A $134.5 million tariff refund lifted Lululemon’s second-quarter gross margin by 560 basis points. Strip out that disclosed effect and the quarter tells a much harder story about demand, product relevance, and the difference between cash recovery and repeatable retail economics.
A September 3 leadership update connects AI, global sourcing, product, architecture, real estate, and logistics; the signal for smaller labels is decision ownership, not a promise that AI will fix execution.
Reducing launch frequency can concentrate product, inventory, service, and learning—but only when a brand measures the whole operating cycle instead of treating scarcity as a strategy.
A visible spike becomes a category question only when search, sales quality, reorders, returns, customer use, supply, margin, and time point in the same direction.
A healthy assortment is not a universal category percentage; it is a cash-aware system that separates core, test, seasonal, and exit inventory while preserving size and customer truth.
Early Spring and Cruise 2027 collection notes point toward controlled lines, adjustable volume, drape, and movement—but the public sample is too early to call a market-wide trend.
Climate information can improve a fashion plan only when long-term baselines, near-term forecasts, observed demand, inventory flexibility, and customer promises stay separate.
A defensible forecast needs a traceable claim, archived source, observation window, transformation, baseline, uncertainty, counterevidence, owner, and expiration date—not a confident trend label.
Replace the viral screenshot with a weekly evidence routine that separates relative attention, persistence, customer questions, commercial relevance, and uncertainty.
Inventory-responsive markdowns can be useful. Secretly estimating what one shopper will tolerate is a different practice with a different trust and regulatory risk.