In VC, we review thousands of pitches each year. Sometimes, you keep hearing similar pitch over and over. Hundreds of founders go for the same business idea, and run into the same barriers when scaling. That’s why I’m writing this series: Sharing what works and what doesn’t around popular business ideas, to help founders avoid common challenges. And perhaps to jointly find solutions. You can find part 1 here.
One of the pitches we’ve seen dozens of times in the past year is in women’s health: An app or agent that understands your cycle with AI and guides you through life and health accordingly. That would be amazing, right?! Unfortunately that idea hasn’t scaled so far, so let’s discuss why.
For context: Before joining Heal Capital, I briefly worked at early-stage femtech startup theblood, where I spoke with many people dealing with menstrual cycle-related conditions like PCOS or endometriosis. These days I sit on the other side of the table with a more commercial lens. I see the challenges and opportunities from both sides.
I’m aware that it’s strange when a guy speaks about women’s health apps. But I believe sharing the VC view and sparking a debate ultimately helps us advance the field.
How it works (in theory)
The idea of tracking the menstrual cycle digitally and drawing conclusions from it has been around for decades. The first startup wave here were classic pre-AI cycle trackers. And hey, we got at least one unicorn out of it: Flo Health raised a $200 million Series C in 2024.
But let’s be precise about what Flo actually is: A consumer content and engagement machine, in a shark tank of competing apps. It’s also ironic that Flo was founded by a man (vs 100s of other female-founded apps who never made it) and they made most money by selling sensitive user data in the background.
Then, AI happened. The rise of LLMs has led many founders in the past 3 years to pitch a new wave of cycle apps. More intelligent, more actionable, more personalized. The pitch usually looks like this:
They build an app, or perhaps use a chat-based interface (Whatsapp & Co)
It captures user input like mood, pain scale, and start of menstrual cycle, etc.
It also pulls in wearable data, especially temperature and heart rate
Optionally, you import medical records and blood tests
AI algorithms (and sometimes knowledge graphs, or even custom-trained LLMs) find the perfect treatment and lifestyle advice for your cycle phase and symptoms
Perhaps you can book clinician or coaching appointments on top
Some also sell supplements or other consumables
The promised result: A MUCH better experience than conventional healthcare, online clinics, or old cycle apps.
When these ideas move from theory to practice, the following happens: First they gain impressive social media traction and get thousands on the waitlist. Then commercialization starts, and free-to-paid conversion remains low, B2B revenue stagnates, and retention sucks. In the end, it’s tough to raise VC money.
Why do so many people want this product, but it doesn’t work out?
We’re looking at the wrong layer
My core assumption: The underlying data and research isn’t there to sufficiently personalize these health journeys. The logical chain from data to treatment is broken.
Think about it - to give someone useful and personalized health advice, four layers need to add up:
The data layer: First, we need objective and measurable input. We need to capture what happens in somebody’s body and mind (symptoms, vitals, mood, ...). Currently we’re great at capturing subjective symptoms and feelings, but we’re not equipped to fully capture the biological data. Hormone levels, for example, are notoriously hard to track in their monthly fluctuations
The interpretation layer: Once we have the right data, we need to identify what’s wrong, and the pathomechanisms behind it. Unfortunately, we face a century-long research gap here. We often have no real clue how female-specific conditions work. Data isn’t worth much if we can’t interpret it!
The treatment layer: Once we know what’s wrong, we ideally match those findings with proven recommendations (lifestyle, supplements, medication, ...). The problem: We often neither have good evidence-based options, nor can we say who benefits most from them. The matching itself could probably be solved, AI is excellent at that. But matching without an evidence base underneath is just hollow
The delivery layer: The resulting guidance then needs to reach the user in a convenient, responsible, and timely way. That’s the part we’ve figured out best: LLM-based chatbots and wearables are a great surface to deliver care. Care delivery can always improve - but it’s likely not the bottleneck in this case
My point? All four layers need to be intact. You can have the most powerful AI system in your pocket - fuel it with incomplete data and missing research, and the outcome will be generic.
In my perception that’s what many women’s health startups out there struggle with. They’ve focused on innovating the surface - while the research base is 50 years behind.
Where should we look then?
My working hypothesis is that we’ll need to re-focus on layers 1-3, and each for different reasons:
The more treatment options (layer 3) we find, the more it makes sense to diagnose and funnel patients into those options. The more understanding we have (layer 2), the better we can identify treatment options. And for that, we need the right data and sensors (layer 1).
