The Lab Has a Diversity Problem. Your Devices Are Paying for It.

Alright, let's talk about who's not in the lab.

I spent four years in hardware QA. Most of those years, the team was overwhelmingly male, overwhelmingly similar in build, and overwhelmingly convinced that their own hands, faces, and wrists were a representative sample of the human population. (They weren't.) The prototypes that came through our lab got stress-tested by people who looked a lot like the people who designed them. And when something felt off to someone who didn't fit the default mold — too heavy, too wide, strap too stiff — the note usually got filed under "edge case."

That's not a culture piece. That's a calibration error. And I've been watching it ship in $1,200 flagships for years.

International Women's Day is March 8. I'm writing this March 5. I'm not going to do a listicle of inspiring women who "smashed glass ceilings." That's someone else's format. What I'm going to do is show you the hardware evidence — the actual, measurable cost of running homogeneous design labs — because the myth that this is a soft HR problem is killing product quality and eating your money.

The IPD Problem Is Not a Coincidence

Let's start with VR, because it's the most documented case.

Interpupillary distance (IPD) — the gap between your pupils — matters enormously for VR headset optics. Get it wrong and users experience eye strain, headaches, and nausea within minutes. The average male IPD is roughly 64mm. The average female IPD is closer to 61mm — a 3mm gap that sounds trivial until you're trying to hold a focal plane steady.

Meta's Quest 2 shipped with IPD adjustment at three fixed positions: 58mm, 63mm, and 68mm. The Quest 3 improved this to continuous adjustment between 53–75mm, which is actually good. But here's the thing: Meta shipped nearly 20 million Quest-line headsets. For the years Quest 2 dominated the market, its optical sweet spots were tuned around three data points — not a distribution. The 63mm default worked great for the median male user and left a significant chunk of everyone else with a headset that caused eyestrain after 20 minutes. Multiple published VR comfort studies have tracked this demographic gap in reported discomfort rates.

Oculus Rift S is a starker example. It shipped in 2019 with a completely fixed IPD of 63.5mm — no adjustment mechanism at all. That's a single number, baked into the lens geometry, representing the default male mean and leaving anyone outside that narrow band with a product that simply doesn't work correctly for their face.

Sony's PSVR2, by contrast, launched in February 2023 with a physical IPD dial adjusting from approximately 58–71mm. That's the right call — and a meaningful improvement over the original PSVR's fixed-lens approach. The adjustment exists because lens geometry affects every user differently. A dial costs almost nothing at production scale. The original PSVR didn't have one.

Valve's Index has dial-adjustable IPD from 58–70mm. The pattern across these headsets is consistent: the teams that treated IPD range as an engineering requirement rather than an edge case shipped products that work for more faces.

Phone Ergonomics: The One-Hand Assumption

Here's a number worth knowing: the average male hand span is about 189mm. The average female hand span is about 172mm — roughly 17mm narrower.

The original iPhone was 115mm tall and 61mm wide. It was designed to be used one-handed. That was a core design principle, explicitly stated by Jony Ive's team.

The iPhone 16 Pro Max is 163mm tall and 77.6mm wide. It is not designed to be used one-handed by anyone. But the "Pro" marketing language still implicitly frames size as performance. Bigger screen = more serious user = this is the Pro device.

Nobody made a deliberate choice to exclude smaller hands. What happened is more interesting and more insidious: the default user in the room got bigger phones, found them comfortable, and the ergonomic feedback loop closed on a set of hands that represented roughly half the population.

The Galaxy S25 Ultra is 162.8mm x 77.6mm. The Pixel 9 Pro XL is 162.8mm x 76.6mm. The largest flagship tier from every major Android manufacturer has converged on dimensions that require two hands to operate and called it aspirational.

Apple's iPhone 16 is 147.6mm tall — and actually fits in most hands. It's frequently described in reviews as the "base model" option, carrying an implicit downgrade framing. It is not the downgrade. It is correctly sized for the median hand. The fact that 163mm feels normal to reviewers is what a decade of male-dominated ergonomic feedback gives you.

Wearables: The Band Isn't the Problem. The Default Is.

Smartwatch shipments ran into the hundreds of millions in 2025, and market research consistently shows women make up a substantial share of buyers — roughly four in ten by most estimates. Yet most flagship wearables shipped for years with sensor-facing default band fits and case sizes calibrated toward the upper end of the female wrist range.

