Beyond the Numbers: Universal Hierarchical Multi-Tagging in obseed

Your Oura ring tracks your sleep. Your Garmin counts the calories — and Whoop measures your recovery, down to the percentage point. And yet you still don’t know whether you slept poorly because you had coffee at 5 PM. Whether your recovery suffered because you ignored a knee injury. Whether that run went well because the conditions were right — or because you’ve actually gotten better.

Wearables measure what happens. But why? That’s still your job. It’s a problem we kept running into internally — and one that ultimately pushed us to do something about it.

Around 28% of adults worldwide wear a wearable (DemandSage, 2025). Most of them face the same problem every morning: numbers without a story.

That’s why we built Universal Hierarchical Multi-Tagging. And today it’s live.

TL;DR: Around 28% of adults worldwide track their health with wearables — but no device explains the “why” behind the numbers. Universal Hierarchical Multi-Tagging adds structured context to your data: over 205 system tags in 11 categories, six flexible time models, and cross-device correlation — automatically, without manual analysis. (DemandSage, 2025)

What We Built

Over 337,000 health apps exist worldwide — and the vast majority operate in isolation, unaware of data from other services (IQVIA Institute, 2024). That’s the problem we wanted to solve. Not yet another app building its own data world — but a system that establishes context across all devices.

What we built isn’t simple labels. These are queryable attributes that transform your wearables from data collection tools into a real health operating system. When you log a supplement, you don’t just write “magnesium” — you specify the exact dosage (400 mg), the time (8 PM), and the form (glycinate). Then obseed automatically correlates that with your Oura deep sleep score, your Whoop recovery score, and your Withings blood pressure values. No manual analysis.

We didn’t start small: over 205 system tags across 11 categories — supplements, sauna, symptoms, mood, equipment, nutrition, medications, environmental factors, sleep quality, stress interventions, and lifestyle events. Each tag is hierarchically organized: “Caffeine” sits under “Stimulants,” which sits under “Supplements.” This makes it easy to filter, search, and analyze at any level of detail — and a single entry automatically correlates with all connected devices.

Six Time Models That Make Tagging Practical

A tagging system that demands more precision than real life allows is a tagging system you’ll abandon after three days. That was our starting point — and honestly, something we noticed ourselves with early prototypes. Too much effort, too little flexibility, and within a week the motivation was gone.

Real life is messy. You might not know exactly when you started taking magnesium. You spent three days at altitude but didn’t tag every evening. No problem.

The six models:

• Instant: Took a supplement at 7 PM. Done.
• All-day: Did an elimination diet on Tuesday. Tag the whole day.
• Interval: Spent three days at altitude. Define a start and end.
• Date range: Tracked a cold for a full week. Capture the entire arc.
• Approximate: Not sure when you started? Tag it loosely — we correlate based on date patterns.
• End + Duration: Fasted for 14 hours. Define how long it lasted and when it ended.

This flexibility means tagging adapts to your life — not the other way around.

Attributes That Actually Matter

“Sauna” as a tag alone isn’t enough — we realized that quickly. What matters: How hot was it? How long? How often per week? Only with these attributes can you answer a real question — like how a 15-minute sauna at 80 °C affects your HRV compared to 20 minutes at 90 °C.

Every tag can therefore carry structured attributes. “Symptom: Headache” includes severity (1–10) and location. “Equipment: Running shoes” includes age (miles worn) and shoe model. The data flows automatically from the tag into comparative analysis — you don’t need to manually combine anything.

We weren’t surprised when we came across a systematic review of 33 studies in the journal Translational Behavioral Medicine that confirmed exactly this: context-aware digital interventions demonstrably improve health behavior in areas like exercise, nutrition, and medication adherence (PubMed, 2021). Raw metrics alone aren’t enough. Structured context is what makes data actionable — and that’s precisely what we’re building here.

A/B Comparisons Built In

When building the tag system, we kept asking ourselves: why doesn’t this exist anywhere? 92% of smartwatch owners use health and fitness tracking as their primary function (DemandSage, 2025) — yet systematic comparisons of whether evening caffeine actually disrupts sleep are nowhere to be found in existing apps.

That’s why we’ve implemented negative/positive tag pairs throughout. “Caffeine” sits alongside “No Caffeine.” “Foam Roll” alongside “No Foam Roll.” “Cool bedroom” alongside “Warm bedroom.” You ask the system to compare days with caffeine against days without. Instantly. Cleanly. No spreadsheets, no complicated filtering logic.

Cross-Device Correlation by Default

This was the core idea from the start: tag once, correlate everywhere. When you log your magnesium supplement in obseed, we instantly correlate it with:

• Your Oura deep sleep percentage and REM latency
• Your Whoop recovery score and strain
• Your Wahoo training load and performance data
• Your Withings blood pressure and body weight trends

All in one view. No switching between apps, no manual merging.

By 2029, 740 million people worldwide will wear a wearable — up from 562 million in 2025 (DemandSage, 2025). We think that makes the question of the “why” behind the data not smaller, but larger. More devices, more data, more silos — and that’s exactly why you need a system that automatically correlates a single tag entry with all connected sources.

Who This Is For

When building the system, we kept thinking about specific people. The global biohacking market grew to $24.5 billion USD in 2024 and is projected to reach over $111 billion USD by 2034 — 16.5% annually (GlobeNewswire, 2025). Behind that number aren’t biohackers in labs — they’re people who simply want to understand why their body responds the way it does.

We’re thinking of serious athletes, who know their training matters but can’t figure out why some recovery protocols work and others don’t. Of biohackers and n=1 experimenters, who’ve been running their own protocols in spreadsheets and want a platform that actually understands what they’re doing.

We’re also thinking of people managing chronic conditions, who need to understand which interventions actually move the needle. Consider: depending on the country, between 44% (Portugal) and 81% (South Africa) of gym-goers regularly take supplements (PMC/MDPI, 2024) — but almost no one can say whether they actually make a difference. That’s exactly what we want to change.

And of health-conscious professionals, who know that sleep, recovery, and nutrition are their foundation — but can’t find a system that brings all those threads together into one coherent picture.

Why This Matters

The quantified self movement has given us remarkable access to our own bodies in a short time. What we observe: people with more data tend to ask more questions — not fewer. The sense of being able to measure almost everything doesn’t make the missing “why” easier to bear — it makes it more frustrating.

We hope Universal Hierarchical Multi-Tagging is the answer to exactly that. Not the perfect answer — but a first, real step toward a system that understands why your body responds the way it does.

Your wearables capture the numbers. obseed captures the story behind them.

Frequently Asked Questions