The Top 10 Things HairAI™ Has Found in Its First Billion+ Hair Strands

What happens when we stop guessing about hair—and start measuring it?

For more than a decade, I’ve been obsessed with one question: What if we stopped telling people what their hair needs based on what it looks like—and actually measured what the eyes can’t see?

That question became the foundation for MYAVANA and ultimately for HairAI™ and HairID™. We built our technology because I believed the hair industry had been trying to personalize an inherently individual experience using categories that were never designed to capture individuality. And as we’ve analyzed more hair, collected more data, and compared what people believe about their hair with what their strands actually tell us, the evidence has become impossible to ignore. HairAI™ is revealing where the hair industry’s assumptions break down—and why measurement is the foundation for true personalization.

Hair is far more individual than the categories we've built around it.

After analyzing data from hundreds of thousands of consumers, laboratory strand analyses, HairAI™ scans, and product catalogs, we've uncovered findings that challenge some of the most deeply held assumptions in hair care.

These are the top 10 things HairAI™ has found.

1. Curl pattern and water testing doesn't accurately predict porosity—and the relationship runs backwards from what the category assumes

For decades, the category has operated on a simple assumption:

The tighter the curl, the more porous the strand, the heavier the product it needs.

Our lab analysis says otherwise.

Among clients whose strands were measured directly, kinky-textured hair was the group most likely to test low porosity—36.4%, more than any other pattern. Wavy and straight hair skewed in the opposite direction, testing high porosity at roughly 55%.

Low porosity means a tightly sealed cuticle that resists moisture entry.

The heavy butters and sealing oils marketed almost exclusively to this group don't necessarily penetrate. They can sit on the surface, contributing to the very product buildup that the same customer is later told to clarify away.

An entire cycle of frustration can be built on an assumption nobody measured.

The lesson: Curl pattern tells us what hair looks like. It does not tell us everything about how that hair behaves.

2. We found 181 different hair profiles among a subset of 399 people

When we combine everything we measure—curl pattern, strand texture, porosity, elasticity, density, and condition—the picture stops looking like categories and starts looking like fingerprints.

Among 399 people in a random subset with complete lab strand analyses, we found:

• 181 unique HairID™ profiles

• The single most common profile represented just 3.5% of the population

• 58% of profiles appeared exactly once

Compare that with the traditional typing system, which offers twelve primary buckets.

Six measured attributes create a possible space of more than 1,600 combinations. We found 181 of those combinations in a group of just 399 people.

That is why personalization at MYAVANA has never been a marketing posture.

It's what the measurements force us to do.

If two people can have the same curl pattern but completely different porosity, strand thickness, elasticity, density, and condition, they shouldn't automatically receive the same recommendation.

Your HairID™ is more than a type. It's your hair's fingerprint.

3. Fine-stranded hair is in the majority—in every curl family, including Type 4

The word "coarse" has been attached to coily hair for so long that the product architecture built around it went largely unquestioned:

Rich. Heavy. Weight-forward. Protein-dense.

Our data doesn't support that assumption.

Nearly half of Type 4 scans—48.1%—returned fine-stranded.

Across all patterns, fine was the plurality at 44.6%, while coarse was the smallest group at 22.3%.

Our laboratory analysis independently supported the finding. Only 5.8% of kinky-haired clients graded coarse, compared with 37.6% graded fine.

This matters because fine strands and heavy formulations don't always make a great combination.

If you have tightly coiled hair that seems to go limp, stiff, brittle, or weighed down no matter what you use, the problem may never have been your hair.

It may have been the assumption behind the product recommendation.

4. Dryness is the challenge that refuses to segment

Almost everything about hair concerns divides along pattern lines.

Frizz is especially prominent among wavy and curly respondents—46.7% and 38.6%, respectively—but affects only 20.4% of kinky-haired respondents.

Breakage runs the other direction, affecting 58.1% of kinky-haired respondents compared with 34.2% of wavy respondents.

Oily hair barely registers for coily textures at 1.7%, but reaches 19.4% among straight-haired respondents.

Dryness does something completely different.

It ranked as the #1 challenge across every hair type we measured:

• 68.3% of kinky-haired respondents

• 63.1% of curly/coily respondents

• 48.3% of wavy respondents

• 45.2% of straight-haired respondents

Overall, 58.5% named dryness among their top three challenges.