This matters because band fit affects sensor accuracy. Optical heart rate sensors, SpO2 readers, and the EDA sensors on Fitbit's stress-tracking devices all require consistent skin contact. A band that ships too loose for your wrist doesn't just slide around. It gives you worse health data.

Apple introduced 40mm watch cases alongside the larger 44mm and 45mm options — a direct acknowledgment that the default case size was excluding a significant user segment. These smaller variants exist because someone measured the market and realized the ergonomic assumptions were costing sales. (That's the depressing version of "inclusion" — it happened when the revenue gap became visible.)

The Galaxy Ring launched in July 2024 with sizes 5–13 — a genuinely useful range that Samsung subsequently expanded through size 15 in 2025. The Oura Ring (gen 4) covers sizes 4–15. Those are both products that actually addressed the measurement problem from the start. Some hardware teams are getting this right. But it took the wearables category roughly a decade and multiple product generations to get there.

Who's Actually Missing and Why It Matters for Engineering

Here's where I'm going to be specific, because vague gestures at "diversity" are useless.

According to the U.S. Bureau of Labor Statistics (Current Population Survey, Table 11, latest published data), women represent:

• Approximately 27.5% of computer and mathematical occupations overall
• About 10.1% of electrical and electronics engineers
• About 22.5% of computer hardware engineers specifically

Read those numbers carefully. The electrical engineers — the people designing circuits, board layouts, and sensor packages — are 90% male. That number is the one that matters for this conversation. The people deciding on button placement, chassis dimensions, band hardware, and lens adjustment ranges are overwhelmingly male. The people running QA on those decisions were, at least in my experience, also overwhelmingly male.

This produces measurable errors. Not because men can't empathize with different users — but because user testing on your own team is a terrible substitute for diverse design representation. If the person who designed the grip texture of your controller has never struggled to reach the face buttons on a wide gamepad, that feedback lives outside the team entirely.

Fairphone has received 10/10 repairability scores from iFixit across multiple product generations — a distinction no other major smartphone manufacturer has matched. Fairphone explicitly includes non-engineer users in design reviews and has built a gender-balanced workforce as part of its operating philosophy. Their modular approach, which drives those repairability scores, is downstream of a user-centric design process. Correlation, not causation — but it's consistent correlation.

The Bright Spots (Real Ones, Not PR)

I'm not interested in tokenism. Here are hardware outcomes that hold up under scrutiny.

Apple's Accessibility Shortcut architecture is now standard across iOS. This wasn't a charity project. It was a recognition that rigid one-size-fits-all hardware interaction was leaving users behind. The result is features that work for everyone — and that Apple has consistently extended and improved across hardware generations.

Microsoft's Adaptive Controller (Xbox) was built with users with disabilities in the design loop from prototype stage. The result is hardware that's modular, documented, and actually serviceable — iFixit gave it an 8/10, compared to 1/10 for the standard Xbox One controller. That gap reflects a design philosophy difference, not a coincidence.

The Quest 3's continuous IPD adjustment — compared to the Quest 2's three-position step system — reflects user feedback that accumulated over the Quest 2's product life. When users with faces outside the three default positions consistently reported discomfort, the spec for the next generation changed. That's how engineering feedback loops are supposed to work.

What Actually Needs to Happen

I'm not going to tell companies to "do better" because that's not actionable. Here's what actionable looks like.

Ergonomic testing standards need explicit variance requirements. If you're shipping a wearable to hundreds of millions of users, your QA testing should include hand, wrist, and face measurements across the full distribution — not just the median. This is an engineering standard problem, not just a hiring problem. Though it's also a hiring problem.

Default user profiles in internal testing need to expand. Automotive safety researchers have documented for decades that crash testing historically relied on male-proxy dummies as the primary standard, and that female occupants show different injury patterns in equivalent crashes. Regulators and academic researchers have pushed for better representation, but the process remains incomplete. Consumer electronics has even less external pressure. That's a gap with consequences.

Hardware teams need to track this. Companies don't improve what they don't measure. If no one is counting how many women are in hardware engineering roles versus UI/UX roles versus software roles, the gap stays invisible — and the calibration errors stay in the product.

The Verdict for Your Wallet

The hardware industry's diversity gap is not an abstraction. It shows up in headsets that give you migraines, bands that slip during workouts, and phones you need two hands to operate. These are quality problems with quality solutions: measure more users, hire from a wider pool, and build feedback loops that include people who don't look like the team.

This is not a soft topic. It is a precision problem. And precision is the only thing that should matter when we're talking about products that cost over a thousand dollars and collect your biometric data.