AI Mode: Aura Home Goods’ 2026 Discovery Success

The so-called AI Mode in Google and Meta’s ad platforms has completely changed the game for discovery campaigns, giving us a new toolkit for reaching audiences through paid channels. This means we have to rethink our old PPC playbooks, especially if the goal is to grow beyond our current customer base. The job now is to figure out how to pilot these new AI features to actually build brand awareness and get new customers without the budget getting out of control.

Campaign Teardown: Launching “Aura Home Goods” with AI Mode Discovery

Back in Q1 2026, we ran a brand discovery campaign for Aura Home Goods, a new DTC brand doing sustainable home decor. Our main goal was to get the name out there with environmentally conscious shoppers who probably weren’t searching for things like “recycled glass vases” yet. We were looking for a big jump in unique website visitors and a clear lift in branded search volume within three months. The primary objective was building an audience for later, not chasing immediate sales, though we were definitely tracking them.

Strategy: Blending Broad Reach with AI-Driven Precision

We built our strategy around the AI Mode features in Google Ads’ Demand Gen and Meta’s Advantage+ Shopping Campaigns. Our theory was simple: these AI systems, with all the signals they process, should be able to spot early interest patterns way faster than we ever could with manual keyword or demographic targeting. So, we started with a wide net, giving the AI room to learn who to target instead of restricting it from the get-go, which is a common mistake people make with these systems.

The campaign ran for 10 weeks, from January 8th to March 18th, 2026, with a total discovery phase budget of $45,000. We put 60% of that into Google Demand Gen for its broad reach and the remaining 40% into Meta Advantage+ Discovery, banking on its strength for visual brand storytelling.

Creative Approach: Beyond Product Shots

For a discovery campaign to work with AI, you need a ton of creative. We built out a library of over 70 unique assets. This wasn’t just product photos. It included:

• Short-form video (15-30 seconds): These showed the Aura products in real-life settings, hitting on the sustainability angle. They worked especially well on Meta.
• Carousel ads: We used these to show off product collections and talk about the brand’s ethical sourcing story.
• Static image ads: These were aspirational shots of beautiful rooms with Aura products, paired with quick, punchy copy about the brand’s values.
• Interactive polls and quizzes: We ran simple polls asking about home decor styles or sustainability habits to get engagement and collect interest signals without being creepy about it.

Everything had to tie back to the brand narrative: “Sustainable Living, Beautifully Crafted.” We used a consistent warm color palette and authentic photography. We also tested CTAs, and found that softer ones like “Explore Collections” worked much better than hard-sell CTAs in this top-of-funnel phase.

Targeting: AI’s Learning Curve

In Google Demand Gen, we started by feeding it broad signals like “Home Decor Enthusiasts” and “Sustainable Living,” but the key was letting the AI expand beyond those. We uploaded a hashed list of our newsletter subscribers to give the algorithm a starting point, but the main instruction was to go find new people who looked like them. We also configured the campaigns to optimize for “New Customers,” which tells the AI to prioritize people who haven’t been to the site before, even though sales weren’t our main KPI.

On Meta, we turned on Advantage+ Discovery and just gave it our full product catalog and a similar broad interest set. The platform’s AI took it from there, creating its own ad variations and audience buckets. We watched the insights dashboard and saw it finding segments we would have missed, like “Urban Gardeners” and “Minimalist Design Aficionados.” This is the real value of the AI, it uncovers these little pockets of customers you would’ve never thought to target yourself.

Performance Metrics: What Worked and What Didn’t

Here’s the raw data from the 10-week campaign:

Overall Campaign Performance

• Total Impressions: 18,500,000
• Total Clicks: 210,000
• Overall Click-Through Rate (CTR): 1.14%
• Unique Website Visitors: 175,000
• Cost Per Unique Visitor: $0.26
• Total Conversions (Micro-conversions like email sign-ups, catalog downloads): 8,500
• Cost Per Micro-Conversion (CPL): $5.29
• Return on Ad Spend (ROAS) from Direct Purchases (Secondary Goal): 0.8x

Platform-Specific Performance

What Worked:

1. The AI Found New People: The AI on both Google and Meta was great at finding audiences way outside our initial ideas. For example, Google found a big group of users interested in “sustainable architecture” and the cost per unique visitor from that segment was 15% lower than the campaign average. That’s a group we wouldn’t have targeted manually.
2. Video Was King: Short-form videos, especially on Meta, killed it compared to static images. The ones showing products in a real home with nice natural light got a 25% higher CTR. For a brand that’s all about aesthetics, showing the product in motion was the right call.
3. Tracking Small Wins: Focusing on email sign-ups and catalog downloads as our main conversion goal was the right move for this phase. These actions told us who was actually interested (even if they weren’t ready to buy), which in turn gave the AI better signals to optimize toward finding more of those people.

What Didn’t Work:

1. Vague CTAs Flopped: Our initial tests with generic CTAs like “Learn More” were a waste of impressions. People need more direction, even in discovery. When we switched to more specific calls to action like “Discover Sustainable Home Decor,” engagement went up.
2. We Let Creative Get Stale: In the first month, we only did one creative refresh, and we saw performance dip, especially on Meta. The algorithms seem to get bored. We learned you have to swap in new creative (at least 30% of your assets) every two weeks to keep things from getting stale.
3. Attribution Is Still a Mess: It’s hard to prove the long-term sales value of a discovery campaign. We saw a 12% lift in branded search for “Aura Home Goods,” which is great, but tying a specific sale three months later back to a specific impression from this campaign requires multi-touch attribution models that are still more theory than practice for most companies.

Optimization Steps Taken

We were tweaking things constantly over the 10 weeks. Here are the main changes we made:

1. Bi-Weekly Creative Refresh: Starting in week 5, we got on a strict schedule of rotating in new videos, photos, and polls every two weeks. This simple change boosted our overall CTR by 10% for the second half of the campaign.
2. Adding Negative Audiences: We were constantly checking the audience reports. On Google, we saw we were getting clicks from people interested in “cheap home decor,” but they never signed up for anything. So we excluded that segment, which made the budget more efficient and dropped our CPL by 8% in the last few weeks.
3. Adjusting Bid Strategies: We started on “Maximize Conversions” for our micro-conversions. Once we saw that video ads were doing really well, we moved those ad groups to a “Target CPA” strategy, which let us tell the AI to bid more aggressively for those high-value engagements.
4. Landing Page Tests: We ran A/B tests on our landing pages. The winners were always the ones with big lifestyle photos up top and an obvious email sign-up form. Pages that looked like a product grid or made people scroll too much had high bounce rates.
5. Feeding the Machine: We kept uploading our first-party data (like lists of recent buyers) as custom audiences. This helps the AI get better at its lookalike modeling because it has a constantly updating picture of what our best customers look like. Giving these systems high-quality, fresh data is probably the most important thing you can do.

The Aura Home Goods campaign proved that AI Mode’s discovery features aren’t just a minor update. They change the whole top-of-funnel marketing process. It takes a big up-front investment in creative and you have to watch it like a hawk, but the way these systems can find entirely new customer segments makes them a must-use for any brand trying to grow in a crowded space.

So yes, use AI Mode for discovery. Just know that it’s not a set-it-and-forget-it tool. Success depends on you feeding it a steady stream of good creative and providing enough oversight to guide it. The future of customer acquisition is clearly going to be a partnership between human strategy and AI execution, not just manual button-pushing.