How AI Search Is Recommending Clean Makeup Brands: Monthly Trends

Executive Summary

The clean makeup category moved in September 2026, with three brands recording significant declines while the leader held steady. e.l.f. Cosmetics remains the category leader at 45.6% valid recommendation coverage in September 2026, down 0.1 points from 45.7% in July 2026. Rare Beauty holds second at 43.7%, up 1.2 points over the same period, narrowing the gap between the top two brands to 1.9 points. The strongest upward mover this month was Thrive Causemetics, which rose 1.7 points from August 2026 to September 2026, while the sharpest decliner was Glossier, down 5.0 points in a single month.

This was not a quiet month. The benchmark records significant movement for three brands. Glossier fell 6.2 points from July 2026 to September 2026, a two-month decline that crossed the significance threshold. ILIA Beauty dropped 8.5 points over the same window, and Tarte Cosmetics fell 7.6 points. All three are classified as significant decliners across the three-month series.

The leader's stability contrasts with the broader category. e.l.f. Cosmetics recorded a notable rise in rank-one recommendations, up 4.1 points from 10.2% in July 2026 to 14.3% in September 2026, even as its overall coverage held flat. The commercial picture is one of a stable leader, a narrowing race for second, and a widening separation between the leading cluster and three brands losing ground across multiple placement signals.

Each monthly run begins with 800 prompt-surface observations (627 unique questions in July 2026, 632 in August 2026, 618 in September 2026) across the benchmark's defined AI/search surface universe. Of those, 800 mentioned a tracked brand or competitor in all three months; 737 were relevant and 63 were irrelevant in July 2026, 715 were relevant and 85 were irrelevant in August 2026, and 699 were relevant and 101 were irrelevant in September 2026. The public metrics use the 626 qualified observations in July 2026, 640 in August 2026, and 645 in September 2026 that survive both qualification stages.

Key Findings

Benchmark Context

The report separates the raw collection universe from the qualified analysis set. Brand-level recommendation percentages are calculated within the qualified benchmark set.

August 2026 sat between these months with 640 qualified observations and 632 unique questions. The intermediate month's relevance count of 715 sits between the July and September figures, showing a gradual drift toward more prompts being classified irrelevant to the vertical.

Benchmark-Level Metrics

The table below aggregates category-level coverage and shortlist metrics for both months.

The valid recommendation shortlist share fell from 78.3% in July 2026 to 74.2% in August 2026, then to 69.3% in September 2026. The recommendation-shaped answer share moved in the same direction, from 49.0% to 47.5% to 46.0%. Both declines occurred within the same six qualified AI surface families, so the shift reflects answer composition rather than a change in surface coverage.

AI Recommendation Trend

Valid recommendation coverage is separating at the top while three brands decline significantly

The three-month series shows a leading cluster of e.l.f. Cosmetics and Rare Beauty holding near 45%, a middle tier of Tower 28, ILIA Beauty, and Kosas between 30% and 36%, and a lower tier where Glossier, Tarte Cosmetics, and others have lost ground. The significant declines are concentrated among brands that entered the series in the middle and upper-middle tiers.

This is not a broad category shift; it is a concentrated pattern of decline among Glossier, ILIA Beauty, and Tarte Cosmetics, each showing a two-month downward streak. The remaining seven brands held within normal month-to-month variation.

What Changed This Month

e.l.f. Cosmetics

e.l.f. Cosmetics held category leadership at 45.6% valid recommendation coverage in September 2026, down just 0.1 points from 45.7% in July 2026. In the immediate prior month, the brand had reached 48.0%, meaning its September figure represents a 2.4-point pullback from August. The brand remains stable overall, but the monthly pattern is one of an August peak followed by a September retreat.

The more telling movement is at the top of the recommendation structure. e.l.f. Cosmetics' rank-one recommendation rate rose 4.1 points from 10.2% in July 2026 to 14.3% in September 2026, a notable increase. Its raw mention presence also rose, from 57.5% to 61.4%, even as its top-three rate fell 2.7 points to 21.7%.

The distinction to notice is between overall coverage and placement quality. The brand is being mentioned more often and recommended first more often, yet its total coverage is flat because it is appearing in the top three less frequently. The rank-one gain is the notable signal, and it happened while the brand's August peak in coverage softened.

