Why Skincare Without Records Is Guessing
Anyone who has worked through a multi-step AB routine long enough to have introduced actives, added an essence, and started experimenting with pH adjusting toners knows the specific frustration of not knowing what changed. Your skin looked notably better for about two weeks in October and then things went sideways, and you have no idea whether it was the new AHA you started, the sheet mask you used four times that week, seasonal humidity change, or the fact that you stopped using your first essence. The timeline lives in your head, approximately.
A daily log with product links per step — not just "used exfoliant" but a linked record to the specific Cosrx BHA with its ingredient list and the purchase date — creates the longitudinal record that makes it possible to actually run controlled experiments on your own face rather than just applying products in sequence and hoping.
The Testing checkbox per product step is the variable-control mechanism. When you're introducing a new product into the rotation, mark it as Testing. That flag in the log lets you filter the days where the new product was present against the days before it was introduced, and compare skin appearance ratings across those two populations. This is the closest a home user can get to a single-variable test in a domain where everything is confounded.
The Twelve-Step Structure as a Data Architecture
The template mirrors the standard Korean skincare layering order: 1st Cleanser (oil cleanse), 2nd Cleanser (water-based), pH Adjusting Toner, Actives, First Essence, Toner, Essence/Serum/Ampoule, Emulsion or Light Moisturiser, Medium or Heavy Moisturiser, Facial Oil, Sleeping Pack, Sunscreen. Each step carries a linked product entry from the Stash library and its own Testing flag.
This architecture captures something that a simple "products used today" log cannot: the order matters, and what you used at which layer matters. A BHA introduced at the Actives step responds differently depending on whether you've already applied a pH adjusting toner to prep the skin's surface. An oil-based serum applied under a water-based emulsion versus over it is a different outcome. The step structure makes these positional decisions part of the record.
The AM/PM radio field is the field that catches the routine divergence that most people don't track. Your PM routine probably includes actives, a sleeping pack, no sunscreen. Your AM routine is probably shorter, sunscreen is mandatory, actives may or may not be present depending on sensitivity. These are not the same routine happening twice a day — they're different product sets with different purposes, and logging them together in a single "today's routine" entry loses the distinction that matters for effect attribution.
Reading the Skin Outcome Data
Overall Skin Appearance rated 1 to 5 is the daily endpoint measure. Not a qualitative note, but a number. Numbers are filterable.
Active Acne as a 4-level scale — 0, 1, 2-3, 4+ — gives a daily breakout count that's objective enough to trend. The Left, Right, and Center photo fields make the assessment spatial: left cheek, right cheek, and center panel (T-zone, chin). Left-cheek dominant breakouts have different causes than chin breakouts, and the photo documentation creates a visual record that shows you not just how many but where, which affects what interventions to consider.
After 60 days of entries, filtering by Active Acne count and looking at which products were marked Testing during the high-count days versus the clear-skin days is the closest approximation to a cause-and-effect analysis you can run on your own skin without a dermatologist's controlled study design.
The Notes field is where the real confounders live: "stayed up until 3am," "high humidity day," "used a new pillowcase," "ate a lot of dairy this week." These are the variables the Testing flags cannot capture, and they're the reason that no routine correlation is ever clean — but having them in the record means you can at least see them in context rather than pretending they don't exist.