How we turned 451 white papers on female physiology into the most powerful algorithm turning raw data into insights and recommendations for women.
In 2019, no one had heard the sentence “cycle syncing”.
And the world thought women were men. Small ones.
But as our founding partner Dr. Stacy Sims says: “women are not small men”.
There wasn’t much research. But there was a bit.
My first hunch was that there might be a correlation between resting heart rate and the menstrual cycle.
More a question than a hunch.
And this is what I found indeed.
Very noisy correlation: the RHR being heavily impacted by sleep, alcohol and general lifestyle.
But it was there.
So I wanted to dig deeper.
Could we draw correlations?
A graph and general knowledge.
This was my first attempt to portray those thoughts.
And from there, it became a tedious, in-depth work.
With our internal team of researchers, with founding member Dr. Sahana Gopal working alongside Dr. Stacy Sims to validate which research we would rely on, depending on the journals they were published in, the reputation of the researchers and peer reviews, the quality of the research itself: how many women, how much information about their life stage was gathered, etc.
We reviewed thousands of articles to start having headlines.
We started with the menstrual cycle, which we split in 5:
- Early follicular, with periods
- Late follicular, no periods
- Ovulation
- Early luteal
- Late luteal, with pre-menstrual symptoms (“PMS”)
And then, in an Excel spreadsheet, started to put all the research that was relevant for that phase, and what it covered : RHR, knee laxity/injury risks, HRV, lung capacity, immunity, nutritional needs, energy levels, etc
This is now organized in Zotero, which we organize per life stages and physiological aspects of female health:
As an example, this is our subfolder of all the research relating to immunity and females
We then draw a baseline of how the female body and it’s underlying metrics fluctuate with the hormonal fluctuations of the menstrual cycle.
And then the needs in terms of nutrition, supplements and hydration.
And then the optimal exercises and recovery protocols depending on where she is in her cycle.
We generate recommendations based on conclusions from the different studies, or through the compilation of conclusions from different studies through our in-house scientific interpretation of the results.
We cross-check with our scientific advisor for the reliability of the recommendation.
We incorporate this recommendation alongside multiple variables including the life stage, body stats, training types of the female athlete, that would appear as the user inputs various training and hormonal data into the app.
This is what made the backbone of our first model. Turned from white paper research into data modelling by Dr. Theodosis Georgiou, Phd in AI.
We built a female readiness score taking into account the important metrics that vary in women, impact her performance, and need to be taken into account.
We drew the bell curve of the typical readiness score of women: peak at ovulation, drop at late luteal.
And that was already game changer, and never seen.
But then, we elaborated: we analyzed and criticized our own internal research.
Some women had the opposite curve: dip at ovulation, having ovulation pain and reacting strongly to the shift of hormones.
And it wasn’t an outlier: it’s another profile of women.
We found that fascinating, as one size does not fit all, and we should always be on top of our own data, and as a company, analyze and critique constantly our assumptions and learnings.
Another anecdote (not yet researched on our end!) Is that HRV would be a less good indicator of readiness for women than for men.
To continuously advance research, we also partner with universities. Concussion and menstrual cycle, strength training and menstrual cycle, etc.
And women have sex.
They often use birth control.
So we got asked : can I used Wild.AI if I use the combined pill? What about the copper IUD?
So we did the research, too.
How does each of the 148 birth controls we cover impact the hormonal fluctuations of women, what does her body need depending on those fluctuations, and how to take advantage of it.
Same for perimenopause, and menopause.
There is a huge amount of research behind the practical recommendations in our app.
We wanted to take the best, most advanced research that is made on women, and make it accessible at their fingertips.
And we have so many more life stages to cover.
Egg freezing: what does it do to the body? What can I do before, during, and after to help the process, get the best outcomes, and for my body?
Huge doses of hormones — imagine a man getting massive shots of testosterone — and we have no idea what’s going on.
IVF cycles.
Pre-partum: we see it as a race, with a date, and a preparation up to it: we want a healthy, low inflammation, low stress body. We have numbers showing that — your rhr or hrv.
Miscarriage.
Abortion.
Postpartum.
Endometriosis, PCOS, 10% of women have it. IBS.
Hysterectomy, full or partial (1/3 of women by the age of 60 in the US, or 600k new ones per year).
We are fascinated about the depth of life stages women go through, we are the experts, we know how to methodically run the research while being extremely focused.
This knowledge and research needs to be translated into actionable insights for women.
Women crave solutions.
They love reading about what they might encounter, and test if for themselves.
We always say “some research says …”, with the references.
And then they test. Yes, true for me. No, not this.
But you can’t imagine how much we crave it.
We are continuously advancing research thanks to the engrained feedback loop we have in the app, from us recommending optimal exercise, nutrition, and recovery, and women ticking it if they’ve done it or not — and then overtime seeing the evolution of symptoms: objective with wearables, and subjective by self-reporting in the daily check-ins.
So much done, so much more to do.
I am so excited about this inflexion time we’re in now and the opportunities ahead of us.
