Dentsu’s 2026 Vision: 3.2x ROAS for CPG

At Dentsu, Chrissie Hanson’s media vision is pretty straightforward: stop talking about impressions and start delivering measurable business outcomes. We’re expected to build campaigns that are accountable from day one, which sounds great in a pitch deck, but what does it actually look like when you’re trying to get results for a client in today’s messy media environment?

Campaign Teardown: “Future-Proofing Retail”

I just wrapped a campaign at Dentsu that’s a perfect real-world example of how we’re applying Chrissie Hanson’s philosophy. We called it “Future-Proofing Retail,” and it was for a major CPG brand in personal care that needed to fight off new competitors. They wanted to see real sales growth, both online and in stores, not just a report full of vanity metrics. So we dove headfirst into retail media networks and tied them into a bigger, integrated plan.

Strategy: Beyond the Click

Our strategy wasn’t about just buying cheap media placements. We mapped the entire path to purchase and built a multi-touch attribution model because last-click attribution is a dead end for understanding what actually works. The only goal that mattered was a measurable lift in sales. Everything had to point to that. We built the plan on a few key actions:

1. Integrated Retail Media Networks: We used the retailer’s own first-party data to get our ads in front of people already shopping on their sites.
2. Programmatic Precision: We ran programmatic display and video to find audiences that were actively in-market for our product category.
3. Connected TV (CTV) Engagement: We used non-skippable CTV ads to build the brand story and get people interested in the first place.

We had a $2.5 million budget to work with over a six-month duration. I put the biggest chunk (40%) into retail media because that’s where you get the most direct purchase data and the fastest return. Programmatic got 35% to find new customers, and CTV got the remaining 25% to build up that top-of-funnel demand.

Creative Approach: Contextual Relevance

Creative had to match the context. On retail media platforms like Walmart Connect and Amazon Ads, people are there to buy, so the ads showed product benefits and deals for that specific store. For programmatic, we used dynamic creative optimization (DCO) to swap out messaging based on browsing history and user data. CTV was different. It was all about lifestyle imagery showing the product in use, telling a story instead of a hard sell. We ended up producing over 50 unique creative assets and ran constant A/B tests to see what actually worked.

And the tests paid off. One of the clearest wins came from our A/B tests on a big retail media network, where we found that simply adding a clear “buy now” button and a price comparison boosted conversion rates by 22%. It’s a simple fix, but it proved that when someone’s on a retail site, abstract brand messaging just doesn’t perform as well as a direct call to action.

Targeting: Data-Driven Precision

Our targeting had to be incredibly precise. With retail media, we got to use the retailer’s first-party data, which is gold. We built segments of people who bought from the category before, were loyal to the brand, or even bought from competitors, letting us run specific win-back offers or loyalty plays. For programmatic, we layered on third-party data from Nielsen and eMarketer to hit the right demos and find people showing in-market signals for personal care. CTV was targeted by household income and what genres they were streaming. We spent a lot of time on exclusion lists too (a seriously underrated tactic) to stop showing ads to people who just bought something or were clearly not our customer, which saved a ton of wasted spend.

We even got as granular as geo-fencing our programmatic buys around specific neighborhoods like Atlanta’s Buckhead and Sandy Springs to push local in-store offers. This let us react to sales data from individual stores and tweak the campaign on a hyper-local level, something you can’t do with a broad national buy.

What Worked: Integrated Attribution and ROAS

The number that mattered most at the end of the day was the overall return on ad spend (ROAS) of 3.2x. We got to that figure with a real attribution model connecting ad views across every channel to actual sales, both online and in-store. It wasn’t a guess. We could clearly see how a CTV ad led to a brand search, which then led to a purchase driven by a retail media ad. Our programmatic cost per lead (CPL) came in at $1.75, which is way better than the typical $2.50 to $3.00 benchmark for CPG. And while we got over 150 million impressions and a solid 0.85% CTR, those were supporting stats, not the headline.

The real business impact was the 15% increase in market share we saw in syndicated sales data for the category. That’s the kind of number that gets you budget next year. That’s what Chrissie Hanson means by driving growth. Our final cost per conversion across everything was $12.30, giving us a hard number to beat for the next campaign.

What Didn’t Work: Over-reliance on Broad Demographic Targeting

We definitely made one mistake out of the gate. A slice of our initial programmatic budget went to broad demographic segments that didn’t have much behavioral data layered on. The performance was bad: ROAS was only 1.8x and the CPL shot up to $3.10. It was a classic case of chasing reach over relevance, and it just wasted money on people who weren’t in-market. It’s a common trap, but it’s a good reminder that reach by itself is just expensive noise.

Optimization Steps: Refining and Reallocating

Once we saw that poor performance from the broad segments, we acted fast. We pulled 15% of the programmatic budget and moved it into highly specific, intent-based audiences we built using real-time bidding data to find people actively looking at personal care content. We also doubled down on lookalike audiences built from our best-performing retail media converters. At the same time, we set a strict frequency cap of 3 impressions per user per week for programmatic and CTV to stop annoying people. Just making those changes boosted our campaign efficiency by 7% over the final three months.

We also got into the weeds on ad placements. We pulled reports to see exactly which sites and apps were working and which were garbage, then we started blacklisting the bad ones and putting more money into the good ones. It’s the kind of detailed work people sometimes skip, but it pays off. Blacklisting just 10 mobile apps that were full of bots or had zero engagement saved us about $50,000, which we immediately funneled back into placements that were actually driving sales.

This “Future-Proofing Retail” campaign shows what happens when you follow a philosophy like Chrissie Hanson’s and focus only on what you can measure. You stop chasing vanity metrics and start delivering actual business results like a 3.2x ROAS and a 15% market share gain.

The only way to get real results is to build an integrated plan that ties your media spend directly to sales, making every dollar accountable. If you want to see how AI fits into this, check out how AI transforms 2026 ad strategy for a 15% ROI boost. It’s a clear reason why CMOs must own AI strategy to succeed in 2026. And for more context on where media is heading, this piece on Dentsu’s Chrissie Hanson’s 2026 Media Reboot is worth a read.