And dryness rarely travels alone.

It appeared alongside:

• Frizz: 42.3%

• Breakage: 29.2%

• Weak edges: 18.7%

Dryness isn't simply one complaint among many.

It behaves like a root problem with a trail of symptoms behind it.

5. In majority of scans, our AI detected more than one hair type.

When HairAI™ classifies a scan, it doesn't simply return a label. It also tells us how confident the model is in that classification.

In 89% of scans, the model's second-choice answer falls into a completely different curl family—not simply a neighboring subtype.

When HairAI™ appears it’s hesitating, it isn't necessarily deciding between 3B and 3C.

It may be deciding between curly and coily, or wavy and straight—across boundaries that the traditional system treats as fundamental.

For a long time in training, we viewed uncertainty as something the model needed to eliminate.

Then we realized: The uncertainty was data.

Hair doesn't always fit neatly into the boxes we've created for it. When a classifier is forced to select one discrete label from a taxonomy that doesn't fully reflect the underlying continuum of hair characteristics, this is exactly the kind of signal we'd expect to see.

Then we had an aha moment. HairAI™ isn't failing when it says, "I'm not completely sure."

Sometimes it's telling us something much more valuable: The one category itself isn't precise enough.

6. What people believe about their hair and what the lab measures are two different things

Ask people to rate their own hair health and the answers cluster comfortably in the middle.

62% say their hair is "average."

16.8% say it's "great."

Only 21.2% describe it as unhealthy.

In other words, roughly four out of five people rate their hair as average or better.

Then we measure the strands.

Among our laboratory-analyzed clients:

75.4% graded "needs more care" or "challenged."

Only 4% graded "good."

Half tested low elasticity—strands that stretch and snap rather than recover.

Nearly 40% tested high porosity.

And 22.6% had both high porosity and low elasticity, a combination associated with increased vulnerability to breakage.

There is an important caveat: people who seek out a strand analysis are often doing so because something is already wrong. That means this population is not representative of the general population, and we would expect the challenge rates to be elevated.

But the gap is still significant.

Self-assessment is useful. Measurement is different.

You can know what your hair feels like.

Technology can help reveal what your hair is actually doing.

 

7. Dryness has become so normal that it stopped meaning anything

Buried in the same dataset is perhaps the sharpest finding of all.

We compared reported dryness with how people rated their own hair health.

We expected people who described their hair as unhealthy to report substantially more dryness.

They didn't.

Dryness was reported by 63% of people who called their hair unhealthy—and 62% of people who called their hair "great."

Statistically, those numbers are nearly indistinguishable.

Think about what that means.

Dryness has become so normalized that it no longer registers as a signal of poor hair health.

People have accepted it as the baseline condition of having hair.

They have priced it into their expectations.

They've stopped counting it as a problem.

But it is a problem.

It's the most common challenge we see, and it's one of the conditions consumers have been most thoroughly trained to tolerate.

That changes the conversation.

Instead of asking, "What product helps me live with dry hair?"

We should be asking:

"Why is my hair dry in the first place?"

Hint: there’s something happening on the inside.

 

8. Six in ten people applying heat are applying it unprotected

52.7% of respondents use at least one heat tool.

Of those, only 40.3% use a heat protectant.

That means roughly 60% are applying heat to bare hair.

The numbers don't improve where the thermal exposure is highest.

Among people using flat irons and curling wands—tools involving direct contact with the hair—46% still report using no heat protectant.

And here's what makes this finding particularly interesting:

The same population tells us they want healthier heat styling and damage repair.

The intention is there.

The education is there.

But one step is missing.

And it may be one of the simplest steps in the entire regimen.

Personalized hair care isn't always about adding another product.

Sometimes it's about identifying the one behavior that matters most.

9. The people using edge control are the people still reporting weak edges

Weak edges were named as a top-three challenge by 19.3% of respondents.

But the rate isn't evenly distributed.

Among edge-control users: 25% report weak edges.

Among non-users: 17.4% report weak edges.

We need to be careful about what this does—and does not—tell us.

This is a correlation, not proof of causation.

The most obvious explanation may also be the simplest: people may reach for edge control for styling purposes but not caring for the health condition.

We are not claiming that edge control causes weak edges.

But the finding raises an important question:

If the product is being used specifically to address a concern, why does the concern remain so prevalent among the people using it?

