Every marketing professional can recall facts or myths about viral social media events supposedly making or breaking a business. However, it's often unclear whether this conventional wisdom is artificially crafted by PR campaigns or simply stems from speculation based on incomplete data. The field lacks rigorous, objective studies—not commissioned to serve a specific brand's interests—that attempt to draw reliable, generalized conclusions. For marketers and brand reputation specialists, understanding the precise patterns, magnitude, and nature of media events' impact on target audience search and purchasing behavior would be invaluable for shaping corporate communications. Furthermore, there is a significant shortage of global, cross-cultural research to help brands adapt successful strategies to new markets while accounting for local nuances. These considerations formed the foundation for our Institute's large-scale, multi-category study.

Surges (or peaks) in brand mentions should correlate with increased demand for its products or services. This could range from a rise in search query frequency (as the most readily available metric) to, ideally, an increase in actual consumer demand—the latter being provable only with access to open-source market data.

We analyzed over 2 million mentions of 39 global brands across a 24-month period (December 2023 – November 2025).

Search query statistics are provided by services on a monthly basis; therefore, a single month was selected as the unit of time (t).

Brands were selected based on the presence of significant mention peaks (or viral spikes) during the specified period. This selection itself relied on standard deviation statistics.

Approximately 68% of observations fall within ±1 standard deviation; therefore, values outside this interval were considered deviations from normal behavior.

On average, each brand exhibited about 2 such peaks.

Examples of peaks:

1. Sydney Sweeney’s American Eagle jeans ad: July 2025
2. Jet2 Holiday viral sound: July 2025
3. SKIMS’ Bush Hair Thong: Oct 2025
4. Kylie Cosmetics’ King Kylie revival: Oct 2025
5. Cracker Barrel’s rebranding backlash: Aug 2025
6. Pop Mart’s Labubu Craze: April 2024

Examples of correlations:

This correlation may manifest either at time t0 (almost simultaneously) or at time t+1 (with a certain delay).

1. Initial data collection was conducted across major social networks and user-generated content platforms: TikTok, YouTube, Instagram, and Medium.
2. Identification of articles, videos, and other content summarizing the most significant viral moments and scandals (e.g., brands that went viral in 2024/2025, controversial brands of 2024/2025, brand scandals in 2025).
3. Collection of search query frequency data over the past two years (December 2023 – November 2025) using system of monitoring.
4. Gathering more extensive social media mention data using the system of monitoring tool for the same two-year period (December 1, 2023 – November 30, 2025). This stage focused on identifying mention peaks for an expanded list of brands.

All brands in the study were categorized using a traditional framework—grouped into broad categories of goods, service sectors, and consumer experience domains.

This article presents the first significant finding of our study. This finding concerns the emotional intensity of media peaks. We believe this result is substantial enough to merit its own dedicated article as the first in our series.

It became apparent almost immediately that not every mentioned peak is equally effective in influencing search and purchasing activity. Mentions can range from expressing various emotions to containing nothing more than standard product descriptions, often formulated by the company itself.

The first key discovery: 96.5% of the 273 million brand mentions collected over the two-year period were neutral in tone. Only approximately 10 million of them expressed either positive or negative sentiment. Consequently, each identified peak contains a varying mix of both neutral and emotionally expressive mentions.

Thus, the study's first task was to classify the peaks based on the proportion of expressive content within each one. After all, emotions are the driving force behind people reconsidering their stance toward brands.

We proceeded by calculating, for each peak, the proportion of emotionally expressive mentions, which are divided into two subgroups:

• Positive Expression: Encompasses commendable statements, enthusiastic reactions, and favorable discussions about the brand.
• Negative Expression: Typically includes criticism of the brand, negative reviews, and discussions of scandals or controversial situations.

The chart illustrates the distribution of these mentions across social networks from November 2023 to December 2025 (based on a sample of 39 brands).

Based on the obtained data, we identified six levels of expressive intensity for brand mention peaks, ranging from completely non-emotional to highly emotional. This detailed segmentation allows us to uncover significantly different response patterns that would remain hidden with a simpler grouping (e.g., high, medium, low).

