Research · March 2026
Why Merrell Dominates AI Hiking Boot Recommendations — and What the Signal Data Shows
Merrell scored 74.55/100 across 60 buyer-intent prompts. Salomon was second at 70.0. The next brand — Hoka — scored 46.3. A 24-point gap between second and third place is not a coincidence.
Merrell scored 74.55/100 across 60 buyer-intent prompts on Perplexity Sonar in W13 2026. When mid-tier brands were not recommended, Merrell appeared in 74.4% of those same prompts. We looked at the structural signal data to understand what might explain this — and where we are still inferring.
There is approximately a 24-point gap between the second-place brand (Salomon at 70.0) and the third-place brand (Hoka at 46.3) on Perplexity Sonar in this dataset. A gap that large, between two brands with comparable marketing resources and distribution, does not appear by accident.
This article is our attempt to explain what the data shows — and to be precise about where the data is conclusive versus where we are making inferences.
Disclosure: Envoyra sells the AI visibility scanning tool used to collect this data. We have a financial interest in findings that demonstrate the value of AI visibility measurement. The methodology is disclosed in full below, including the limits of what the data can and cannot show.
Dataset summary
Platform: Perplexity Sonar (standard tier). Prompts: 60 buyer-intent queries. Brands: 25 hiking boot brands. Timeframe: W12–W13 2026 (2 weeks with complete data). Total prompt-response pairs: 1,372. Total brand mention events: 262. Evidence snippets coded: 247 AI-generated recommendation fragments.
An evidence snippet is the AI-generated text segment from a Perplexity response where a brand was mentioned — extracted from the first 500 characters of each response. Signal frequency scores are calculated from these snippets.
Note: Hoka and several other brands have no W11 data in our dataset. W12 and W13 are the two weeks with complete cross-brand coverage.
What we observed
74.55 vs 70.0 vs 46.3 | Merrell vs Salomon vs Hoka — Signal Scores, Perplexity Sonar, W13 2026
Score vs mention share — what the difference means: Signal Score reflects weighted structural presence across prompts. A brand mentioned authoritatively in 12 of 20 prompts scores differently than one mentioned weakly across 15. Mention share reflects raw inclusion frequency. A brand can lead on raw mention share while scoring lower on Signal Score if its mentions appear in lower-confidence contexts.
Salomon scored 70.0/100 with a slightly higher raw mention share of 16.0%, versus Merrell's 15.6%. Merrell's higher Score despite lower raw mention share suggests its citations carry more structural authority.
The week-over-week trend
Between W12 and W13, Merrell's inclusion rate moved from 68.3% to 75.0% while Salomon's declined from 79.8% to 71.7%. By W13, Merrell's Signal Score (74.55) exceeded Salomon's (70.0).
This divergence — higher raw inclusion for Salomon in W12 but higher weighted Score for Merrell by W13 — is worth noting. It suggests Merrell's structural signal profile was compounding more effectively than raw mention frequency alone would predict.
Hoka had no data in W11. In W12 its inclusion rate was 0.0%. In W13 it was 48.3% — a 48-point swing in one week with no observed change in Hoka's own content.
We note this not as a structural signal finding but as a volatility finding. If a brand's Perplexity inclusion can move 48 points in one week without brand-side action, that limits how confidently we can attribute any brand's current score to specific structural choices. It is also the primary argument for weekly monitoring over single-snapshot measurement.
The displacement pattern
For Oboz (rank 6, 32.8% inclusion across its relevant prompt subset):
In the 39 prompts where Oboz was not recommended, Merrell appeared in 29 of those same prompts — 74.4% of Oboz's missed prompts. Salomon appeared in 25 of those same prompts — 64.1% of Oboz's missed prompts.
74.4% | of prompts where Oboz was absent, Merrell appeared — the clearest displacement pattern in our dataset
If you're not in the shortlist, Merrell and Salomon are taking your slot. Every single time. Explore the research →
Whether this displacement pattern holds at similar rates for other mid-tier brands is something we have not yet measured with the same granularity.
Two distinct signal measurements — kept separate
Our dataset contains two distinct signal measurements that must not be conflated.
Measurement A: Evidence snippet signal frequency — what AI actually said about brands across 247 recommendation fragments. This reflects downstream AI behavior — what signals appeared in Perplexity's generated text.
Category averages in AI-generated text: Waterproof certification language (53.4%), Structured spec formatting (53.0%), Use-case positioning (26.7%), Numeric measurement density (14.2%), Traction/sole certifications (10.5%), Third-party authority citations (0.8%).
Measurement B: Brand page crawl signals — what structural signals were detectable on Merrell's product pages. This reflects the brand's own content architecture.
Merrell's page crawl scores: Numeric measurement density (100%), Waterproof certification language (42.9%), Third-party authority citations (28.6%), Traction/sole certifications (13.3%), Use-case positioning (14.3%), Structured spec formatting (0%*).
*Merrell's product pages appear to be JavaScript-rendered. The 0% structured spec formatting score reflects what was accessible to a standard crawler — not what a human visitor sees.
The third-party citation gap is the most striking finding across both measurements. Category average in evidence snippets: 0.8%. Merrell's page crawl score: 28.6%. We observe this as a correlation — we do not claim it causes Merrell's higher score.
What this suggests for brands in the 20–40% range
Three structural areas appear worth examining based on this dataset:
First, third-party authority citations. The gap between 0.8% category-average snippet frequency and Merrell's 28.6% page crawl score is the largest structural differential we found across both measurement types.
Second, waterproof certification language in the first 120 words of product content. This signal appeared in 53.4% of category-wide evidence snippets — the highest frequency of any signal class in AI-generated responses.
Third, comparison page architecture. Direct comparison pages targeting prompts where displacement is highest ("Brand vs Merrell," "Brand vs Salomon") address the specific query types where the 74.4% displacement pattern is most concentrated.
These are observations from one platform, one vertical, two weeks of complete data. We are continuing to track weekly and will update findings as the dataset grows.
Explore the research
Most brands assume they show up in AI recommendations. Most don't.
The research documents which brands appear, which are consistently absent, and what the structural differences are.
Data: W12–W13 2026 · 60 buyer-intent prompts · 25 brands · Perplexity Sonar 1,372 prompt-response pairs · 247 evidence snippets coded · 262 brand mention events Envoyra Engine v1.0 · envoyra.com · March 2026