Peaks of Mentions – a surge in brand mentions that significantly exceeds the baseline average within a specific timeframe (statistically defined as exceeding one standard deviation).

Expression – the verbal articulation of sentiment toward a brand, whether positive or negative.

Intensity of Expression – the degree to which these peaks of mentions are saturated with emotionally charged statements, irrespective of their positive or negative valence.

Informational Search Queries – user searches aimed at gathering more details about a brand. This includes queries about the specific model or product, its price, specifications, and reviews.

Transactional Search Queries – searches specifically intended to find points of sale, such as physical store locations, official websites, or booking platforms.

Spontaneous Purchases – non-essential, low-cost items bought without a practical necessity. These purchases carry minimal social, psychological, or health risks. Examples include indulgent snacks and drinks, small accessories, free-to-download apps, and lower-priced cosmetics.

Routine Purchases – essential goods with a low to medium price point that require regular replenishment. They carry moderate health and social-psychological risks. Examples include groceries, clothing, monthly subscriptions, and skincare products.

Careful Purchases – significant, periodic (but not frequent) purchases with a high price tag. These come with considerable financial and social-psychological risks, often tied to lifestyle choices. Examples include electronics, hotel stays, airline tickets, and concert tickets.

Risky Purchases – major, typically one-time purchases with a distinctly high price and critically high financial and social-psychological stakes. These are characterized by a high cost of replacement or are often irreversible. Examples include real estate, automobiles, medical procedures, and education.

In our previous article, we laid the groundwork for this research initiative. We detailed the core questions driving our study, outlined our robust dataset and methodology, and presented our initial hypotheses regarding how spikes in social media brand mentions influence consumer purchasing behavior.

Our analysis examined over two million mentions of 39 global brands across a two-year period, from December 2023 to November 2025.

Due to the monthly aggregation of search query data by analytics services, we adopted the month as our standard unit of time (t).

Brands were selected for the study based on the occurrence of what we term peaks of mentions—or viral spikes—during this period. This selection was statistically grounded in the analysis of standard deviations.

Given that approximately 68% of observations fall within one standard deviation of the mean, we defined our peaks as any value falling outside this range, representing a significant deviation from normal brand conversation volume.

On average, each brand in our study experienced two such peaks over the 24-month timeframe.

Results of the first study:

Our initial analysis yielded a compelling insight into the role of emotional expression. We found that peaks characterized by moderate to high expressiveness—where emotional content constitutes between 10% and 39% of all mentions—have a significant and sustained impact on subsequent search behavior. This influence persists for at least one month following the initial media spike.

Conversely, the effect of non-expressive peaks—those where a surge in volume is not accompanied by any significant emotional sentiment—on user search activity is weak. Equally marginal is the impact of what we term extreme expression, where overtly evaluative or emotional statements make up more than 40% of the total mentions.

Hypotheses for the Second Phase of Research:

A rise in brand mentions affects different types of search queries—informational and transactional—in different ways. The first type of query arises from a need among users to gather information about a brand when that information is not sufficiently covered in the peak mass of mentions. Certain product categories require a deliberate purchase decision and necessitate the formation of a compelling brand attitude, one that is balanced between its emotional and cognitive components.

It’s anticipated that informational and transactional search queries will exhibit different trajectories following a media peak, both in terms of their rate of growth and the timing of their peak activity. Furthermore, in theory this relationship is moderated by product category. Specifically, the more complex and inherently risky the purchase, the more independent and protracted the consumer's information-gathering phase will be. However, this extended deliberation is expected to yield a more durable and sustainable impact on long-term brand equity and downstream marketing effectiveness.

Preparing for the Second Phase of Research:

The initial step of this phase involved a meticulous separation of search queries into two distinct categories: informational and transactional.

Informational search queries were defined as those where users expressed a general interest in the brand or its products. This includes top-of-funnel research such as "what is SKIMS," "SKIMS products," and "SKIMS reviews."

Conversely, transactional queries were classified as all searches with direct purchase intent, guiding users toward a point of sale. Examples include "buy SKIMS," "SKIMS official store," and "SKIMS Amazon."

To ensure our transactional category accurately reflected the diverse business models of the brands in our study, we manually augmented the data set. This involved incorporating unique, platform-specific action verbs, such as "download Spotify," "install Spotify," "book Jet2 flights," and "Jet2 tickets." Furthermore, given the presence of major U.S.-based brands, we included queries containing "Amazon" and "Ebay." These platforms remain dominant forces in e-commerce, and their role as direct sales channels for many brands makes their associated search traffic a vital component of any transactional analysis.

With these two query types defined, we proceeded to aggregate search volume data for each of the 85 identified brand mention peaks. This data was then normalized using the same methodological framework established in our first article. This consistency allowed us to rigorously test our hypotheses by reintroducing the variable of expression intensity as a key moderating factor.

In addition to query type, we introduced a second layer of analysis by classifying products and services into four distinct categories. These categories are based on the underlying psychology of the purchase decision, specifically the interplay of price, frequency, and perceived social, psychological, and financial risk:

• Spontaneous (32 peaks analyzed)
• Routine (21 peaks analyzed)
• Careful (31 peaks analyzed)
• Risky (no such peaks in our dataset)

Spontaneous purchases characterized by a low price point and no practical necessity. These purchases carry minimal social and psychological risk. Common examples include indulgent food and beverages, small accessories, free app downloads, and lower-priced cosmetics. The underlying consumer logic is often, "I'll only know if I like it by trying it myself."

Routine purchases may also be relatively low-cost, but they are distinguished by the need for frequent repurchase to replenish supplies. They carry a slightly higher social and psychological risk than spontaneous purchases. This category typically includes groceries, clothing, monthly subscriptions, and skincare products. The consumer mindset here is more utilitarian, "I'll see it for myself when I actually need it."

