Trillions of signals and billions of people: how systems score goals
Picture the stadium at a World Cup game. Did you envision the 80,000 people, or just a building? What about the billions watching from home? The games come to life thanks to the people, and campaigns similarly find resonant performance amid trillions of data signals calling out like fans chanting and cheering in the stands.
Every single fan is a different signal — different language, different jersey, different reason for being in that seat. The energy they create together is something no single fan could generate alone. And the best teams don't ignore that noise. They are energised by it, adapt to it, and play differently because of it.
That's what modern signal modeling actually looks like. Not one flavor of data. Not a rigid playbook built before kickoff. A living system that reads the room in real time — flexible enough to shift when the momentum shifts, rich enough that no single data point carries the whole weight. The identity isn't the IP address. The crowd isn't one person. The model isn't one signal.
That's the difference between a team that survives the group stage and one that lifts the trophy. Between a marketing team that gets a budget cut in January and one that's collecting accolades and bonus checks in December.
So what does that mean in practicality? How can a marketer reach the right people — whether among the 80,000 in a stadium or among the billions watching online — to score with their campaigns? Let’s dive in.
Read the crowd, not the seat number
For decades, the industry tried to identify fans by where they were sitting. The IP address was the seat number — a single coordinate that told you almost nothing about who was actually in the chair. Was it the season ticket holder? Their kid? A scalped seat sold three times over? The seat doesn't know. And when you build a campaign on the seat, you end up serving the same ad to the season ticket holder and the scalper, then wondering why your conversion math doesn't work.
That's roughly the state of CTV identity today. A recent industry study found that IP-to-email accuracy in connected TV sits at just 16%[1], and IP-to-postal accuracy at 13%[2]. Relying on a limited number of datapoints was never going to be enough. Campaigns must leverage and dynamically adapt to more signals.
Layering the Possibilities
First, let’s explore what’s actually possible in CTV targeting. Instead of focusing on the fallibility of a single match field, marketers finding success are layering far more signals than ever before. For instance, instead of pre-assembling a sport enthusiast audience based on a few static demographic signals — gender, age, income, watches sport, etc — savvy CTV marketers are letting algorithms run with the ball.
Now, models evaluate dozens of signals in parallel for every screen, on every bid, throughout the campaign. This includes dimensions like cross-device behaviour, prior conversion patterns, content context, time of day, mobile engagement history, and so much more. In mere milliseconds, 45 signals layered together produce more than tens of trillions of possible configurations[3] to model against.
Adapting to Win
Of course, the play that worked in one moment may not work the next. Consider the defender who’s a step slow, the conversion pattern that just shifted, the wing that just opened up, or the household that just became in-market.
Just as playbooks cannot replace swiftly reading the field and adapting, so marketers and their campaigns must adjust in real-time. For a human, it’s a lot to demand, and it’s no wonder that, per the IAB, 44% of buyers state that “adapting to evolving customer behaviour” is their top challenge. Prediction, measurement, and adaptation are the state of the game.
For the CTV marketer who has flipped their strategy from static lists to outcomes-based bidding, adaptation is natural. And as these campaigns deliver better performance, expect budgets to follow. Indeed, per the IAB, CTV advertising spend is expected to grow 13.8% this year — while linear faces a mild decline[4].
Why is it working? Modern CTV campaigns get smarter every minute the ball is in play, every moment the campaign is in flight. Every conversion sharpens it. Every miss sharpens it. Before the campaign ends, your model knows things about your customer that your brief couldn't have told you in March.
Considering the Cup
Whoever your target audience is, and whichever signals best predict their conversion likelihood, marketers must seriously consider and account for the 2026 FIFA World Cup in their strategies this summer. FIFA estimated that over six billion people will engage with this year’s tournament[5] — the largest live audience in sport history, fragmented across linear, streaming, broadcaster apps, and phones.
In short, there’s a high likelihood that a significant portion of your audience will be watching. The question is: can you reach them cost effectively? Streaming CPMs could rise sharply as demand concentrates around the tournament.
That combination — enormous audience, fragmented attention, premium pricing — is exactly the environment where a weak signal model gets exposed and an excellent one pulls away. When inventory is cheap, you can afford some waste. At World Cup CPMs, every impression you serve to the wrong household is a measurable hole in your P&L.
This is the moment that separates the brands buying reach from the brands buying outcomes. The first group will spend big and tell a great story in the recap deck. The second group will spend smarter and tell a better story to their CFO.
Winning the Moment
Whether on the pitch or in the #Marketing channel at work, the winning teams during the World Cup will be those that feed off the energy (and signals) available to them.
It will be the halftime ads that drive same-day installs and next-morning purchases. It will be a second-screen retargeting wave that converts the household that paused on your CTV spot. It will be knowing which fans in that 80,000-person stadium — and which of the billions watching online — are actually your customers, and meeting them where they already are.
So, treat your audiences like a living, breathing crowd rather than a static spreadsheet. Consider and leverage dozens of layers of possibilities about each bid. Stadiums aren’t just concrete, and exponential growth is not on linear.
[Editor's note: This is a contributed article from Moloco. Streaming Media accepts vendor bylines based solely on their value to our readers.]
[1] Source: Truthset, State of Data Accuracy 2026. February 2026
[2] Source: Truthset, State of Data Accuracy 2026. February 2026
[3] Source: Basic math. 45 binary signals produce 2^45 possible configurations, which is 31,184,372,088,832 (or 35.184 trillion)
[4] IAB. 2026 Outlook: A Snapshot of U.S. Ad Spend, Opportunities, and Strategies for Growth. Jan 2026.
[5] Source: StackAdapt, Connected TV Statistics Roundup. March 2026. AND
FIFA, 500 days to go: excitement builds for FIFA World Cup 26. Jan 2025. AND
Sports Illustrated, 2026 World Cup: The Most-Watched Sporting Event in History? Dec 2025.
Related Articles
Globo director of adtech Ana Beaklini says latency has traditionally been the biggest problem the Brazil-based Globo has faced in its efforts to provide seamless ad insertion for high-stakes, high-concurrency, ultra-low-latency live events like the World Cup. In this conversation with Streaming Media contributing editor Nadine Krefetz at Streaming Media Connect 2026, Beaklini explains how Globo has addressed this issue and the best practices the company will implement at massive scale for upcoming World Cup live streams and OTA broadcasts and AI's role in hyper-personalising Globo's live ad decisioning.
27 May 2026
The World Cup's return to the U.S. will be defined by moments, whether it's a last-minute goal, a watch party with friends, or a casual fan tuning in for the first time. While the event brings global scale, its impact is deeply personal and local. For marketers, the opportunity isn't just to be part of the event, but to connect with how those moments unfold in people's lives.
06 May 2026
Gaming is built as an always-on system designed to maximise session frequency, retention, and lifetime engagement. Live sport, on the other hand, is still optimised for scheduled viewing windows where engagement peaks during the match and drops off immediately once it's over. For streaming product leaders and sport media executives responsible for retention and average revenue per user (ARPU), the opportunity isn't to change the live event. It's to capture the audience before and after it.
16 Apr 2026