What 90 Days of Wearable Data Taught Me About Perimenopause

Three years ago, at 39, I started forgetting words.

Not big words. Normal words. The kind you've used a thousand times. I'd be mid-sentence and just... blank. The word was gone. I'd stand in front of the fridge and forget why I opened it. I'd walk into rooms with zero recollection of what I came for.

Then came the rest. Hair thinning. Not falling out dramatically, just... less of it. My nails became brittle for the first time in my life. I developed a skin allergy. Out of nowhere. And food intolerances I'd never had before. Foods I'd eaten my entire life suddenly didn't agree with me.

I blamed it all on post-COVID. Everyone was saying it. Long COVID, they said. It happens.

But my gut said no. Something deeper was off.

The hunt for answers

I went to doctors. I got blood work done. Thorough, multiple panels, everything checked. I introduced a systematic daily movement routine. Dozens of supplements. Meditation. It helped. Some of it genuinely helped.

But the symptoms kept growing.

Some were unexpected in ways I didn't anticipate. Like increased libido. Not exactly what you read about in "things falling apart after 40" articles. But there it was. While other changes made me feel bad about myself: the fog, the fatigue, the sense that my own body was becoming unfamiliar. Some were just... surprising. The whole experience didn't fit any single narrative.

And the wearables (yes, plural, because by that point I was wearing more than one) were no help.

When your data gaslights you

Here's what drove me crazy: my wearable would show a terrible restoration score and an awful night of sleep, and I'd feel perfectly fine all day. Or I'd wake up feeling like I'd been hit by a truck, check my scores, and see green across the board. 82% recovered. 7 hours of sleep. Everything "normal."

Sometimes the scores aligned with how I felt. But sometimes isn't good enough when you're trying to understand what's happening to your body.

The problem isn't the hardware. Apple Watch, Oura, WHOOP. They collect excellent data. The problem is what the software does with it. As much as I enjoy the thought of being compared to a 30-year-old marathon runner, the comparison doesn't exactly serve me. They measure duration when fragmentation matters more. They compare your HRV to a population average, not to your own baseline from two weeks ago. They don't know, or care, where you are in your cycle.

When your wearable says you're 88% recovered after a night where you woke up at 2am, 3:30am, and then stared at the ceiling until 5? That's not a score. That's a misunderstanding.

The hypothesis no one suggested

Perimenopause wasn't on my radar. Why would it be? My cycles were regular. I have PCOS, and despite that, my periods came like clockwork. Same as they always had. Perimenopause means irregular cycles, right?

Not always, as I learnt. Research has shown that a significant number of women maintain regular menstrual cycles throughout the entire perimenopause transition. The full spectrum of symptoms. Cognitive fog, sleep disruption, mood shifts, vasomotor changes. All present. Periods perfectly on schedule. A 2021 study by Coslov et al. and the 2025 AMY Study by Islam et al. both confirmed that many women experience perimenopause symptoms while their cycles remain regular.

If your only signal is cycle irregularity, you miss these women entirely. I was one of them. And honestly? It felt unfair. Like the diagnostic criteria themselves weren't built for my experience.

Once I had that second hypothesis, things started clicking. The brain fog clustering at certain points in my cycle. The sleep disruption that wasn't random but followed a hormonal rhythm. The HRV drops that aligned with my luteal phase. Patterns that no single doctor's appointment or blood panel had surfaced. Because nobody was looking at three months of daily data through that lens.

The coffee moment

Fast forward to a few months ago. I'm sitting with my morning coffee, preparing for a doctor's appointment about what had become near-depression-like states. A new addition to my growing list. I'm going through past recommendations, supplement lists, notes from previous visits.

And this is what that looked like in practice: digging through emails. Paper documents. Notes in the Notes app. Lists in Reminders. Screenshots of blood work. A supplement spreadsheet I started but never maintained. Three different wearable apps with three different interpretations of the same night's sleep.

All of this. During one of the worst brain fog episodes I'd had in months.

The irony was not lost on me.

I thought: wouldn't it be wonderful to have something that just... holds this part of my life? Something that tracks my data, stores my history, knows my context. That replaces the three wearable apps I'd grown frustrated with. That looks at my sleep, my HRV, my cycle, my supplements, my symptoms, all together, and actually tells me what it sees.

Not a symptom log. Not another tracker. Something that reads my data and understands what perimenopause does to it.

What 90 days of real tracking revealed

Once I started looking at my data this way, holistically, with hormonal context, three things became clear:

Fragmentation was the real sleep story. My total sleep averaged 7+ hours. Fine on paper. But my wake-after-sleep-onset averaged nearly an hour. On bad nights, 90 minutes. As I dug into the research, I found the SWAN Sleep Study, which used in-home polysomnography on 368 women to measure actual brain waves during sleep. It confirmed what my body had been trying to tell me: sleep fragmentation predicts next-day fatigue, cortisol disruption, and cognitive fog more reliably than total sleep duration. My wearables were measuring the wrong thing.

And I get it. The business logic is clear. Why would these companies redesign their algorithms for a subset of their user base? The math is against it. Understandable. But still unfair.

My baselines shift every two weeks. This was a genuine revelation. My HRV isn't one number. It's a wave that rises in my follicular phase and drops in my luteal phase, by as much as 30-40%. Comparing me to a population average is meaningless. The only useful comparison is me vs. me, adjusted for where I am in my cycle.

Nothing was random. Before I tracked the patterns, my symptoms felt chaotic. Brain fog on a Tuesday. Terrible sleep on a Thursday. Fatigue that came and went with no apparent logic. And most of us would just give up on trying to make sense of it at that point, right? Too much. Too scattered. Too exhausting to even think about when you're already in a fog.

But 90 days of data showed that almost all of it followed a rhythm. The fog clustered in specific cycle phases. The worst sleep correlated with elevated resting heart rate. The fatigue followed high-fragmentation nights.

Once you see the pattern, something shifts. It's not about blaming yourself less. It's about soothing the anxiety of your body behaving in new, unpredictable ways with no signs of any well-known disease. When the data shows you a rhythm where you only saw chaos, you stop wondering if you're losing it. It's physiological. It has a shape. And once you can see that shape, it becomes something you can work with.

Why I'm sharing this

There are roughly 55 million women in perimenopause across North America right now. Most are having some version of this experience: confusing symptoms, unhelpful wearable scores, scattered health records, and a medical system that often doesn't connect the dots.

The data is already being collected. Millions of women wear Apple Watches and Oura rings to bed every night. But the algorithms reading that data were not built for hormonal transitions. They score your sleep on duration. They flag your HRV without context. They don't know what perimenopause does to the numbers they're measuring.

That coffee-fueled frustration turned into something I've been building for a while now. But if you're in a similar place, confused by your symptoms, frustrated by scores that don't match how you feel, wondering if what you're experiencing is even real, I want you to know: it is real. The data proves it.

And once you can see the patterns, everything starts making a lot more sense.