How Thrive Causemetics is using Haus’ AI agent to unlock $700K+ in incremental revenue
- 2.4%
- lift over business as usual in new customer Shopify revenue
- ~$700K
- annualized and nationalized incremental new customer DTC revenue
- 17 days
- head-to-head GeoLift test, Architect vs. business as usual

- Industry
- Beauty and Personal Care
- Channels
- YouTubeAppLovinMetaGoogleTikTok
- Features used
- ArchitectCasual MMMGeoLift
The challenge
Would you trust an AI agent to know where to make marketing investments?
At Thrive Causemetics, Head of Data and Business Analytics Barron Kennedy II wanted the answer to be “yes,” – after all, his team had invested significant time, effort, and resources into learning from their Haus incrementality experiments, and in theory, the promise of Haus’ platform learning from these experiments to make budget recommendations made sense. But before trusting an agent to guide the team in the right direction, Barron and his team wanted to be certain the agent would outperform their business as usual (BAU) decisions.
Haus has been an outstanding partner to our team. We were excited to test Architect’s recommendations because it’s built on a foundation of causal data we already trust. But agentic decisioning is new for everyone, and just like we’d test a channel’s incremental performance with a geo-experiment, we figured it’d make sense to test the agent’s performance against BAU. So we did.
The solution
A head-to-head GeoLift test of Architect against business as usual
Architect is Haus’ decisioning agent – among a suite of broader capabilities, it generates immediate, specific recommendations for how marketers can improve their mix, measures whether the shift was effective, and adapts.
All those untold hours spent downloading spreadsheets and turning knobs and levers and pulling your hair out? That’s what Architect is good at doing. It’s embedded directly into Haus’ Causal MMM, providing answers to critical questions like, “What’s the most impactful action I can take today to hit my monthly, quarterly, or annual goals?” and “Did that budget shift work?”
With the support of Thrive’s executive team and the Haus team, Barron set out to test the first experiment of its kind: A head-to-head test pitting Thrive’s BAU media allocation against Haus’ Architect recommendations. “This was scary,” acknowledges Barron. “This was the first time anyone had rigorously tested an AI agent’s media allocation recommendations, and we know there’s a cost to getting it wrong.”
The experiment was an 80/20 split: 80% of the country received Thrive’s BAU allocation and 20% received Architect’s allocation. The test ran three weeks and focused on Thrive’s New Customer revenue on Shopify across the US.
We were glued to our Haus dashboard every day. It was hard not to be skeptical, but we were also hopeful. The goal was to figure out if Architect could show us insights and recommendations we wouldn’t have found or known about otherwise. So for example, from running Haus incrementality experiments, we knew, directionally, we were probably over-invested in AppLovin and TikTok. But we didn’t know where to re-allocate that spend, and how much to re-allocate, for it to work harder for us. Our hope was that Architect’s recommendations could answer that question.
During the 17 day experiment, Thrive implemented Architect recommendations that gradually right-sized their overspending on AppLovin and TikTok, reallocating that spend to Google Shopping and upper-funnel Meta Prospecting campaigns.
The result
Architect generated 2.4% lift over business as usual
After nearly three weeks of implementing Architect’s recommendations, the results were in: Architect generated 2.4% lift over BAU in New Customer Shopify Revenue for Thrive.

When annualized and nationalized, that tracks to nearly $700K in incremental new customer DTC revenue from changes to Meta, Google, TikTok, and AppLovin. And that’s a floor: The test ran during a low-spend evergreen period on DTC alone. Applied across Thrive’s full media mix, sales channels, and peak seasons, Thrive believes Architect could drive multiple millions in incremental revenue.
Architect has been a huge win for the business. We didn’t go in expecting fireworks. What surprised us was how simple it was on our end: Architect kept making small, smart recommendations, and by acting on them we were able to decisively outpace our BAU. We couldn’t have asked for a better outcome.
Going forward, the Thrive team expects to rely on Architect recommendations during their most critical business moments.
Next steps
About Thrive Causemetics
Thrive Causemetics is an independent, female-owned beauty brand that creates high-performance vegan cosmetics and supports a community of giving. All of Thrive Causemetics’ products are free of parabens and sulfates, in addition to being 100 percent cruelty-free. Through its Bigger Than Beauty™ program, every product purchased results in a donation to help communities thrive.
Published: November 19, 2026