Highest-priority diagnostic: Which prompts are driving the rank-one increase, and why does the brand's top-three rate decline even as its rank-one rate rises?

Rare Beauty

Rare Beauty rose 1.2 points in valid recommendation coverage, from 42.5% in July 2026 to 43.7% in September 2026, closing the gap to the category leader to 1.9 points. The brand's raw mention presence rose 5.0 points to 66.0%, the highest presence rate in the category and a sizable increase from its baseline.

The brand's top-three recommendation rate rose 1.0 point to 21.1%, but its rank-one rate fell 0.6 points to 6.0%. Rare Beauty is surfacing in more AI answers than any other tracked brand, yet its share of first-place recommendations trails e.l.f. Cosmetics' 14.3% by a wide margin. In September 2026, the brand recorded 282 valid recommendations.

The distinction to notice is between visibility and top placement. Rare Beauty leads the category in presence but converts that presence into rank-one placement at less than half the rate of the category leader. The gap between the two brands at rank one is the structural difference in their recommendation profiles.

Highest-priority diagnostic: Which recommendation contexts place Rare Beauty in the answer but not at the top, and why does its high presence rate not translate into more rank-one placements?

ILIA Beauty

ILIA Beauty fell 8.5 points in valid recommendation coverage, from 40.9% in July 2026 to 32.4% in September 2026, the largest decline in the category and a significant two-month streak. The brand dropped from the third-ranked position in July 2026 to fourth in September 2026. In 209 valid recommendations, the brand is still being recommended, but far less often and far less prominently.

Every placement signal declined. The rank-one rate fell 10.1 points from 17.4% to 7.3%, and the top-three rate fell 11.5 points from 26.8% to 15.3%. Raw mention presence fell 10.6 points from 51.4% to 40.8%. In July 2026, ILIA Beauty was the most frequently recommended brand at rank one; by September 2026, it had lost more than half of that share.

The distinction to notice is that sentiment held steady while placement collapsed. The brand's net sentiment score stayed near 0.9 across all three months, meaning AI systems still frame ILIA Beauty positively when they mention it. The loss is in how often it is surfaced and where it is ranked, not in how it is described. The notable gap between ILIA Beauty and Rare Beauty widened from 1.6 points in July 2026 to 11.3 points in September 2026, growing every month.

Highest-priority diagnostic: Which competitor is taking the rank-one and top-three positions ILIA Beauty previously held, and which product categories or prompt types account for the sharpest losses?

Glossier

Glossier fell 6.2 points in valid recommendation coverage, from 20.3% in July 2026 to 14.1% in September 2026, a significant two-month decline. The brand fell from 7th position in July 2026 to 8th in September 2026, and its position relative to the category weakened considerably. In the immediate prior month, Glossier fell 5.0 points from August to September alone, the sharpest single-month drop in the category.

The decline spans every placement signal. Glossier's raw mention presence fell 5.6 points from 34.0% to 28.4%, its top-three rate fell 6.2 points from 10.2% to 4.0%, and its rank-one rate fell 1.5 points from 2.1% to 0.6%. All three metrics moved in the same direction as its overall coverage decline. In September 2026, the brand recorded just 91 valid recommendations, down from 127 in July 2026 and 122 in August 2026.

The distinction to notice is the breadth of the decline. Glossier is losing ground on presence, top-three placement, and rank-one placement simultaneously. The benchmark cannot distinguish platform behavior from measurement effects, but the breadth of the decline across every placement signal is the notable feature of this month's data. The brand's rank position fell from 7th to 8th in the tracked set, and its coverage now sits closer to the lower tier than the middle tier it occupied in July 2026.

Highest-priority diagnostic: Which AI surfaces or question types have stopped returning Glossier in their answers, and which brands are appearing in its place?

Tarte Cosmetics

Tarte Cosmetics fell 7.6 points in valid recommendation coverage, from 16.8% in July 2026 to 9.2% in September 2026, a significant decline and the second consecutive month of significant movement. The brand dropped from 8th to 9th in rank order. In 59 valid recommendations, the brand is being recommended at less than half the rate it was in July 2026.

