How to Overcome Marketing's ROI Measurement Crisis
Juliana Rosso Frediano breaks down the impact of AI on data integration and common pitfalls in B2B channel attribution.
Fragmented data and the rush for immediate returns distort marketing campaign evaluations. In a discussion on measurement, analysis shows how statistical modeling and automation help evaluate channels without relying solely on the final click.
Juliana Rosso Frediano, CMO at First Answer, addressed data fragmentation and ROI measurement bottlenecks in a conversation with Bruno Gimenes on Marketing - The Podcast. The debate draws on market research featured on the show: data from Nielsen's CMO Outlook 2026, via Business Wire, indicates that 83% of CMOs remain confident in the value of their brands, while coverage of CMO Summit 2026 by Mundo do Marketing reveals that 54% of leaders view connecting fragmented data sources as their main hurdle, and 23% admit to a crisis of confidence in ROI.
To understand the full context of the discussion—part of a broader episode on enterprise revenue architecture—watch the video above to follow the full dynamic between the participants.
Decentralized reporting and the role of AI
Assessing the landscape of dispersed metrics, Frediano argues that advances in AI tools have reduced reliance on traditional analytics structures. According to her, a process that once required internal technical ticketing queues can now be operated directly by marketing teams through automations and structured queries to repositories.
Frediano recalls the sector's previous workflow: "We used to have a BI department where you had to get in line and ask the BI team to build dashboards and reports to help us show campaign performance, investment returns, and everything else." For her, having direct query mechanisms at hand simplifies report creation without the service bottlenecks that historically delayed proving value.
Timeline misalignment and attribution models
The perceived inefficiency in calculating ROI stems, according to Frediano, from attempting to measure distinct channels across the exact same timeline. In a B2B environment with complex solutions, such as ERP systems or CRMs, demanding immediate conversions from relationship-building initiatives ignores the true length of the sales cycle, which can easily span three to six months.
Highlighting the pressure for quick returns from in-person activations, Frediano notes the frequent friction with finance teams: "Oh, we ran an event this month. And the very next week, finance is asking how many new contracts were closed from that event. Forget it—that's not how you measure event ROI." She points out that assessing the metric in the subsequent month fails to reflect the reality of the business.
The issue persists in digital media when last-click attribution is treated as the absolute measure of success. Frediano underlines the shortcomings of this approach: "The person who saw a campaign didn't necessarily go there, click last, and justify attributing that sale or new business solely to that channel. Once again, ROI is being measured incorrectly."
Statistical modeling and causal inference
To navigate the complexity of multiple touchpoints, Frediano points to Marketing Mix Modeling (MMM), citing market tools such as Purple Metrics, founded by Guta Tolmasquim and mentioned by Bruno during the conversation. The model correlates online and offline channel investments with sales volume to estimate statistical impact over time.
Bruno notes growing interest in applying causal inference across the tech sector, advocating for a more rigorous understanding of the logic connecting each marketing action to closed deals. Frediano agrees that decoding cause-and-effect relationships beyond simple numerical correlation is key to ensuring performance metrics remain credible.