This study was developed within the CSV (Cultural & Semantic Value analysis) framework — the ICDS proprietary methodology that reconceptualises conventional media monitoring. Rather than tracking aggregate mention volumes, the framework shifts analytical attention to the internal structure of a media event: its sentiment profile, trigger type, content genre and the dynamic trajectory before and after the peak.

Previous studies in this series established that it is not the volume of mentions but their emotionally evaluative component that serves as the primary mediator between a media event and a change in consumer behaviour. The present paper extends that finding by posing a different question: to what extent is the character of media events itself determined by a brand's product category and country of origin? In other words — is the structure of a media peak universal, or is it shaped by categorical and cultural conditions?

The unit of analysis is a media peak — a month in which a brand's mention volume exceeded the monthly baseline by more than one standard deviation (≥25% deviation from the mean). This operationalisation is consistent with the definition adopted across the ICDS research series and ensures comparability across brands with substantially different baseline media footprints.

For each observation, the following variable blocks were recorded:

The sample was constructed through purposive sampling: brands were included if they satisfied two criteria — the presence of a documented media peak during 2023–2024 and sufficient public visibility of the event to allow trigger verification through open sources. The final sample comprised 66 observations across 66 brands in 26 countries.

The distribution across product categories broadly mirrors the structure of the global FMCG market: the largest sub-sample falls in Food & Beverages (12 observations), the smallest in Baby & Child Products (2) and Confectionery (3).

The ten product categories were aggregated into three analytical clusters on the basis of consumer psychology and the social function of the product:

The initial sample of 50 brands contained a material country-group imbalance: 63% of observations in the hedonic cluster came from developed markets (five brands from the United States alone), while 60% of the functional cluster comprised emerging-market brands. Left uncorrected, this disparity risked conflating category effects with the concentration of large global brands with disproportionate media budgets.

To remove the confounder, 16 additional brands were purposively added, prioritising under-represented country groups within each cluster. Following this rebalancing, a 50/50 parity was achieved across all three clusters, with the full sample reaching an exact equilibrium of 33 developed-market brands and 33 emerging-market brands.

Between-cluster differences were tested using the Kruskal–Wallis test (three-group comparisons); pairwise comparisons between country groups employed the Mann–Whitney U test. Significance thresholds: p<0.05 — statistically significant finding; p<0.10 — consistent tendency, requiring a qualifying caveat. Results with p>0.10 were not interpreted as patterns.

Three key limitations should be acknowledged. First, the sample is based on purposive rather than random selection; generalisation to the full population of FMCG brands requires caution. Second, the Identity & Lifestyle cluster contains 13 observations — sufficient for statistical analysis but at the lower bound; findings for this cluster should be treated as preliminary. Third, sentiment and volume metrics were derived from automated NLP monitoring systems, which carry a classification error of approximately 5–15% depending on language and platform.

The central question of this section is whether a brand's product category determines the character of its media events. By "character" we mean the set of measurable parameters: the dominant trigger type, the prevailing content genre, the sentiment profile of the peak, and the structure of its media dynamics.

The analytical starting point is the null hypothesis: a media peak is structurally uniform regardless of category, and observable differences are explained by the country-group and budget composition of the sample. Following the balancing procedure described in Section 1.5, this hypothesis was tested statistically.

The sole metric for which inter-cluster differences reach statistical significance (Kruskal–Wallis H=7.33, p=0.026) is the share of emotionally evaluative mentions within the total volume of the media peak.

The gap between hedonic and functional categories stands at 6.2 percentage points — moderate in magnitude but reproducible. It persists after sample balancing and is not explained by country-group composition.

Explanatory model. Alcohol, soft drinks, snacks and confectionery are products embedded in social rituals — shared meals, celebrations, sporting events. Mentions of these products in the media carry a consistently higher emotional charge not because brands invest more in emotional content, but because the consumption itself is public and socially salient. Functional products — cleaning agents, dairy, staple foods — lack this ritual dimension: they are consumed privately, and their media discussion defaults to informational rather than affective register.