The decline is visible across the recommendation structure. Raw mention presence fell 8.8 points from 24.3% to 15.5%, and the top-three rate fell 3.2 points from 6.9% to 3.7%, both notable declines. The rank-one rate fell 0.3 points to 1.6%, a smaller movement that did not cross the significance threshold. The notable gap between Rare Beauty and Tarte Cosmetics widened from 25.7 points in July 2026 to 34.5 points in September 2026.

The distinction to notice is the steady bleed across three months. Tarte Cosmetics declined from 16.8% to 13.4% to 9.2%, with the August drop of 3.4 points and the September drop of 4.2 points both registering as significant. With only 59 valid recommendations, the small-count context means each percentage point represents fewer than six observations, and the movement should be read as directional within this three-month record.

Highest-priority diagnostic: Which product categories or prompt types account for the concentrated loss, and which competitor is most often recommended where Tarte Cosmetics previously appeared?

Tower 28

Tower 28 held third position at 36.3% valid recommendation coverage in September 2026, down 0.8 points from 37.1% in July 2026. The brand peaked at 39.2% in August 2026, meaning its September figure represents a 2.9-point pullback from the prior month. The brand remains stable overall.

Raw mention presence held essentially flat at 50.4%, down 0.2 points from the July baseline. The brand's coverage movement came from ranking shifts rather than changes in how often AI systems surface it. Its top-three rate fell 1.8 points to 11.6%, while its rank-one rate rose 0.1 points to 2.5%. In September 2026, the brand recorded 234 valid recommendations.

The distinction to notice is that Tower 28's presence rate of 50.4% remains the third-highest in the category, but its conversion of that presence into coverage has softened. The brand is still being surfaced in half of all qualified observations, yet its rank-one rate of 2.5% places it in the middle of the tracked set rather than near the top.

Highest-priority diagnostic: Why has Tower 28's coverage pulled back from its August peak even as its presence rate held steady, and which ranking contexts shifted?

Kosas

Kosas held fifth position at 30.5% valid recommendation coverage in September 2026, down 1.4 points from 31.9% in July 2026. The brand peaked at 33.4% in August 2026, making its September figure a 2.9-point pullback from the prior month. The brand is classified stable.

Raw mention presence fell 2.3 points to 41.5%, and the top-three rate fell 2.5 points to 12.7%. The rank-one rate rose 0.6 points to 2.8%, a modest counter-trend within an otherwise downward month. In September 2026, the brand recorded 197 valid recommendations, down from 214 in August 2026.

The distinction to notice is the compression in the middle tier. Kosas, Tower 28, and ILIA Beauty all lost ground in September 2026, even as the top two brands held their positions. The separation between the leading pair and the middle tier is the structural feature of this month's data.

Highest-priority diagnostic: Which prompts drove the August peak and the September pullback, and is the movement concentrated in specific product categories?

Milk Makeup

Milk Makeup rose 1.2 points from August 2026 to September 2026, a modest gain, though the brand remains down 1.8 points from its July 2026 baseline. The brand holds sixth position at 23.4% coverage. In 151 valid recommendations, Milk Makeup reversed a two-month slide with a modest September gain.

The brand's raw mention presence fell 4.4 points from 37.1% in July 2026 to 32.7% in September 2026, but its rank-one rate rose 0.3 points to 2.2%. The coverage gain from August to September came despite a continued decline in overall presence, pointing to a ranking improvement rather than a visibility increase. The brand's net sentiment score rose from 0.8 to 0.9.

The distinction to notice is the divergence between presence and placement. Milk Makeup is being surfaced less often across the three-month series, but when it is surfaced, it is being recommended at a slightly higher rate. That combination kept the brand stable even as its presence fell.

Highest-priority diagnostic: Which ranking contexts improved between August and September 2026, and can that improvement be traced to specific prompts or product categories?

Thrive Causemetics

Thrive Causemetics rose 1.7 points from August 2026 to September 2026, from 13.3% to 15.0% coverage, the largest upward move this period, though the brand remains down 1.8 points from its July 2026 baseline of 16.8%. The brand holds 7th position. In 97 valid recommendations, the September gain reversed part of the August decline.

