3D for e-commerce5 min read

A short history of AR in e-commerce, with dates you can check

From ARKit and IKEA Place in 2017 to Quick Look, Scene Viewer and headsets. What actually happened in AR shopping, with a source for every date.

By CharpstAR · 4 Oct 2023 · Updated 22 Sept 2026

OCT2023

A short history of AR in e-commerce, with dates you can check

Augmented reality in shopping has been announced as the next big thing roughly once a year since 2017. Some of those announcements turned into infrastructure that millions of product pages now sit on. Others turned into nothing. It is worth separating the two, because the pattern tells you where to put your money today.

Everything below has a date and a link. Where we do not have a source, we have left the claim out.

2017: the phones learn where the floor is

The important year. Apple shipped iOS 11 on 19 September 2017, and with it ARKit, which Apple described as bringing augmented reality to hundreds of millions of iOS devices. Before that, putting a virtual object into a camera view at the right size required a printed marker or a specialist app. After it, the phone could work out where the floor was on its own.

Retail moved fast. IKEA announced IKEA Place on 12 September 2017, built on ARKit and requiring iOS 11, with more than 2,000 products at launch and a stated scaling precision of 98 percent. The launch range was deliberately chosen: sofas, armchairs, footstools, coffee tables and top-selling storage. In other words, the things where the shopper's question is whether it will fit.

That choice of range is the most durable lesson in this whole history. Nine years later, the categories where AR earns its keep are still, overwhelmingly, the ones IKEA picked first.

2018: Android catches up

Google announced ARCore 1.0 on 23 February 2018, with support across 13 phone models and a stated reach of 100 million Android devices. Developers could publish AR apps to the Play Store from that day.

The gap between September 2017 and February 2018 explains a lot of the awkwardness of early retail AR. For those months, an AR feature was an iPhone feature, so brands built apps rather than web experiences, and apps require a download that almost no shopper will do to look at one sofa.

2018 to 2020: AR moves out of apps and onto the web

This is the shift that mattered commercially and got the least attention. Two pieces of platform plumbing made it possible.

Apple's AR Quick Look lets Safari, Messages, Mail and Notes open a USDZ file directly in AR on iPhone, iPad and now Apple Vision Pro, with support from iOS 12 onwards. No app. Apple also added Apple Pay, custom banners and material variants to it, which is Apple saying plainly that this is a commerce feature.

Google's Scene Viewer does the equivalent job on Android, reading glTF 2.0 and GLB, and can be launched straight from a web page. Google's own documentation recommends staying under 100,000 triangles and 10 materials.

Sitting on top of both is model-viewer, a web component maintained by Google that renders a 3D model in a page and hands off to whichever AR system the device has. It is the reason a modern 3D product page is a few lines of HTML rather than a project.

2020: the first widely quoted numbers

Shopify posted in September 2020 that interactions with products having 3D or AR content converted 94 percent more often than the same kind of products without. It is the single most cited statistic in this industry and it deserves a caveat every time: it compares product interactions across merchant data, not a controlled trial, and merchants model their better products first.

Shopify's more granular retailer figures are easier to trust because they are per-brand. Rebecca Minkoff shoppers were 44 percent more likely to add to cart after viewing a 3D model and 27 percent more likely to order, with a 65 percent purchase lift on AR-enabled pages. Gunner Kennels reduced return rates by 5 percent after adding AR.

Two things happened close together that are usually discussed separately.

Apple Vision Pro went on sale in the United States on 2 February 2024 at 3,499 dollars. Samsung's Galaxy XR, the first Android XR device, launched at 1,799 dollars. Both are real products with real app stores. Neither is a shopping channel at those prices and those unit volumes.

Meanwhile Meta shut down the Meta Spark platform on 14 January 2025, and third-party AR effects on Instagram, Facebook and Messenger stopped working. A large amount of brand AR, all those try-on filters agencies built between 2019 and 2023, simply switched off. Meta kept its own effects and moved its investment towards glasses.

If you want one lesson from this history, it is that one. The AR that survived is the AR that lives on your own product page in an open format. The AR that disappeared lived inside somebody else's platform.

Where it actually stands now

The honest summary in 2026: web AR on the product page is mature, cheap and boring in the good sense. The formats are settled, GLB and USDZ. The viewers ship with the operating systems. The cost of a model has fallen far enough that subscription pricing works, which is why our own pricing starts at 10 dollars per product per month on Basic with a 12-month term and a 100-variant minimum, and 20 dollars per model per month on Value with a 24-month term.

What has not happened is mass adoption of AR as an experience in its own right. Shoppers do not go looking for AR. They go looking for an answer to a question about size, colour or construction, and AR is one of the better ways to give it to them. Our clients treat it that way. Sweef used it to make a modular sofa understandable. Contura used it for a variation-heavy stove range. MELIMELI used it to move part of the showroom experience online.

If you want to see where your own catalogue sits in that story, we will build one 3D model of one of your products for free. Send us a product and it is yours to keep. What we do beyond that is on solutions.