Your Whoop Can't Tell You If You're Losing Muscle or Fat on GLP-1

You started semaglutide. The weight is dropping. Your Whoop recovery score is green every morning. Garmin says your sleep is great. Everything looks perfect.

Except nobody is telling you what you’re actually losing.

Studies show that 15 to 39 percent of weight lost on GLP-1 medications is lean mass. Not fat. Muscle. The stuff that keeps your metabolism running, protects your joints, and lets you function at 70 the way you did at 50.

Your wearable has no idea this is happening. And that’s the problem.

Why your recovery score looks so good (and why that’s misleading)

When you eat less, your body does less work digesting food overnight. Heart rate drops. HRV improves. Sleep metrics trend up because your sympathetic nervous system isn’t grinding through a big dinner at 2am.

Your Whoop sees all of this and gives you a green score. Recovering great! Keep going!

But recovery from what? The score measures overnight autonomic balance. It doesn’t measure whether the weight you lost last week was 80% fat and 20% muscle, or the reverse. It can’t see that your basal metabolic rate just dropped because you lost contractile tissue. It doesn’t know that the “recovery” it’s celebrating is partly your body cannibalising itself.

This isn’t a flaw in Whoop. It’s a flaw in trusting any single data source to tell you the full story.

The data you’re already generating (that nothing connects)

Here’s what’s wild. Most GLP-1 users are already tracking more health data than the average athlete:

Wearable data: Whoop, Garmin, or Oura tracking sleep, HRV, resting heart rate, strain. Daily.

Nutrition: MyFitnessPal, Cronometer, or Lose It logging calories and macros. Most GLP-1 communities hammer protein tracking because the research is clear: 1.2 to 1.6g per kg of body weight is the minimum to slow lean mass loss.

Body composition: Many users get DEXA or InBody scans every 8 to 12 weeks to check what’s actually changing under the surface.

Blood panels: Fasting glucose, A1C, lipids, sometimes thyroid. Quarterly for most.

Training: Apple Watch or Garmin logging resistance sessions and cardio. Strava for runs.

Five data streams. Zero synthesis. Each one lives in its own app, its own dashboard, its own silo. You’re supposed to look at all of them, correlate the trends yourself, and figure out whether your current approach is preserving lean mass or burning through it.

Nobody does this well. Not because they’re lazy, but because no tool connects these dots.

The question that matters (and nobody can answer easily)

The question every GLP-1 user should be asking after month three is not “am I losing weight?” They already know the answer to that.

The real question is: given my current protein intake, training load, sleep quality, and recovery trend, am I on track to retain lean mass? Or am I losing the wrong weight?

Right now, answering that requires manually cross-referencing your nutrition logs, your wearable trends, your last body comp scan, and your training volume. Then doing it again next month and comparing.

Some people do this in spreadsheets. Most people just trust the scale and hope.

What this looks like when it goes wrong

A typical pattern: You’re four months in. Down 15kg. Feeling lighter, sleeping better, recovery scores consistently green. You assume it’s working.

Then you get a DEXA scan and find that 5kg of that 15 was muscle. Your metabolic rate has dropped accordingly. The weight loss is about to stall because your body now burns fewer calories at rest. And the “great recovery” your wearable has been reporting? Partly a symptom of having less active tissue to recover.

Nobody warned you because no single tool had the full picture.

The gap is synthesis, not more data

GLP-1 users don’t need another tracking app. They need the interpretation layer that sits across all the data they’re already collecting and answers one question: is this working the way you think it is?

That means correlating protein intake with lean mass trends over time. Flagging when recovery scores improve but training volume drops (which might mean deconditioning, not recovery). Showing the relationship between resistance training frequency and body comp changes between scans.

This is the same gap that exists for endurance athletes who track everything and still overtrain. Different audience, same problem: data exists, interpretation doesn’t come standard.

That’s what we’re building with P247.