Adaptive Weekly Fitness Goals: Why Garmin, Google, and the Science Moved Off Fixed Targets

The fitness industry has been quietly abandoning fixed targets for years. Garmin built Training Load into its devices almost a decade ago. Google's newly redesigned Google Health (formerly Fitbit) now centres its main metric on an adaptive weekly goal called Cardio Load. A randomised trial found that people given a fixed 10,000-step goal moved less over time than people given a personalised adaptive target. The evidence isn't mixed on this. It's converging.

This post explains why fixed targets fail, how the adaptive weekly goal model works, and where the industry's thinking has landed in 2026.

---

Why fixed fitness goals backfire, and what the research shows

Fixed goals fail for a specific, measurable reason: they can't account for the person in front of them. A 10,000-step target is the same whether you're a sedentary office worker who averages 3,000 steps or a postwoman who covers 15,000 before lunch. For the first person it's overwhelming. For the second it's trivially easy. Neither outcome helps.

A randomised controlled trial published in JMIR mHealth and uHealth tested this directly. One group got a fixed 10,000-step daily target. The other got a machine-learning system that personalised targets to each person's actual activity history. After 10 weeks, the fixed-goal group had lost 1,350 daily steps compared to baseline. The adaptive group lost only 390, nearly 1,000 fewer steps per day.^[1]

The mechanism isn't mysterious. Fixed targets above your current level feel unachievable, so people stop trying. Fixed targets below your level don't push you, so you plateau. A goal that adjusts to what you actually do keeps the bar in the right zone: hard enough to matter, close enough to reach.

The 10,000-step number, for context, was never evidence-based. It originated as a marketing slogan for a Japanese pedometer sold ahead of the 1964 Tokyo Olympics. The character for 10,000 (万) resembles a walking figure, which made it memorable.^[2] It worked as a campaign. No clinical evidence ever backed it as the right target for any individual.

---

How Garmin and Google both landed on adaptive weekly goals

Garmin arrived at this model first. Its Training Load system builds on the Banister TRIMP model. TRIMP stands for Training Impulse and measures how hard your heart works during exercise. It tracks your short-term load against your long-term average to see how your body is adapting. The key insight from Banister's work is that fitness adapts to a rolling average of what you've done, not a single fixed target. Garmin's Training Status feature uses this to show whether your current load is building fitness, maintaining it, or pushing you toward overtraining. All relative to your own recent history.^[3]

Google's Cardio Load, which lives in the redesigned Google Health app (the Fitbit rebrand went live on May 19, 2026),^[4] uses the same approach. It's a heart-rate strain score based on the same TRIMP model. The score comes from how high your heart rate goes during exercise, measured as a share of your personal heart-rate range from resting to maximum. The Target Load it generates isn't a fixed number. It's an adaptive weekly range based on about your previous four weeks of activity, recalibrating each week as you get fitter or fall behind.^[5]

Google switched Cardio Load from a daily to a weekly target in October 2025. Users had complained that rest days were triggering undertraining warnings. That's exactly what happens when an adaptive metric gets applied to a daily frame instead of a weekly one.^[6]

The weekly frame matters. Life doesn't run on 24-hour fitness cycles. You have busy days and quiet days. A bad night's sleep on Tuesday gets made up for on Thursday. Weekly goals absorb that variability in a way daily goals can't.

---

What's working in Cardio Load, and where it hits limits

Cardio Load is a well-designed approach. One of the world's largest tech companies chose it as the core metric of a redesigned fitness platform. That alone confirms something sports scientists have known for years: personalised, adaptive, weekly targets work better than fixed ones.

That said, it has four real constraints worth understanding:

• Requires a Pixel Watch or Fitbit. The metric runs on HR data, and without it you don't get a score.^[7] The May 2026 rebrand also removed Fitbit's social layer entirely: challenges, adventures, groups, community feeds, badges, and direct messaging are all gone.^[8][9] The backlash was loud: review bombing, 1,500+ upvote complaint threads, and a Google damage-control roadmap published May 27.^[10]

• No social comparison. Cardio Load is a private, solo number. There's no comparison against others, no social layer, no competition. The adaptive target tells you where you stand relative to yourself. That's valuable, but it's only half the picture for most people.

