Havan boosted Product ROAS by 25.7% with AI-powered SMART Segments

With over 350,000 SKUs live at any given time, Havan's Advantage+ Shopping campaign faced a structural problem common to high-SKU retailers: the algorithm had no signal on which products actually deserved more visibility. Every item in the catalog was treated equally, which meant ad exposure — and by extension, budget — tended to concentrate on the most-shown products rather than the ones that converted best.

This created an efficiency gap that couldn't be solved by simply shrinking the catalog or hand-picking hero products. Havan's team needed a way to direct spend toward genuinely strong performers without sacrificing the breadth of their assortment or taking on the impossible task of manually managing prioritisation across hundreds of thousands of products. The goal was clear: drive more efficient revenue from the products already proven to convert, without adding operational overhead.

To solve this, Havan deployed Segment Prioritisation with AI-powered Smart Segments, and validated the approach through ROI Hunter's Experiments A/B testing flow.

Two newly launched Advantage+ Shopping campaigns were set up with identical targeting, placements, and budgets — the only variable was the segment strategy. The Control campaign ran exactly as before: full catalog, no prioritisation, Meta's algorithm distributing exposure on its own terms. The Test campaign applied the Boost ROAS Segment Prioritisation strategy, layering Smart Segments' automated, daily-refreshed scoring on top of the same full catalog.

Rather than restricting the feed to a curated product set, the Test campaign kept every SKU active and simply boosted the products Smart Segments classified as Top Performers based on real, continuously updated performance signals. This meant Meta's algorithm never lost visibility into the full catalog and kept learning from every product, while spend was nudged toward the items already proving they could convert. No manual filters or thresholds were applied — by design, the whole point was to let Smart Segments do the work that manual catalog management never could at this scale.

The two campaigns ran side by side with virtually identical spend between the Control and Test cells, which meant any difference in outcomes could be attributed cleanly to the Segment Prioritisation strategy itself, not to budget or targeting differences.

The Test campaign delivered a 25.7% uplift in Product ROAS and a 25.5% increase in Product Revenue compared to the Control — at virtually identical spend. In other words, Havan generated meaningfully more return from the same budget simply by telling Meta which products deserved a boost.

The most striking secondary result was RPDV (Revenue Per Data View), which jumped by 50.7% in the Test campaign — a strong signal that the boosted segment wasn't just getting more clicks, but was driving traffic to products people were genuinely ready to buy. This was reinforced by a 16.3% increase in item quantity sold alongside a 7.9% growth in average selling price, showing that the boosted products converted more often and carried more value per sale. Interestingly, CTR dipped 7.8% while CPC rose 7.1% — consistent with Meta becoming more selective under the new strategy, trading raw click volume for fewer, better-qualified clicks that ultimately paid off in revenue and ROAS.

With a clean, controlled test proving that Segment Prioritisation could lift returns without shrinking the catalog or adding manual work, Havan is well positioned to extend the use of SMART Segments across more of its Meta campaigns — scaling smarter spend allocation to match the scale of its catalog.