Online glasses fitting fails because webcam PD tools have 2-3mm error margins

Every online eyeglass retailer now offers a 'virtual try-on' and a webcam-based PD measurement tool. The problem is these tools have error margins of 2-3mm, which is clinically significant. For single-vision lenses, a 2mm PD error causes the optical center to shift, inducing unwanted prismatic effect. For progressive (no-line bifocal) lenses, even 1mm of error can make the reading zone misaligned, causing the wearer to tilt their head unnaturally or experience blurred vision at certain distances. This is why online retailers have return rates of 10-15% compared to 2-3% for in-store purchases. The downstream pain is significant. A person orders progressives online for $120, waits 7-14 days for delivery, tries them on, and discovers they cannot read comfortably. They initiate a return (another 7-14 days), reorder with adjusted PD (another 7-14 days), and may still get it wrong. After a month of back-and-forth, they give up and go to LensCrafters, paying $400+ for what should have cost $120. The failed online attempt wasted 3-4 weeks and created frustration that makes the consumer unlikely to try online again. This experience is common enough that it serves as word-of-mouth marketing for expensive retail optical -- 'don't buy glasses online, I tried and they were terrible.' This persists because accurate PD measurement requires either a trained technician with a pupillometer or a calibrated reference object in the photo. Webcam-based tools rely on algorithms that estimate distance using the width of a credit card held to the face or facial landmark detection, both of which introduce error. Phone-based AR tools are improving but still struggle with progressive lens fitting, which also requires segment height, vertex distance, and pantoscopic tilt -- measurements that cannot be captured from a selfie. No one has solved the fundamental physics problem of precise spatial measurement through a consumer-grade camera.