• The most useful features cost extra. The base adaptive weekly target is free. But the dynamic targets that shift between Recovery, Maintain, and Build modes sit behind Google Health Premium at $9.99 a month. That's up from Fitbit Premium's old $79.99 a year.^[11]

• Only measures heart-strain during exercise. That's useful for athletes who want to track how hard they're pushing. But it misses a beginner's 20-minute walk, a gentle yoga session, or the steps you rack up on the school run. If you just want to keep moving more each week, a heart-rate strain score isn't the right tool.

---

Why social comparison on a fixed target never worked fairly

Social fitness challenges break down when everyone competes on the same absolute number. The STEP UP trial found that a competitive social group logged 920 more daily steps per person over 24 weeks than a control group.^[12] Social motivation works. But the design of the competition matters.

If you're competing against someone who walks twice as far as you do by default, the contest isn't motivating. It's demoralising. Google removed its social features entirely rather than fix this problem. Fitbit's old challenges were fun but flawed: they compared absolute step counts, so the least active participants were structurally destined to lose every time.

The version that works is effort-based: you're competing against what you can do, not what someone else does. A 70-year-old hitting 80% of her personal weekly goal should beat a 30-year-old who hits 60% of his. That's the design that keeps everyone in the game.

For more on this, the post on choosing the best fitness app for perimenopause covers the evidence on effort-based social features in detail.

---

How Motion applies this model, and where it differs from Cardio Load

Motion's adaptive weekly goal lands on the same core idea as Garmin Training Load and Google Cardio Load. Your goal should come from your own recent history, update regularly, and run on a weekly frame rather than a daily one.

Motion uses a 12-week rolling window of your actual activity to set your personal weekly goal. The target adjusts automatically: nudging up when you've been more active, easing back when you've had a quieter period. You don't need to reconfigure anything. The goal just moves with you.

A few key differences from Cardio Load:

• No heart-rate watch needed. Motion's effort-based scoring works from any movement data: steps from your phone, data from Apple Watch, Garmin, Fitbit, Samsung, or any major tracker. A brisk walk counts. So does a gentle jog, a gym session, or an afternoon in the garden. The score is your percentage of your personal weekly target, not a heart strain number. A beginner's effort is as valid as an athlete's.^[13]

• The social layer is still there. Activity Battles let you challenge friends to week-long competitions. The scoring is effort-based: whoever hits the highest percentage of their own personal goal wins. Someone who averages 4,000 steps a week can beat someone who averages 12,000, if they push harder relative to their own starting point. That's the design that makes social motivation fair. It's also the exact social layer Google removed from its platform.

• The adaptive goal is the free product. You don't pay extra for personalisation. It's the core feature Motion offers to every user.

• It's quiet. Motion doesn't give you a Gemini-powered AI coach narrating your data. It gives you a target and a simple view of where you stand. There's a virtual pet that responds to your effort, and a competition with your friends if you want one. Common forum complaints about Cardio Load include "targets ridiculously high even though I work out every day" and "I don't understand what it wants from me."^[14] People who've switched to a simple weekly effort percentage tend to find it easier to act on.

If you're coming from Fitbit or Google Health, Motion connects to your existing device. You keep your tracker, get a friendlier goal, and get your challenges back.

---

Adaptive weekly goals: the right framework for long-term fitness progress

Adaptive weekly goals work because they solve the right problem. The goal isn't to hit an arbitrary number. It's to keep moving at a level that's doable, slowly improving, and matched to your actual life.

Garmin built this into training hardware. Google built it into a consumer platform. Both confirmed what the trial evidence showed years ago: personalised adaptive targets beat fixed ones by a clear, measurable margin.

Motion is the consumer-friendly version of that model. No HR-monitor needed, no paywall for the core personalisation, and a social layer built to motivate rather than demoralise. If you've been stuck on a fixed goal that hasn't moved in months, or you're looking for something to replace Fitbit's old challenges, it's worth a try.

Try Motion free →