reads

With Jet Coach

X days

Control group

Y days

Days saved

Y minus X

Based on

N trips

The pilot fills in the numbers. Days saved is the headline.

Ship if

  • Days saved: At least the amount set before the pilot
  • Adaptation Rate: No lower than the control group's
  • Outcome Coverage (trips with enough strap data to score): Above the set minimum
  • Scenario Pass Rate: 100% (Validation & Evals)
  • Sleep and effort: No less sleep than the control group, and under two minutes a day

Every bar is set before the pilot, so the result cannot choose its own bar.

The Core Bet: The fixed-plan group is compared with the Jet Coach group. If Jet Coach doesn't win, re-planning isn't earning its place.

After launch

  • Days to Adapt is tracked month by month, for the same kind of trip.
  • When the engine changes, half of users get the new version and half keep the old. The faster one ships, as long as users don't sleep less.

Future Roadmap

The prototype proves the concept. This is how I'd take the real product from here, with WHOOP's team.

  1. Immediate solution

    Prove it works
    • Build nightly re-planning for real. In the prototype it is simulated.
    • Validate the body clock model and light rules with circadian scientists
    • Closed pilot with invited members against a control group
  2. Near-term enhancements

    Fits real life
    • Work and personal time blocks the plan works around
    • Plans that re-anchor automatically when a flight is delayed or cancelled
    • Algorithm tuned on pilot data
    • UX refinements from real usage, like haptic cues on the strap
  3. Long-term vision

    Every trip, learned
    • Plans that learn from each user's past trips
    • Multi-stop and return trips, planned as one journey

GTM Strategy

Jet Coach shows up when the trip does. It isn't sold on its own, so it spreads inside WHOOP. It starts with the pilot, for invited members who travel often, and goes further only if the pilot clears its ship bar.

For Peak and Life members

  • Launch: A home-screen card introduces it to every memberTrip Activation Rate
  • On landing: WHOOP notices the time-zone change and offers a plan for that day, plus one for the flight homePrompt Tap Rate
  • Trip end: A trip report shows how fast they adapted, or how long it took without Jet Coach. Next time they add the flight themselves, or connect their emailRepeat Trip Rate

For everyone else

  • One members: The same launch card and landing prompt, leading to a locked entry point with the upgrade one tap awayTier Upgrade Rate
  • New members: A result card members share on the morning they adapt, WHOOP's athlete brand ambassadors, and advertisingResult Share RateNew Member Lift
  • Companies: Offered through WHOOP Unite to teams that travel for workIts own test

Afterthoughts

I'm into biohacking, and jet lag kept coming up in the communities I spend time in, with no real answer. I wear WHOOP, and the more I used it, the more I took it apart. I wanted to know how it scores, what feeds what, and why it says what it says. Jet Coach grew out of putting the two together, shaped the way WHOOP's own team might build it.

The science was overwhelming at times, and most of the work was making it simple, usually by cutting more than I added. I questioned every line until I could explain it in plain words. What you see is a fraction of it, and I enjoyed every part. I hope a wearable company, ideally WHOOP, solves this properly, and I'd love to help build it.

Here are my socials

In collaboration with Dhananjay

More projects

OTP Pulse. Status: Building.
Nomzo. Status: Building.