Publicis could simulate the world, person by person.

There’s an unresolved debate about the two divergent data strategies that the holdcos are pursuing - deterministic 1st party data solutions (Publicis and IPG) and probabilistic 3rd party data solutions (the rest).

With IPG’s announcement last week, we can see a way to settle it.

You see the debate has been all about what I’d call ‘downstream’ applications of data, specifically how to target and track audiences of different levels of resolution. There are pros and cons oft argued: the cost and complexity for using 1st party for hyperpersonalisation, versus the use of 3rd party targeting at a level before diminishing returns kick in.

Last week, IPG just pulled ahead of Publicis for the first time in the AI race (despite being distracted by the Omnicom merger). IPG unveiled their "Simulation Studio" powered by AI startup Aaru who bring technology that creates digital twins of real people to test marketing scenarios before campaigns run.

The technical capability: Aaru's platform generates AI agents that simulate human responses with 85% accuracy, scaling from small focus groups of 5 agents to population-level studies of 100,000+ simulated people. Instead of waiting weeks for traditional research, agencies can test unlimited creative concepts, media strategies, and audience targeting in minutes.

The strategic integration: IPG's Simulation Studio connects directly with Acxiom's database of 2.3 billion consumer profiles. This means they're not simulating anonymous demographic segments, they're creating digital twins of actual customers whose behavioural patterns, purchase history, and preference signals they've tracked over time.

The speed claim: Aaru promises to "reduce four weeks of research into 40 seconds with higher accuracy than any survey, poll or focus group." More importantly, they can simulate specific scenarios: how will affluent parents in suburban Chicago respond to this creative concept versus that media timing versus this competitive context?

What makes this different: While other agencies are building AI tools for content creation or media optimisation, IPG is using AI to simulate human psychology at scale. They're not just improving execution, they're predicting behavioural responses before campaigns exist.

As I said earlier, the marketing industry has debated two fundamental data approaches:

Everyone focused on the downstream difference - targeting accuracy. Deterministic data targets real customers, probabilistic data targets statistical approximations. The performance gap was measurable but not transformational.

Simulation technology just revealed the upstream difference - and it changes everything.

With deterministic customer profiles, you can now simulate emotional responses, predict behavioural changes, and model decision-making processes before any campaign runs. You're not just targeting better - you're understanding human psychology at scale.

Probabilistic data cannot do this. Ever. You cannot simulate the emotional response of "statistical audience segment 18-34 female urban high-income." You can only simulate actual people whose behavioural patterns you've observed and recorded.

IPG's breakthrough: they can now simulate how Sarah Johnson from Portland will respond to different creative concepts based on her actual purchase history, brand interactions, and behavioural profile. They can test thousands of creative and media scenarios on her digital twin before spending a dollar on actual advertising.

This isn't incrementally better targeting. This is predictive human psychology at scale - something platforms do with their user data but agencies have never been able to match.

Until now.

1. Emotional Response Prediction You can simulate how specific personality types respond to different emotional appeals, testing creative concepts on digital twins of actual customers before production. Probabilistic audiences cannot generate emotional intelligence - you cannot model the feelings of "affluent urban females 25-35."

2. Purchase Decision Modelling You can simulate the decision-making process of customers whose purchase history, research patterns, and influencer preferences you've tracked. This reveals the cognitive pathways that drive conversions, not just demographic correlations.

3. Content Resonance Forecasting You can test messaging, creative concepts, and brand positioning on simulated versions of customers whose actual brand interactions you've observed. This predicts cultural and personal relevance at individual scale.

The strategic breakthrough: agencies with comprehensive 1PD can now understand why campaigns work before they run them. This transforms marketing from reactive optimisation to predictive behavioural science.

WPP and Omnicom, dependent on probabilistic targeting, remain limited to testing what works after campaigns run. They can optimise delivery but cannot simulate human psychology because they don't own the behavioural data that makes psychological modelling possible.

Publicis should be watching IPG's move very carefully. They have the deterministic data infrastructure through Epsilon and the AI investment through CORE AI. But IPG just demonstrated how to activate this combination for genuine strategic advantage.

The question is whether Publicis will follow this upstream intelligence strategy or continue focusing on downstream optimisation that every agency can eventually replicate.

Epsilon has far more profiles at significantly higher resolution than IPG’s Acxiom. Now imagine if they deployed the technology to have privacy compliant simulated versions of the IDs they have that they could use for research, innovation and content optimisation.

Now that is a new strategic moat.

The strategic opportunity is immediate: Use Epsilon's deterministic profiles to power simulation technology that can predict emotional responses, model purchase decisions, and forecast content resonance before campaign creation. This transforms CORE AI from execution tool to strategic intelligence platform.

IPG moved first, but they don't have an insurmountable advantage. The question is whether Publicis will follow and improve on this upstream intelligence strategy or continue focusing on downstream optimisation that delivers incremental improvements rather than transformational capability.

Brandflow is written by Justin Billingsley, who has spent his career on all three sides of the industry's table: senior client, global agency leader, technology founder. First published 12 August 2025 in the Brandflow newsletter on LinkedIn.