Cosmetics and Expansion: KIKO Milano Enters a Game of Positions

The same dynamic works differently within shopping centers. If existing brands remain while others enter, adding a new operator does not necessarily refresh the category but can instead increase competition within it. The decision to bring in new brands therefore depends not only on space availability, but also on the value a new brand can add to the tenant mix alongside chains that are likely to remain at the property.

That increased competition is also occurring in an expanding market. Mordor Intelligence estimates that Mexico’s beauty and personal care industry will reach $17.6 billion in 2026 and nearly $22.4 billion by 2031, growth that could expand the available business even as more brands compete for it.

That growth, however, has not yet translated into a comparable expansion of the category within the shopping centers analyzed. New locations can be added alongside existing competitors at one property while, at other times or properties, positions in the category disappear. Since mid-2023, the number of brands present has therefore remained virtually unchanged, at between 15 and 16 per quarter, while their combined locations edged up from 160 in the first quarter of 2024 to 164 in the second quarter of 2026⁴.

The difference between these two scales changes what growth means. In a segment occupied by established competitors, gaining ground requires differentiating product, price and commercial proposition enough to generate the profitability needed to justify each store. Under that logic, expansion ceases to be an end in itself, as what matters less is how many positions a brand can open than how many it can maintain without compromising the business that supports them.

Want to learn more about Mexico’s retail market performance? Visit SiiLA Market Analytics or email us at contacto@siila.com.mx.

***

¹ SiiLA calculations based on quarterly data for cosmetics and makeup establishments in shopping centers monitored by SiiLA in Mexico City, Guadalajara and Monterrey. For this analysis, each location corresponds to a unique brand–property relationship in a given quarter; multiple records for the same brand within a property are counted as a single presence. The annual change in each network is calculated by comparing the number of properties where a brand is present with the number observed for that same brand four quarters earlier. Among networks present at both endpoints of each annual comparison, 331 observations were analyzed; the median number of new locations was zero, as was the 75th percentile, while the 90th percentile was one location and the 95th percentile was two. The annual net change likewise had a median of zero locations. Annual retention is calculated as the share of brand–property relationships observed in one quarter that remain at the same property four quarters later, considering only properties comparable in both periods. As a robustness check, the calculation was replicated using a constant panel of 71 shopping centers observed over 24 quarters, yielding an average annual retention rate of 95.9%. For the five-year horizon, brand–property relationships observed in Q2 2021 were compared with those recorded in Q2 2026 within the comparable universe: of 159 initial relationships, 126 remained at the end of the period, equivalent to a 79.25% retention rate. Figures reported in the text are rounded. Data period: Q4 2019–Q2 2026.

² New entrants are limited to brands whose first presence can be observed within the period analyzed. Brands present from the beginning of the series are not considered new entrants because their entry date predates or cannot be determined from the available data. Given the small number of observed entries, the analysis cannot establish a statistical relationship between time in the market and network expansion.

³ To assess whether brand entries coincided with competitor departures within the same property, unique brand–property events were constructed and entries were compared with departures of other brands using windows of 0, ±1 and ±2 quarters. The results were replicated using a constant panel of properties and a 10,000-iteration permutation test in which departures were randomly reassigned among properties within each quarter. The share of coincidences remained low as the time window widened and showed no statistically significant evidence of a spatial association greater than expected under random assignment (p=0.1767 for coincidences within the same quarter). Of the 31 brand–property entries observed in the full universe, 29 (93.5%) did not coincide with the simultaneous departure of another brand in the sample from the same property. The analysis identifies temporal coincidences within a property, not causality or replacement of the same physical retail space.

⁴ The number of brands corresponds to the total number of distinct operators observed in each quarter, while total locations corresponds to all unique brand–property relationships recorded during that period; multiple records for the same brand within a property are counted as a single location. The figures describe the aggregate presence of brands included in SiiLA’s sample, not the entire universe of cosmetics establishments in Mexico. Changes between periods should be interpreted within the coverage of properties observed by SiiLA.