OTC retinoids are real — the fair framing is a potency-and-evidence ladder, not a real-vs-placebo split. The deep clinical record belongs to prescription tretinoin (retinoic acid): the 1988 double-blind, vehicle-controlled trial that started it all showed significant photoaging improvement in every participant who finished (https://pubmed.ncbi.nlm.nih.gov/3336176/). Retinol and retinaldehyde (retinal) are cosmetic cousins your skin converts toward retinoic acid — retinal is one conversion step away, retinol two. Fewer steps generally means more potency and more irritation; all of them work along the same pathway, with progressively less direct trial evidence as you move down the ladder (https://dermnetnz.org/topics/topical-retinoids).
For sensitive skin, that ladder is a feature. A retinoid only works if you use it for months, and the version you tolerate three nights a week beats the version that puts you on a two-week recovery break every time you use it. Plenty of people build from retinol to retinal and stop there, entirely satisfied; others graduate to tretinoin later. None of these rungs is a placebo.
Worth knowing if even retinol proves too cranky: in a 12-week randomized head-to-head, 0.5% bakuchiol matched 0.5% retinol on wrinkle and pigment outcomes with less scaling and stinging (https://doi.org/10.1111/bjd.16918) — one study, so hold it loosely, but a reasonable escape hatch.
Whichever rung you pick: pea-sized amount, start slow, moisturize freely, sunscreen daily, and judge at three months, not three weeks.