We know this playbook works in cancer, where cutting-edge research has produced more and more personalized treatment options, including for breast cancer. Pharma money drove a lot of that, since oncology treatments are expensive. Most women’s health conditions don’t have that engine - yet! I’d bet we see the same story in conditions like endometriosis once pharma options hit the market.
Beyond medication, here are some topics that I expect will materially change the market:
Any type of hormone measurement, best case continuous and easily wearable, combined with research around it! Teams like Level Zero Health, Monix, and Impli are working on exactly this. I’m not a fan of indirect methods (non-invasive wristbands and such), I simply don’t trust them to deliver medically useful, accurate results. They’re guessing instead of measuring - a topic for another post. I have similar issues with saliva and urine… although I know some people insist otherwise
Research into uterine tissue and physiology, including menstrual fluid. Such an early field, but we need to know more about what’s going on in the uterus
Any work around metabolism vs women’s health. Many diseases are inherently linked to metabolic changes, and women sometimes respond differently. PCOS (Polycystic Ovary Syndrome) for example was officially renamed to PMOS (polyendocrine metabolic ovarian syndrome) recently
Not everything mentioned above is automatically a great venture investment, unfortunately. Some of it is fundamental research. Time to discuss this.
What VCs can (or cannot) back
Many founders in this space are disappointed from VCs. Why are VCs not seeing the potential here, given the huge need and market?
Some femtech founders I’ve met tend to explain this with the predominantly male VC scene, claiming they don’t understand the severity of the problem. They also point to the biases and systemic disadvantages for female founders when raising.
There’s certainly truth in these explanations, but they’re missing the structural reasons why VCs struggle to invest. Not all layers are equally attractive, and not for the same investors:
The data layer is up and coming. Since deeptech is cool now, VCs are warming up to new sensor technology to capture more data. But they’ll only back companies that a) have world-leading scientific talent, b) are born fundraisers, and c) build actual new breakthrough tech. Throwing Claude at existing sensor data won’t cut it. But it remains challenging: These companies stack scientific, engineering, timing, and market risk all at once. Taking new medical hardware to market can cost hundreds of millions, meaning more dilution for early investors than in other models. That drives exit expectations into the multi-billions. Also, VCs have only 10 years to return a fund, which conflicts with the long research timelines
The interpretation layer is hard to fund: Doing the fundamental research to understand pathomechanisms is usually not VC territory. It belongs to academic, public or industrial research - that’s been the same for all scientific areas. VC money that requires fast scaling and commercial returns is the wrong instrument for super early and experimental science
The treatment layer has a proven model: It seems biotech investors are waking up to female health. I’ve heard of a few cool rounds recently for drug assets or platforms targeting women’s health conditions. The big advantage here: Biotech has a very established business model and go-to-market
The delivery layer is over-competitive: This is where our AI-driven women’s health apps sit. It’s the most competitive and least differentiated layer. As a result, VCs are skeptical and will only back absolute outliers in terms of founder background or traction
Btw, you could argue the deeptech comparison cuts both ways. Defense and space startups get plenty of capital right now, despite horrendous capital needs and long development timelines. Why not women's health?
The difference is that those markets are much larger money-wise than any individual female health condition, with large industry and government budgets behind them. Unfortunate, but true. Just imagine the German government funneling €100 billion into women’s health instead of defense… that would change VC dynamics. In the end, VCs just follow the money.
Bottom line? You can see my tendency: I’m very curious about the data/sensor layer, although I still struggle with dilution and timing in individual cases. And I won’t say no to founders with great execution, traction, and a smart twist on distribution.
If you’re building here, or have a take on the underlying research, DM me anytime!
Speak soon,
Lucas
This is such a sharp take. You nailed it.
Love this breakdown of the VC angle, makes total sense why B2C + hardware is a tough sell for that model. Still think this space deserves continued investment though, and yeah, the “not proven yet” trap hits basically every preventive health tech out there, not just femtech. Thanks for laying it out so clearly, def gave me a lot to chew on 👏
Hey Lucas. I agree that the science and data need to improve. I also think there’s another basic question: who will actually pay for these products, and what outcome are they paying for? Someone might download a cycle app out of curiosity, but may not keep paying every month for general advice. An insurer or employer will only pay if it clearly improves care or saves money, while pharma will only pay if it helps the right patients reach a diagnosis or treatment. That’s why I suspect the stronger businesses will solve one specific, costly problem from beginning to end, like shortening the path to an endometriosis diagnosis rather than trying to become an all-purpose AI cycle companion.