Careful purchases are most often characterized by a medium-to-high price point. They are made infrequently—perhaps once every few months or annually—and carry a high degree of social and psychological risk. While the financial outlay is significant, the cost of replacement, though notable, is not prohibitive.

This category encompasses electronics, hotel stays, airline tickets, and concert tickets. The consumer is deliberative, operating on the principle, "I'll only try this for myself if I'm confident it's good and that I truly need it."

Risky purchases defined by a distinctly high price and a very low frequency of purchase (annually or even less often). These decisions involve critically high social, psychological, and financial risk.

They are often characterized by a high cost of replacement or, in many cases, irreversibility. We classified real estate, automobiles, medical procedures, and education into this category. The consumer's stance is one of extreme caution, "I will only try this for myself if it is absolutely necessary and promises the highest possible expected utility."

Results 1. The Pivotal Role of Informational Search Queries in Brand Success

By analyzing the various intersections and combinations of our three data segmentation methods an overarching conclusion came into focus.

The highest effectiveness—the strongest conversion from media exposure to transaction—is achieved when a two-stage process occurs. First, users are immersed in a wave of emotionally charged brand content. Then, they proactively supplement this affective experience by conducting their own informational searches to gather the necessary utilitarian details.

However, a critical temporal dynamic is at play in these most successful cases. The surge in transactional searches—queries for "order," "buy," "install," or "store near me"—is significantly more pronounced in the month following the peak media visibility, rather than in the month concurrent with it.

We first introduced this proposition in our initial article: that non-expressive content is typically informational and self-sufficient, effectively preempting the user's need for additional research. The data from this phase supports that claim. The volume of informational search resulting from peaks of non-expressive media actually charts a negative trajectory, confirming that these spikes satisfy curiosity in the moment without fueling further inquiry.

At first, the data might suggest that transactional search volumes generally trend positive following any brand mention peak. However, when compared to the impact of more expressively saturated peaks, the effect of non-expressive content is marginal. This represents the outcome of a purely rational appeal. While such an approach does generate some transactional interest, its potential is significantly limited by the absence of an emotional component to engage the consumer.

Examining the three defined levels of expression intensity reveals a more nuanced picture.

All three levels demonstrate a broadly similar pattern for both search types, with an immediate reaction occurring at time t0 for both informational and transactional queries, followed by a gradual decline in subsequent months. However, the moderate intensity level distinguishes itself from the others. Its decline from t0 to t+1 is noticeably less steep, exhibiting a trajectory that begins to mirror the behavior of the two higher intensity levels. This subtle difference brings us closer to understanding the conditions necessary for generating the long-term media effects that build sustainable brand awareness and equity.

When a peak of brand mentions contains a highly or extremely proportion of emotional expression (20% and above), the increase in transactional queries is substantial but occurs with a one-month delay. Extremely emotional expression exhausts demand over the second and third months. Highly emotional expression, by contrast, satisfies the same demand more quickly because it contains a sufficient share of factual information about the product, thereby laying the cognitive groundwork for the purchase decision.

Both scenarios are preceded by a rise in informational queries in the same month as the peak of mentions.

In that first month, the business—if measuring sales—has no way of knowing that the foundation for its future success is being laid. Notably, an extremely high share of emotional expressions in a peak compels people to gather three times more information through search than when the share is simply high (20 to 40%).

Results 2. Brand Categories Most Responsive to Long-Term Effects

The next key finding to emerge from our analysis can be stated as follows:
The effectiveness of media exposure—measured by its conversion into both informational searches and eventual transactions—increases in direct proportion to the perceived importance, cost, and risk associated with the purchase.

In the previous section, we identified that the high expression intensity level (where evaluative statements constitute 20-39% of total mentions) is the most effective. Here, we compare how media peaks with this optimal level of expression perform across product categories with varying degrees of perceived risk. (As noted, the fourth category, Risky Purchases, was not represented in our dataset.)

The spontaneous category demonstrates not just low conversion into search queries, but an actual decline relative to the norm. We have not yet analyzed media peaks in terms of their positive-to-negative sentiment ratio. At this point, we can only hypothesize that people primarily encountered negative information about spontaneous products—and that this is why they lost interest in them. Confirmation or refutation of this hypothesis will appear in a future article.

The routine category shows a small immediate increase in informational queries, followed by a sharp decline in the following month—falling into negative territory. The careful category, by contrast, shows a fairly high immediate increase in informational queries, followed by a gradual decline that remains positive even into the second month.

At the same time, transactional queries for these two categories, while differing in volume, follow the same dynamic: a threefold increase in the first month after the media peak, followed by a halving in the second month. Despite this similarity, the clear "winner" in this effectiveness race is the more complex and higher-risk category. Media peaks with a high share of emotional expression establish a sustained and elevated trend line for both informational engagement and purchasing activity.

To better understand the optimal approach for the spontaneous category, we continued our analysis and identified a more suitable strategy.

By distinguishing between two different types of search queries, we found that media exposure achieves its highest conversion rate into transactional queries when users first receive maximum emotional impact from it, and then supplement with utilitarian information through informational searches.

It was also established that the conversion of media exposure into both informational search and transactions increases as the importance, cost, and perceived risks of the purchase grow. The highest effectiveness for highly expression-saturated media peaks was identified for categories such as significant periodic (but not frequent) purchases with high prices and substantial financial and socio-psychological risks, often tied to lifestyle choices: electronics, hotels, airline tickets, concert tickets.

In forthcoming articles, we plan to refine our analysis by introducing the critical variable of sentiment, examining how different balances of positive and negative expression within a media peak affect outcomes. In addition, we will augment our dataset with brand valuation metrics to directly correlate these media dynamics with long-term shifts in brand equity.