How OluKai Unlocked a 20% CAC Reduction Overnight from Causal MMM

As OluKai headed into Q4, the team was under pressure to drive strong returns and ensure every marketing dollar was working efficiently. They had been relying on an existing MMM vendor whose models consistently recommended prioritizing view-through heavy channels such as TikTok, Snap, and YouTube. However, their Haus GeoLift results showed that TikTok and Snapchat were less incremental than Meta.

Acting on the legacy guidance, OluKai shifted incremental budget into those channels and reduced Meta spend. But the forecasts weren’t matching reality: CAC was 20% higher than predicted, leadership grew increasingly concerned, and the MMM’s guidance was not aligning with what the team was seeing in their actual revenue outcomes.

With peak season quickly approaching, every shift in spend had a significant impact. Even the smallest of inefficiencies in channel allocation could quickly compound into meaningful margin pressure — especially as auctions became more competitive and customer acquisition costs naturally rose during Q4 ramp up. OluKai needed a measurement approach they could trust to make confident decisions in real time, not weeks after the fact.

Despite increasing investment according to their previous vendor’s model, performance continued to deteriorate. OluKai needed clarity — fast.

Haus’ Causal MMM (cMMM) is powered by OluKai's causal experiments, not just historical correlations. That means channel recommendations are anchored to incremental lift — giving the team confidence they're funding true growth drivers rather than performance that only looks directionally efficient.

OluKai turned to cMMM for a more accurate, scientifically grounded read on channel performance. While their former MMM partner reported strong ROAS for channels like TikTok, Snap, and YouTube, Haus’ causal model revealed a very different reality. cMMM showed that the channels previously recommended for increased investment were not incrementally driving revenue, while Meta and Google were actually the most efficient drivers of incremental sales.  Since cMMM was rooted in their causal experiment results, they trusted the recommendations and made two key shifts overnight:

• Reallocated approximately 50% of view-through channel spend into Meta
• Reduced total marketing spend by ~15%, right before entering their most important quarter

The impact was immediate and measurable:

• CAC dropped ~20% overnight.
• cMMM’s order volume and CAC forecast matched actuals, validating the model’s causal accuracy.
• Performance improvements held for several weeks, even with lower total spend.
• Leadership gained renewed confidence in spend allocation and forecasting heading into the most important part of the year.

OluKai gained efficiency while also improving their forecasting. cMMM’s order volume and CAC projections are now closely aligned to actual results and performance held for weeks, even at ~15% lower spend — improving performance and reducing risk during the most competitive part of the year.

By grounding their decisions in causal truth, not correlation, OluKai was able to course-correct quickly, increase efficiency, and avoid overspending during a critical period.