It turns out that AR isn’t the world-shifting technology it was trumpeted to be in its circa-2017 hype cycle. It’s not a silver bullet in all areas of life and work… but it is proving meaningful and transformative in a few high-value areas such as enterprise productivity and brand marketing.

This list will grow, but these are the early leaders where AR is producing real results today (read: revenue). Zeroing in on brand marketing specifically, sponsored AR lenses and virtual try-ons offer the ability to demonstrate products in immersive ways — from cosmetics to couches.

For example, when looking at paid ad placement to amplify these lenses (just one piece of the AR marketing value chain), our research arm ARtillery Intelligence projects $6.7 billion in revenue by 2025. This is fueled by a feedback loop of positive results from brand campaigns.

But one question looms over all this: how do you measure AR campaign performance? Tried and true 2D metrics like impressions and click-through rates (CTRs) don’t capture AR’s depth of experience. That depth can achieve things like greater memory recall and “ craveability.”

This all boils down to the question of what’s the CTR of AR? How do we standardize and normalize AR marketing with a common meter stick that lets marketers benchmark performance? Snap may have an answer to this question, which it calls the Marketing Mix Model (MMM).

The idea is that AR — or any ad format for that matter — doesn’t happen in a vacuum. Most brand marketers have several irons in the fire (or “omnichannel” if you like jargon) considering social, search, email, display, video, etc. How does AR fit into the mix and stack up?

In short, Snap worked with Nielsen to convert AR performance to metrics that can be compared with the above formats. It tracked 15 brand lens campaigns and correlated their activation to three years of sales data, brand-health tracking and other KPIs (full methodology here).

Essentially, what this all means is that Snap’s answer to the question of how we measure AR is to create a sort of conversion formula that lets us compare it to traditional metrics in legacy media. This lets marketers get an apples-to-apples view of its relative performance.

And the results? AR Lenses achieved a 1.67x return on ad spend (ROAS) across these fifteen brands. This compares with other formats in the media mix such as television (1.04x), other social media (.93x), and the overall average (.77x). Anything less than 1 means a negative ROI.

Using the same 15 brands, Snap applied YouGov’s standard BrandIndex score. This is the “brand-health tracking” referenced above, which applies several KPIs like brand awareness. AR lenses scored 339, versus television (101) other social (43), and the full index (100).

Stepping back, an inherent challenge with new technologies is that they need to be revolutionary (see Dan Bricklin’s law of 100x) to replace what came before them. But that same order of magnitude difference is what causes adoption friction from comfort-bound stakeholders.

And it doesn’t get much more “comfort bound” than Madison Avenue. It took the broader collective ad industry years (to this day) to adequately adopt native mobile and social strategies. And a big part of that isn’t just new formats and their effective execution but how they’re measured.

Put another way, the ad world tends to stick to what it knows, such as impressions and CTRs as noted. We’ve seen some movement in new and native metrics for AR marketing, such as 8th Wall and Poplar’s emphasis on dwell time. On that measure, AR outperforms formats like video.

Of course, the ultimate metric is revenue. It’s universally understood and applies across ad formats. But a direct line from ad performance to revenue isn’t always clear, so the advertising world relies on proxies for, or extrapolations of, revenue impact — again… clicks and impressions.

This is all to say that AR’s challenge isn’t just to develop new metrics, but to rip off the collective security blanket of consumer brands and ad agencies. If history is any indication, that will be a decades-long process. so perhaps it’s more about weaning than ripping.

“With media agencies and creative agencies, it’s hard to reframe your mind in this new medium,” M7 Innovations’ Matt Maher told us. “So you benchmark it based on the past medium. […] But slowly, we’re seeing the smart marketers say, ‘Okay, we’ll create new benchmarks. We’ll start to adapt what we know.’ You don’t want this to be a bolted-on innovation. If it’s measured wrong [the client is] never going to do AR again. So you kind of need that balance of both to justify the spend.”

Data Dive: How Should We Measure AR Marketing? was originally published in AR/VR Journey: Augmented & Virtual Reality Magazine on Medium, where people are continuing the conversation by highlighting and responding to this story.