{{title}}
{{textarea}}
{{task}}
{{select}}
{{team}}
{{input}}
{{solution}}
{{result}}
ICDS Research Brief Media Peaks & Consumer Behaviour
Moneyball for Marketing: Why Media Activity Is an Equation, Not an Art Form
Marketing directors calculate, budget, and forecast — so why does “did the campaign work?” still feel like a matter of opinion? Because the standard measurement stack is answering the wrong question, at the wrong moment.
VM
Vadim Matyushkin Behavioral Scientist & Sociologist · ICDS
In 2002, the Oakland Athletics had one of the lowest payrolls in Major League Baseball. They could not compete for star players. So general manager Billy Beane did something scouts called heresy: he stopped trusting intuition and started trusting equations. He identified which statistics actually predicted winning — not the ones that sounded impressive, but the ones that correlated with outcomes. The Athletics won 20 consecutive games that season. The approach became a book, then a film, then the standard model for evidence-based decision-making across industries.
Marketing is still in the pre-Moneyball era.
Not because marketers lack data — they have more data than ever. But because the metrics that dominate dashboards measure the wrong things at the wrong time. A campaign generates a media peak. The peak has volume, reach, sentiment. The report is written. The budget decision for next quarter is made. And the fundamental question — did this campaign actually move consumer behaviour toward a purchase? — remains, at best, answered by inference.
Three years of research across 66 FMCG brands and 11 fintech companies has produced a different answer. Media activity around a brand is not art. It is a process with a measurable structure — before the peak, inside the peak, and after it. Each phase contains data that standard monitoring systematically ignores. Each phase contains a lever that marketing directors can actually pull.
92%Peaks have a paid foundationSustained digital ads in the 12 weeks before the peak
20–41%Of a peak is noiseInertial content that masks real sentiment by 34–56%
4–8 wksPost-peak windowIn 72% of Cautious brands, intent peaks after the event
22 of 24Ads precede the peakPaid activity leads the organic peak in non-crisis events
Three Questions Your Dashboard Cannot Answer
Every marketing director has encountered at least one of these situations. They are not edge cases. They are structural blind spots in the standard measurement model.
Question 01
Two campaigns, identical sentiment scores, completely different sales outcomes. Why?
You ran two major campaigns in the past year. Both generated comparable media peaks. Both showed similar positive sentiment in your monitoring platform. One drove a measurable lift in sales. The other did not. Your analytics team cannot explain the difference. The creative director says it was the creative. The media planner says it was the placement. Nobody looks at what was inside the peaks.
The finding
Between 20% and 41% of every media peak is inertial content — reviews, discussions, reposts from the preceding month that have nothing to do with your campaign. This inertial layer carries last month’s sentiment and systematically dilutes the real audience reaction to what you actually did. Two peaks with identical observed sentiment can have radically different reactive signals underneath. One campaign genuinely moved your audience. The other mostly moved old content.
Question 02
The campaign ended three weeks ago. Search traffic is flat. Did it fail — or are we looking too early?
Your media peak has passed. Monitoring shows mentions returning to baseline. The paid media has stopped. You are checking Google Analytics and search volumes, and nothing has moved. The conclusion forming in the room is that the campaign underperformed. The brief for the next campaign is already being revised.
The finding
In Cautious purchase categories — premium cosmetics, technology, financial products, baby goods — the consumer does not search to buy at the moment of the media peak. They search to learn. Transactional intent (searches like ‘[brand] buy’, ‘[brand] price’) emerges four to eight weeks after the peak, in 72% of studied brands. If you measure conversion the week the campaign ends, you are measuring the wrong moment.
Question 03
A competitor with the same budget keeps generating viral peaks. We do not. What are they doing differently?
You watch a competitor’s brand generate organic social media peaks regularly — user content, influencer amplification, genuine cultural moments. Your campaigns generate reach, but not that kind of spontaneous response. The assumption in your organisation is that they have better creative, a better agency, or better luck. You increase the creative brief. Nothing changes.
The finding
Organic peaks do not arise on bare ground. Across 27 FMCG brands, 92% of non-crisis media peaks were preceded by sustained digital paid activity in the 12 weeks before the peak. Even purely viral UGC peaks — Heinz, Pringles — had active paid campaigns running in the background. Paid media does not buy organic peaks. It builds the awareness and cultural presence that makes organic peaks possible. Without the foundation, the same event that generates a peak for your competitor generates nothing for you.