The difference in negative-mention share does not reach the p<0.05 threshold, but the Mann–Whitney test registers a consistent tendency: hedonic categories show a structurally lower share of negative mentions (median 16.8%) compared with the identity cluster (25.0%) and functional cluster (24.1%), p≈0.10.

This signal weakened after sample balancing, but the direction was preserved. The interpretation is cautious: hedonic categories may be structurally less exposed to reputational negativity, but this tendency requires confirmation on a larger sample.

The clearest and most robust difference between clusters was found not in quantitative sentiment or amplitude metrics, but in the type of dominant trigger — that is, in the very nature of the events that generate a media peak:

Each cluster operates through a fundamentally different dominant mechanism for entering the media landscape. Hedonic brands predominantly borrow media energy from external events through sponsorship — major sports tournaments, festivals and national holidays. Identity-cluster brands rely on paid influencer collaborations as their primary peak-generation tool. Functional brands are more varied in their mechanisms, but are distinguished by two polar types: on one side, planned marketing campaigns and social initiatives; on the other, crisis-driven peaks that are unique to this cluster.

The second robust structural divergence concerns the prevailing content genre that constitutes the peak:

Hedonic and beauty categories operate almost exclusively in a video-centric media environment. Functional categories operate in a mixed environment with a high proportion of text content (45% combined). This is not an artefact of individual brand marketing strategies — it reflects the broader structure of media consumption: food, household products and dairy regularly become subjects of consumer journalism, regulatory reviews and healthcare publications. The media field around these brands is shaped by forces well beyond the brands themselves.

Scandal and crisis events — product recalls, litigation, regulatory action — are concentrated almost exclusively in the functional cluster (14% of observations, versus 4% in hedonic and 0% in beauty/personal care). Within the cluster the concentration is even higher: 25% of food-category peaks and 50% of baby-products peaks are crisis-driven.

The trust-contract model. Consumers enter into an implicit safety contract with functional brands: basic foods, baby products and hygiene goods are perceived as categories where risk is by definition unacceptable. Any signal of a breach — even a perceived rather than actual one — triggers a media response disproportionate to the actual scale of the event. The Maggi (India), Johnson & Johnson and Nestlé cases in our sample illustrate this pattern: each crisis peak was accompanied by a surge in earned and UGC content, a predominance of news-format coverage, and a sharp spike–sharp drop profile.

The "sharp rise — sharp fall" (V-shaped) dynamic profile appears in 31% of functional-cluster observations, versus 21% in the hedonic cluster. This directly reflects the crisis nature of a proportion of functional peaks: a crisis event ignites rapidly and is equally rapidly displaced by new agenda items. Hedonic brands more frequently display an asymmetric profile with a gradual decline — a consequence of riding the fading media energy of a host event.

The list of metrics that showed no significant inter-cluster differences after sample balancing is itself analytically important: peak absolute amplitude, post-peak decline rate, next-month sentiment rebound, and the overall sentiment index all proved categorically universal. In other words, all metrics describing the intensity of a media event are category-agnostic. Category determines the nature and genre of an event — not its strength.

The previous section established that product category shapes the nature of a media event but not its quantitative characteristics. A further question follows: does a brand's country of origin make an independent contribution to the structure of media events — over and above the category factor? To address this, the analysis was conducted at two levels of aggregation: (1) a binary developed/emerging-market split; and (2) six geographic regions.

At the binary level, two statistically significant findings were identified.

Brands from developed markets receive a demonstrably higher sentiment index (median 0.34 vs. 0.26, Mann–Whitney p=0.060) and a higher share of positive mentions (median 55.3% vs. 46.9%, p=0.050). Peak amplitude, mention volume, post-peak decline rate and sentiment rebound do not differ between the two groups.

The sentiment gap between groups is substantive rather than technical: it is not explained by differences in cluster composition (post-balancing) and is reproduced across both metrics — the direct index and the positive share.

