Marketing ROI: Incrementality Reveals 1.8x ROAS in 2026

Most marketers are stuck on last-click attribution which means they’re not seeing the real impact of their campaigns. This teardown shows how we used incrementality testing to get a completely different ROAS number than the ad platforms were telling us, proving that a lot of businesses are misjudging their own performance by not measuring the true, incremental effect of their ads.

Campaign Overview: Driving E-commerce Sales for a Niche Apparel Brand

We worked with a medium-sized e-commerce brand selling sustainable outdoor apparel that wanted to boost online sales for their new winter collection. The main goal was a 2.0x return on ad spend (ROAS), but it was also about figuring out which channels were actually driving new sales that wouldn’t have happened otherwise. Their historical reliance on last-click attribution meant they kept over-investing in bottom-funnel tactics that just scooped up existing demand, so we convinced them to focus on incrementality instead.

The campaign had a $150,000 media budget and ran for eight weeks, from October 1 to November 26, 2025, timed perfectly to capture early holiday shopping intent. We tracked the core KPIs: purchase conversion rate, cost per acquisition (CPA), and ROAS.

Strategy and Targeting: A Geo-Lift Approach

We ran a multi-channel campaign on Meta Ads (Meta Business Help Center) and Google Ads (Google Ads Help), but we put a lot of the budget into brand awareness and consideration, not just direct response. The whole measurement plan hinged on a geo-lift experiment. We picked 20 matched US regions based on population, past sales, and demographic profiles. Ten became “test” cells that got all the ads. The other ten were “control” cells where we shut off ads for the entire test period.

This setup let us isolate the real impact of our ads by comparing sales in the test regions to the control regions, which gave us a clean baseline to measure against. For the test regions, we targeted lookalike audiences from their existing customer lists, people with interests in outdoor activities, and specific demographic segments (ages 25-54, household income $75,000+). We intentionally ran upper-funnel placements like Instagram Stories and Facebook Feeds on Meta, which we balanced with lower-funnel Google Search and Shopping campaigns.

Creative Approach: Storytelling and Product Utility

For the creative, we focused on storytelling around the brand’s sustainable manufacturing and also hammered the functional benefits of the new winter gear. On Meta Ads, this meant a bunch of 15-30 second video ads showing real people using the apparel outside, playing up the durability and warmth with a user-generated content feel to make it more authentic. We also ran carousel ads to show off collections with direct links to product pages. On Google Search, the ad copy was all about specific features and seasonal keywords like “waterproof jacket” and “thermal base layers,” while Google Shopping ads just showed off good product shots and competitive pricing.

Initial Performance Metrics (Last-Click Attribution)

After the first four weeks, the platform-reported metrics looked just okay. We were seeing a 1.2x ROAS based on last-click, which was way below our 2.0x target. Cost per acquisition (CPA) was averaging $45 against an average order value (AOV) of $110 for the new collection, and the click-through rate (CTR) was sitting around 1.5% across 18.5 million impressions. At this point, the client was getting nervous and started talking about cutting the budget or resorting to heavy discounts.

The Incrementality Test: Uncovering Hidden Value

And this is where the geo-lift test paid off. Once we had four weeks of data, we could finally analyze the incremental impact by comparing the normalized sales growth in the test regions against the control regions. The results were pretty clear.

We saw a 15% incremental lift in sales in the test regions over the control regions. That means 15% of the sales in those test areas simply would not have happened without the ad campaign. When we applied that lift to the sales numbers, the story of the campaign’s performance changed completely.

The last-click models were crediting us for $90,000 in revenue (the 1.2x ROAS on a $75k spend). But our incremental analysis showed the ads had actually generated an additional $35,000 in sales that were truly *caused* by the campaign. This bumped the incremental ROAS to 1.8x. While still shy of the 2.0x target, it proved the campaign was much healthier than it looked, and the real CPA dropped to $32.14. It’s the classic last-click problem: the platforms were taking credit for conversions that would’ve happened anyway.

This is a critical distinction for anyone managing a budget. A marketer looking only at the platform’s 1.2x ROAS might have cut this campaign, killing what was actually a source of real growth.

Optimization and Refinement: Weeks 5-8

With this incremental data in hand, we made some key changes for the second half of the campaign. Our focus was no longer about maximizing last-click conversions. It was about maximizing incremental lift. We shifted an extra 10% of the remaining budget to the video creatives that showed the highest incremental impact which was especially obvious in our Meta campaigns where certain video stories clearly drove a higher sales lift in the test groups.

Specifically, we found that videos with testimonials from real outdoor enthusiasts had a 22% higher incremental conversion rate than our more polished, brand-focused content. We also changed our bidding to prioritize reach in the best-performing geo-segments, even if it meant a slightly higher reported last-click CPA there. We had to have a tough conversation with the client, who saw the platform-reported CPA going up in some areas and wanted to pull back, but showing them the incremental lift data gave us the proof we needed to push forward. We also cut a bunch of Google Search keywords that, while looking good on last-click, showed zero incremental effect. They were just capturing existing demand, so we expanded into broader, awareness-focused terms around sustainable fashion and outdoor gear instead.

By the end, the client understood exactly which ads were making a difference and which were just riding on existing demand. Our optimization phase resulted in an 18% improvement in incremental efficiency, we generated more new sales per dollar spent. The final incremental ROAS for the whole campaign landed at 1.8x. It missed the 2.0x goal, but it was an accurate number that reflected the campaign’s true impact. Relying only on last-click metrics here would have led to bad budget decisions, potentially penalizing campaigns that were driving genuine new demand.

Make Incrementality Testing Continuous

This campaign taught a clear lesson: incrementality testing needs to be an ongoing process integrated right into your measurement framework. Without it, marketers are just guessing, mistaking correlation for causation and likely wasting money on channels that don’t actually drive new business. We told the client to permanently set aside 10-15% of their media budget for dedicated test cells for continuous measurement. This commitment to testing allows for continuous learning, ensuring future campaigns are built on real impact.

Customer journeys are messy, which means you need measurement that can handle it, like incrementality testing. Any marketing team not actively running these tests is leaving a huge blind spot in their strategy. This approach complements last-click attribution by revealing the true causality of your marketing. Industry trends, like the IAB’s MAP Report, also point toward the need for strong, privacy-safe methods like incrementality.

Marketers are often intimidated by the perceived complexity of setting these tests up. And while they do require careful planning and some statistical know-how, the tools and methods are getting much easier to use. Both Google and Meta Ads now have more built-in tools for running controlled experiments, so it’s not just for giant brands anymore. The biggest barrier is getting teams to shift their focus from easy-to-read platform metrics to demonstrable lift.

The investment in testing pays for itself by stopping you from wasting money. The alternative is continuing to pour budget into channels that look good on a last-click basis but are really just cannibalizing organic sales or capturing demand that was already there. For a brand with tight margins, that’s a mistake that can drain a budget fast.

Being able to walk into a budget meeting and say, “This campaign drove exactly this much *new* revenue” gives you a real competitive edge and makes it easier to get more funding. Understanding incremental impact also makes your creative optimization way more effective. Knowing exactly which message or visual actually gets someone to buy lets you refine your content strategy with real precision. For this apparel brand, that meant doubling down on authentic user stories and their sustainability message, not generic product shots. This iterative process of testing, measuring lift, and optimizing is the core of modern data-driven marketing in 2026.

Incrementality testing lets you make data-backed decisions that drive actual growth, getting you past the limits of last-click models to see what really works.