1. Non-expressive: 0% share of expressive mentions.
2. Minimal expressiveness: 1% – 3%
3. Low expressiveness: 4% – 9%
4. Moderate expressiveness: 10% – 19%
5. High expressiveness: 20% – 39%
6. Extreme expressiveness: Over 40%

For readers interested in the data distribution across these levels, we note that in half of all cases, peaks contain emotionally charged mentions at a level below 5.63%. For the majority of peaks, the share of emotional intensity fluctuates between approximately 2% and 12%.

It is important to note that we searched for correlations while accounting for a potential time lag between the independent variable—a peak of brand mention—and the dependent variable, namely consumer behavior. Based on such correlations, one can make reasonably probable inferences about causal relationships between these two phenomena. However, to be certain of their existence, one must proceed further with field experiments where brands deliberately stimulate a mention peak (e.g., via UGC campaigns) and then conduct measurements, controlling for all other marketing activities and tracking actual sales.

The first and, in our view, most significant result was that no substantial link was found between a media peak and an increase in search queries in the category of non-expressive peaks—when a significant surge in content lacks any pronounced emotional expression.

Conversely, a correlation between increased search queries and expressive mention peaks was confirmed, with a typical delay of one to two months following the peak.

While this fact may seem intuitively obvious, it cannot be definitively explained. After all, as we all know, not all search queries are informational; some are transactional and lead directly to online stores. We can propose the following two explanations:

1. The absence of emotional expression has an ambiguous effect on conversion into purchasing activity: it either has no impact or acts randomly.
2. We might be observing a shortened customer journey that bypasses search engines altogether, though this is more likely to apply to existing customers.

To be confident in either explanatory hypothesis, data on sales dynamics or market share in each specific case is required.

The second, less obvious result was the absence of a correlation between search activity and the most extreme peaks in terms of their expressive component. In other words, moderate expressiveness influences search, and its effect persists for at least a month afterward. However, the influence of extreme expressiveness was not proven in our study; it appears inconsistent and unreliable.

Here, too, we can only speculate about the reasons for this disparity:

1. Individuals, who themselves mostly adhere to social norms and are not inclined to express intense positive or negative emotions about brands in public, may distrust extreme expressions. For them, such content is a sign of deviation, and it does not alter their purchasing behavior.
2. Excessively emotional discussions may prioritize social validation and consensus over factual analysis, shifting attention from search engines to the debates among participants. Such content might also be perceived as media consumption or entertainment with its own narrative drama, rather than as a prompt for a purchase decision.

How can our findings benefit brands?

First, measuring a brand's media weight in isolation from its emotional dimension is insufficient. Brands must keep a pulse on the emotional tone of their informational field. As researchers, we are unequivocally confident in this conclusion.

Second, it is advisable to incorporate and carefully manage the emotional component when planning and executing social media mention campaigns. More precisely, brands should actively manage the proportion of this component, ensuring it is neither "overheated" nor completely devoid of expressiveness.

The next step of our hypothesis involves moving beyond the intensity of emotional expression to analyze the valence—positive, negative, and potentially mixed signals—and their distinct impacts.

We have already observed preliminary patterns:

• Positive social media mentions correlate with a moderately higher frequency of search queries in the same month and the following month; some are followed by an additional peak (often a result of sequential brand PR activities).
• Negative social media mentions correlate with sharp spikes in search query frequency in the same month, followed by a somewhat lower, yet sustained, level in the subsequent month.

To deepen this analysis, future work may involve classifying search queries as informational vs. transactional, and investigating whether search peaks for a brand's competitors follow critical attacks on the brand.

A third research direction will explore differences in media dependency across product categories.

Preliminary data indicates that utilitarian brands show a strong correlation between a sharp increase in mentions and a simultaneous sharp spike in search query frequency at times t0 and t+1, which then completely subsides by t+2.

Emotional brands exhibit a less dramatic initial spike, but their search query frequency typically does not drop as sharply by t+2 and sometimes even demonstrates a new peak.

We believe the ideal culmination of this work would be the development of a Media Dependency Index for product and service categories, providing a nuanced tool for strategic planning.