Whatever edge control is doing, the data suggests it isn't resolving the underlying condition for everyone.

And a consumer applying a product every day may reasonably believe that using it is equivalent to treating the problem.

Those are not necessarily the same thing.

10. Consumers have split into two ingredient tribes—and they barely overlap

When asked what matters most when selecting a hair product:

56.3% chose "made with natural ingredients"—and nothing else.

27.7% chose "scientifically advanced formula"—and nothing else.

Only 3.4% chose both.

That's a striking divide.

Two large populations are making purchase decisions using fundamentally different logic, yet much of the industry continues to communicate as if one message can speak equally well to both.

Then we looked at the shelf.

Across thousands of products with complete ingredient decks:

• 78.2% contain fragrance or parfum

• 26.5% contain silicones

• 62.6% contain glycerin

• 3.8% contain SLS or SLES sulfates

Glycerin is particularly interesting because its behavior can vary depending on the characteristics of the hair and the environment—bringing us right back to our first finding about porosity.

And while the industry spent years debating sulfates, fragrance has received far less attention despite its relevance to consumers concerned about sensitive scalps and related issues.

Here's the disconnect:

For the majority of consumers who use "natural ingredients" as their primary filter, four out of five products on the shelf contain fragrance or parfum.

The shelf isn't necessarily organized around what the consumer says matters most.

That's a personalization problem.

What These 10 Findings Tell Us

Ten findings.

One pattern underneath all of them.

Almost every widely held rule about textured hair was built by generalizing from a category label:

Coily therefore coarse.
Tight therefore porous.
Dry therefore in need of more oil.

Those generalizations became product architecture.

Then retail strategy.

Then marketing language.

Then advice repeated so often that it stopped sounding like a claim and started sounding like fact.

But measurement tells a different story.

None of these assumptions are reliable enough to serve as the foundation for truly healthy hair and personalized hair care.

That isn't an indictment of the people who built the industry.

They were working with the tools available to them.

The hair typing system itself was a genuine advance when it was introduced. It gave people a shared language to describe something incredibly complex.

But we've entered a different era.

We can now measure what was previously only described.

And measurement keeps returning the same verdict:

Hair is more individual than any category system can hold.

That's why the MYAVANA Unique HairID™ exists.

We aren't trying to create a better way to sort people into boxes.

We're building a better way to understand the individual.

The future of hair care isn't about finding the perfect product for your hair type.

It's about understanding your hair—its structure, behavior, condition, needs, and changes over time, including you as a person—and using that information to make better decisions and better planned routines.

Because personalization isn't putting a different label on the same recommendation.

Personalization is when the recommendation changes because the person changes.

And that's exactly where we're going next.

Know Your Hair. Prove It.

HairAI™ represents what happens when technology, science, and beauty come together around one simple idea:

You shouldn't have to guess what your hair needs.

You should be able to know.

And increasingly, you should be able to prove it.

This is only the beginning.

We're continuing to expand what we measure, what we learn, and what HairAI™ can understand—not because we want to create more categories, but because we want to make categories less necessary.

The future of beauty will not be one-size-fits-all. 

It will be data-informed, deeply personal, and built around the individual.

Know your hair. Prove it.

My hair. My journey. MYAVANA.

— Candace V. Mitchell
Founder & CEO, MYAVANA

 

---

 

Methodology

These findings draw on several MYAVANA data sources: HairAI™ classifier output, including confidence scores; complete six-attribute laboratory strand analyses conducted by MYAVANA hair analysts; consumer questionnaire responses; and our master product catalog with full ingredient decks.

Cohort sizes are noted with each finding and vary by source. Our analysis population is self-selected—these are people who sought out a hair analysis, often because something was already wrong—so challenge and damage rates may run higher here than they would in the general population.

Where sample sizes are modest, or where a relationship is correlational rather than causal, we've noted directly rather than rounding toward a cleaner story.

The purpose of this analysis isn't to prove that every traditional hair-care practice is wrong.

It is to demonstrate something more important: We now have the ability to measure more.

And when better data becomes available, better personalization should follow.

To scan your strands and start your personalized hair journey with MYAVANA, please click here: https://www.myavana.com/pages/consumer/

To work with MYAVANA on consumer intelligence needs and recommendation intelligence opportunities, please contact us here: https://www.myavana.com/pages/consumer-product-insights