The Equation: Three Phases, Three Measurements
The Moneyball insight was not that baseball could be understood with data. It was that the data everyone was using was measuring the wrong things. The same insight applies here. The problem is not that media activity is unmeasurable. It is that the standard measurement model captures only the middle of a three-phase process.
Weeks −12 to −1 · Before
Paid Foundation
92% of organic peaks are preceded by sustained digital advertising. This is not correlation — it is the mechanism. Continuous paid presence builds the audience awareness that makes a media event land. Without it, the event produces a smaller peak, weaker search response, and lower conversion probability.
Measurement tool: Google Ads Transparency
Month t0 · During
The Layered Cake
20–41% of peak volume is inertial — old content carrying last month’s sentiment. Observed sentiment ≠ real audience reaction. Reactive sentiment (the actual response to what you did) is systematically more intense than the aggregate. For crisis events, the real negativity is up to 56% stronger than the dashboard shows.
Measurement tool: reactive sentiment formula
Weeks +4 to +8 · After
The Purchase Window
In Cautious categories, purchase intent builds after the peak, not during it. Informational search ([brand] reviews, [brand] vs) rises at t0. Transactional search ([brand] buy, [brand] price) rises at t+1. The window is four to eight weeks. Brands that prepare conversion content before the campaign capture it; those that react after the peak miss it.
Measurement tool: Google Trends query intent split
What This Changes for How You Run Campaigns
Billy Beane did not replace scouts with spreadsheets. He changed which questions scouts were asked to answer. The same reorientation applies here.
1
Audit your paid foundation before measuring campaign results.
Before attributing a campaign’s success or failure to creative, messaging, or timing — check the 12-week window before it. Was there sustained digital advertising activity? If not, the baseline awareness required for organic amplification may simply not have existed. This is not a creative problem. It is an infrastructure problem — fixable with a different investment pattern rather than a different agency.
2
Read reactive sentiment, not observed sentiment.
The sentiment number your platform reports includes inertial content from the previous month. For a brand with high continuous paid activity — Coca-Cola, L’Oréal, Pampers, Red Bull — this inertial layer is systematically large, and it softens the real reaction to events. During a crisis, your dashboard may show moderate negativity while your actual audience reaction is severe. Reactive sentiment — observed sentiment corrected for inertial weight — is the number that correlates with market behaviour.
3
Prepare conversion content before the campaign, not after the peak.
In Cautious categories, the post-peak window is when your potential customer is actively looking for reasons to buy — or not to. Comparison content, detailed feature explanations, and review aggregations need to be indexed and visible before your media peak happens. If you create them in response to an unexpected search surge, you are two to three weeks late to a four-to-eight week window. The campaign created the demand; your content infrastructure should have been waiting for it.
The Moneyball analogy, precisely
Beane’s insight was not that some players were undervalued. It was that the metrics used to value them — batting average, RBI, stolen bases — were measuring the wrong things. On-base percentage predicted wins, and nobody was tracking it seriously. The equivalent here: reach and aggregate sentiment are batting average. Paid foundation coverage, reactive sentiment, and post-peak informational search are on-base percentage. They predict outcomes. They are available. Nobody is tracking them seriously.
Research basis
What this brief synthesises
This brief draws on three ICDS studies. (1) The Paid Foundation of an Organic Peak — 27 FMCG brands, Google Ads Transparency API, 12-week pre-peak observation window. (2) The Layered Cake of a Media Peak — 11 fintech peaks, reactive sentiment decomposition, 4 public companies, OHLC data. (3) Beyond Search: Social Media Presence, Consumer Attitudes, and Actual Purchase Behaviour — 71 brands, 6 markets, YouGov Consideration Score, Kantar Brand Footprint, Google Trends. Social media monitoring data: commercial platform.
Full research articles at icds.institute.
VM
About the author
Vadim Matyushkin
Behavioral Scientist & Sociologist · ICDS
Vadim Matyushkin is a psychologist and sociologist with close to twenty years of research into digital behavior, trust, and information dynamics, and a researcher behind the Institute of Communication and Data Science (ICDS). His work has supported organizations including Coca-Cola, PepsiCo, Mars, Danone, Nestlé, and Bayer in moving from self-reported survey data toward direct behavioral evidence.
ICDS — Institute of Communication and Data Science is an independent research institute focused on understanding how trust, behavior, and reputation are formed in digital environments. ICDS operates as an intellectual institution — not an agency, not a platform, and not an educational provider.