What’s Driving Houzz’ AR Success?
A lot can be learned from consumer AR’s early leaders. What are they doing right? How are they engaging users? And how are they making money? These are key questions in AR’s early stages, as there’s no standardized playbook and lots of experimentation underway.
This exercise includes pinpointing fitting AR use cases, as well as more granular strategies around user experience (UX). What types of AR interactions resonate with consumers? And what best practices are being standardized for experience and interface design?
After diving deep in Parts I (Snapchat) and II (Niantic), we turn attention in Part III to the broader set of innovators emerging in today’s market. These are AR’s less proven, but still promising, next class. And to make sense of the competitive field, we’ve divided it in three parts.
Some companies are established players moving into AR (AR converts). Some are AR-native players building tools and technologies to advance the state of the art (AR enablers). And some are players just out of the gate and showing promise to represent AR’s next phase (AR hopefuls).
We’ll dive into the converts first. Picking up where we left off last week in examining Instagram’s AR ambitions, we’ll now dive into another noteworthy AR convert: Houzz.
AR Converts: Houzz
One company that’s received considerable accolades for AR commerce is Houzz. The home furnishings company offers AR to visualize products prior to purchasing. Its 11x boost in conversions from AR visualization (see below) is one of AR’s most cited data points.
But how did it get there? Much of this success came under the tenure of Houzz’ visual tech lead Sally Huang. In fact, her title alone demonstrates the company’s commitment to AR. She joined Houzz four years ago to bring visual tech into the mix, though AR wasn’t the obvious path initially.
This started as a 2D version of its now popular AR “View in My Room” feature. It used flat stickers for in-room furniture visualization. Though primitive, Houzz operational scale gave it large-sample demand signals by which to test and iterate features. All signs pointed to keep moving towards AR.
“The mobile team took product photos, removed the background and used them as stickers,” she said at AWE. “They were able to apply this to just about any photo in the library… so if you were shopping for anything, there was an opportunity to view it in your room as a 2d sticker.”
With just that 2D sticker, Houzz saw a 3x boost in conversions. This was enough to get it thinking about more spatially-advanced integrations. Notably, this was before Apple’s ARkit so like other AR pioneers at the time, it had to do lots of in-house development.
It seemed like too much of a technical challenge, but when we saw the success… it became obvious it was something we had to work towards
If you were able to get such good results with 2D, imagine what it would look like [with] a full 3D integrated experience.
The Shopping Flow
Fast forward several product cycles later, and Huang’s team was ready to hit the ground running when ARCore and ARkit finally did release in 2017. It was able to take all the work it did and integrate these new native libraries for better functionality like tracking, scale and light estimation.
More importantly, all of that previous iteration and integration of View in My Room taught Huang & team an important lesson: AR needs to be in the shopping flow. It discovered through large-scale testing that AR features within users’ existing shopping path outperforms “standalone” AR.
“Part of what we think made our experience so successful was that we really had to answer the question of AR visualization in the shopping flow,” said Huang of the design principle also known as ‘AR as a Feature’. “So you can basically do the visualization, pop right out and check out.”
The result of all of this was the now-famous 11x figure shown above. The tactics that got it there were iterative product cycles, listening to customers, following usage data closely, and not requiring “activation energy” for users to access something new and unproven like AR.
Make AR Predictable
Now, Huang and team are doubling down on these principles. Just as they’ve been conditioned to create clever hacks out of necessity, they’re now doing the same to reconcile ARkit deficiencies like advanced vertical plane detection. After all, lots of Houzz products go on your wall.
People buy ceiling products, people buy wall products
So we basically spent a lot of time really thinking about the simple gestures we can introduce into our features to vastly expand on the amount of functionality and the amount of categories that we can handle.
Another important tactic is to make AR intuitive. ARkit and ARCore provide a great foundation. But it’s just that: Some degree of customization is needed for optimal UX in specific product classes, which is a good lesson for any AR-curious retailers. It’s all about making it predictable for users.
“When it came to user experiences, we basically opted for predictability,” said Huang. “You know, at the end of the day, AR itself has some unpredictable behaviors built-in. So wherever you can, as a design philosophy, [you should] introduce predictability. We basically follow that as a rule.”
Place AR in Users’ Paths
In addition to making AR predictable, it should be easy to find without “activation energy.” As examined last week in light of Instagram, AR is seeing the most success when it is placed directly in users’ paths. It should also be integrated in ways that align with users’ comfort and cognition.
For example, with the transactional functionalities for AR visualization and shopping, the entire purchase funnel should happen all in one flow as explored above. It shouldn’t bounce users to another app or website to further browse products and transact. It should all happen in one place.
This is important in AR’s early stages of consumer adoption when it isn’t yet proven enough to compel users to go out of their way or “work for it.” This involves the AR-as-a-feature approach which integrates AR in sparing ways that are organic and additive to an already-prevalent activity.
We’ll pause there and circle back in the next installment to continue looking at AR Converts.