Raw mention presence rose from 16.2% in August 2026 to 18.1% in September 2026, still down 1.1 points from the July baseline of 19.2%. The brand's rank-one rate of 2.0% and its top-three rate of 6.0% remain modest, and neither moved significantly across the three-month series. The brand's net sentiment score held at 0.9.

The distinction to notice is the parallel movement with Tarte Cosmetics. Both brands entered the series at 16.8% coverage in July 2026, but they have diverged sharply: Thrive Causemetics held near 15.0% while Tarte Cosmetics fell to 9.2%. The separation between the two brands widened across the series.

Highest-priority diagnostic: Why did Thrive Causemetics hold its position while Tarte Cosmetics, which entered the series at the same coverage level, declined by more than 7 points?

Beautycounter

Beautycounter rose 0.8 points in valid recommendation coverage, from 2.6% in July 2026 to 3.4% in September 2026, its second consecutive month of upward movement. The brand remains at the bottom of the tracked set at 10th position. In 22 valid recommendations, the brand continues to operate with very small underlying counts.

Raw mention presence fell 0.3 points to 3.9%, even as coverage rose. The brand's top-three rate rose 0.6 points to 1.4% and its rank-one rate rose 0.3 points to 0.6%, both moving in the same direction as its coverage. Net sentiment rose from 0.6 in July 2026 to 0.9 in September 2026, though this movement is based on a very small number of mentions.

The distinction to notice is the small-count context. With 22 valid recommendations out of 645 observations, the brand's coverage rate of 3.4% represents a small number of actual recommendations. The upward movement across three months is noteworthy for its consistency, but each percentage point represents roughly six observations.

Highest-priority diagnostic: Which specific prompts surface Beautycounter at all, and what common thread explains the limited contexts where the brand appears?

Buyer-Intent Interpretation

All 645 qualified observations in September 2026 fell into the Brand Recommendation cluster, the same as in July and August 2026. The public benchmark cannot yet answer pricing, value, or head-to-head comparison questions because no qualified observations landed in those clusters. Pricing analysis did appear in the raw classification as a response type, but not at a level that supports a published finding. The strategic implication is that the current benchmark measures which brand gets recommended, not how AI systems frame the trade-offs between brands on price, value, or direct comparison.

Brand Opportunity Summary

The benchmark identifies where attention is warranted; a company-level analysis is needed to explain why.

Evidence Behind the Benchmark

The aggregate metrics are built from prompt-level observations (query, surface, recommendation outcome, rank, sentiment, and citations where exposed). Company-level analysis can go deeper into prompt, competitor, surface, and evidence patterns. Source presence is not automatically treated as proof of causation.

About This Benchmark

This report is part of the LLM Authority Index AI Market Discovery research program.

• AI Industry Market Discovery Methodology
• AI Industry Market Discovery Metrics
• AI Industry Market Discovery Standards

Report-Specific Interpretation Notes

• The qualified denominator (645 observations in September 2026) is the public analysis set; the raw collection of 800 prompts is broader and includes prompts that were deemed irrelevant or reserved.
• Small-count brands (Beautycounter at 22, Tarte Cosmetics at 59, Glossier at 91, Thrive Causemetics at 97 valid recommendations) carry more month-to-month variance, and their movements should be read as directional within this three-month record rather than as large-sample findings.
• Movement analysis identifies changes worth investigating. A brand gaining or losing coverage is a benchmark finding, not evidence of a specific cause.
• Three brands are classified as significant decliners this period (Glossier, ILIA Beauty, Tarte Cosmetics); the remaining seven are classified stable on valid recommendation coverage.

Next Step

The Public Benchmark Shows Where a Brand Is Winning or Losing. A Company-Level Audit Shows Why.

Beneath the aggregate coverage percentage sit the questions that matter commercially: which high-intent prompts is the brand winning, which competitor takes the recommendation when the brand loses, what attributes do AI systems associate with each option, and which external sources are shaping those answers. The benchmark shows the scoreboard. It does not reveal the play that produced the result.

A company-specific AI visibility audit maps those prompt, surface, competitor, ranking, sentiment, and evidence-source patterns into a prioritized visibility strategy. That is where the story behind the movement becomes actionable.

Request an AI visibility audit