Two interpretations merit consideration. The first is that brands from developed markets carry higher global brand equity, which translates into a more positive media backdrop regardless of the specific event. The second — methodologically significant — is that norms of public emotional expression differ across markets. In a number of emerging markets, critical online expression about brands is more prevalent, which may systematically depress sentiment index scores without reflecting any deterioration in actual brand perception. Disentangling these interpretations requires a Cultural Expressiveness Baseline (CEB) calibration procedure, as described in the PRI methodology.

Moving to six geographic regions sharpens the sentiment differences (H=9.60, p=0.088) and reveals a concrete structure.

East Asia (South Korea, Singapore) displays a median sentiment index of 0.10 — significantly below both the Anglophone West (p=0.033) and Continental Europe (p=0.028). At the same time, mention volumes and peak amplitudes in this region are standard — the media environment is active, but tonally restrained.

This result is consistent with the Cultural Expressiveness Baseline concept in the PRI methodology: in societies with strong norms of public restraint in evaluative expression, the sentiment metric reflects a culturally conditioned style of online communication rather than actual brand perception. Korean and Singaporean consumers may not hold less positive views of brands — they simply express those views differently in digital public space.

The distribution of content genre across regions reveals a clear gradient: the share of text-based content (articles + news) increases consistently from Asian to Western markets.

This gradient reflects not differences in brand marketing preferences but the structural characteristics of regional media ecosystems. Western markets — particularly the US and Europe — have well-developed consumer journalism in which FMCG brands regularly attract editorial attention: product reviews, investigative reporting, ESG ratings. Asian markets are dominated by short-video platforms (TikTok, YouTube Shorts, Reels), and brand media image is formed predominantly through visual content.

The share of peaks initiated by brand social initiatives (ESG campaigns, advocacy statements, environmental programmes) varies substantially by region: Anglophone West — 25%; Continental Europe — 24%; South Asia — 7%; Latin America — 14%.

This disparity is not a reflection of differing corporate priorities: many global brands execute a uniform ESG strategy across all markets. However, the media response to social initiatives differs significantly by region. In Western markets, consumer audiences have developed a stable appetite for brands taking public positions on social issues, and a social initiative reliably generates a media peak. In emerging markets, this mechanism is considerably weaker.

In South and South-East Asian markets, 27% of media peaks have an exclusively paid source — meaning the peak is generated by paid content with no organic component. In Western markets this configuration does not appear in pure form: all paid activity is embedded within mixed owned-plus-paid structures. This signals a structural difference in audience trust in organic content: in Asian markets, paid influencers are treated as a sufficient media catalyst, while Western audiences appear to expect a combination of paid and editorial components.

Consistent with the category-level findings, several metrics prove to be cross-regional constants. Peak amplitude, post-peak decline rate and sentiment rebound show no significant regional differences. A media peak as a physical phenomenon — a surge and its dissipation — operates identically across all markets. What differs is only what causes it and how the audience colours it.

The analysis of 66 media events across global FMCG brands yields three integrative findings, each carrying both theoretical and practical significance.

First finding: universal mechanics, specific nature. A media peak — its amplitude, decline rate and sentiment rebound — operates identically regardless of product category or country of origin. What differs is not how it unfolds but what generates it and how the audience codes it. This distinction should serve as the operational foundation of media planning: crisis and peak management frameworks can be applied universally; the tools for generating and framing a peak must be calibrated to the categorical and geographic nature of the event.

Second finding: category as a genre predictor. Hedonic categories inhabit a video-centric, affectively saturated media environment and generate peaks primarily through event sponsorship. Functional categories operate in a mixed genre with a high text share, and are the only cluster in the sample for which a crisis peak is a structural rather than incidental phenomenon. Beauty and personal care constitute the category of paid collaborations, marked by a pronounced sentiment spike followed by an equally pronounced rebound.

Third finding: country of origin as a sentiment modifier. Developed markets confer a systematic sentiment premium on brands: higher sentiment index and positive share at identical event mechanics. East Asian markets display structurally low sentiment scores that cannot be interpreted without applying a Cultural Expressiveness Baseline. Western markets construct brand media fields through text content and social advocacy; Asian markets do so through video and paid presence.