"Data Must Serve the Business," Says Laerte Finotelli
IBOPE product lead discusses attention metrics, Goodhart's law, and data governance in marketing and AI.
Laerte Finotelli examines the transition from pure reach metrics to attention quality in advertising. The discussion covers digital measurement blind spots, AI governance, and the impact of cognitive biases on data interpretation.
Laerte Finotelli, product lead at IBOPE, sat down with Bruno Gimenes on Marketing - The Podcast to discuss the technical and methodological hurdles of audience measurement, the shift toward qualitative attention metrics, and the strategic risks of misinterpreting statistical data.
Throughout the conversation, Finotelli highlighted how product science and behavioral neuroscience can prevent common marketing measurement pitfalls. The complete video above covers all the cited examples and in-depth discussions.
Career path: from Volkswagen engineering to IBOPE
Finotelli began his career in the automotive industry in the ABC region of São Paulo after studying mechatronics and production engineering. During his three-plus years at Volkswagen, he observed the sharp contrast between the strictly technical mindset of engineers and the broad emotional connection required by automotive advertising campaigns.
This insight prompted him to pursue postgraduate studies in Big Data and Marketing at ESPM, where he was introduced to IBOPE's sports analytics division. Finotelli joined the company in 2019, building his career across measurement products and establishing multidisciplinary teams with a strong statistical foundation.
Attention measurement, the value of podcasts, and TV vs. digital at the World Cup
The conversation around audience measurement is undergoing a methodological shift: quantitative reach remains fundamental, but it must be enriched with qualitative layers of attention and context. Finotelli pointed out that immersive formats such as podcasts offer a voluntary, receptive listening environment, unlike forced-exposure digital ads.
Analyzing World Cup broadcasts, Finotelli highlighted how broadcast television and digital channels like CazéTV complement one another. "Broadcast television signals, due to established transmission infrastructure, retain a clear latency advantage over digital streams," Finotelli explained, noting the real-time migration of viewers between screens during goals and commercial breaks.
Advertising metrics and data serving business strategy
Finotelli argued that data must strictly serve business strategy rather than dictate direction in a vacuum. Traditional television and video measurement deliver demographic and commercial-break reach data, leaving agencies and brands responsible for correlating those numbers with sales and brand recall.
Inverting that order undermines effective decision-making, Finotelli cautioned. "Data should never come first. The question must always be: what is the actual business need? Data must serve that purpose," he noted.
Goodhart's law: when a metric becomes the target
Goodhart's law manifests when a proxy metric turns into the primary goal, distorting team behaviors and compromising actual business value. Finotelli illustrated this with contact centers that enforce strict targets for shorter handling times and higher call volumes.
Under such pressure, agents tend to disconnect calls prematurely just to hit the numbers, ultimately degrading service quality. "You might meet the target, but you move further away from the real business objective. Goal setters must be careful not to chase numbers for numbers' sake," Finotelli emphasized.
Digital measurement blind spots: walled gardens vs. probabilistic models
The digital ecosystem continues to struggle with fragmentation caused by walled gardens—closed platforms that refuse to share standard cross-channel data. Finotelli contrasted digital deterministic models with traditional probabilistic panel sampling, warning against superficial engagement metrics based on just a few seconds of screen time.
He pointed out that variables such as screen size, active sound, and even thumb placement while scrolling affect how an audience absorbs a message. "Creating shared measurement standards generates value across the entire industry. That is precisely why attention metrics have gained so much traction," Finotelli stressed.
AI governance and auditing automated decisions
Automating business decisions through AI agents demands rigorous governance frameworks that combine automated data processing with human contextual evaluation. Finotelli warned that machine learning models frequently correlate variables that do not align with corporate strategy or ethical considerations.
To prevent scaled distortions, Finotelli advocates using automated tools to provide grounded recommendations while keeping final validation in human hands. "There must be a systematic layer: understanding when removing human bias is beneficial, but also knowing where human oversight is vital to keep models grounded in reality," he questioned.
Neuroscience, heuristics, and essential questions in data analysis
Drawing on his background in behavioral neuroscience, Finotelli addressed how cognitive biases distort reporting. While heuristics serve as efficient everyday mental shortcuts, they can lead to systematic errors when practitioners dismiss statistical evidence in favor of gut instinct.
Before reviewing any data report, Finotelli recommends asking three fundamental questions: What does the raw data literally represent? What precise question does this metric answer without overstepping its scope? And what additional data points are required to provide complete context?