Your exfoliant log is hiding the variable that actually drives rosacea flares.

One field, four missing variables

Somewhere in a rosacea diary right now, a person just typed 'exfoliant: yes' and moved on. That entry will mean nothing in three months when they're trying to figure out why Tuesday's flare was bad and last Tuesday's wasn't.

The problem isn't the diary. It's that 'exfoliant yes/no' treats every alpha hydroxy acid as the same substance at the same depth, doing the same thing to the same tissue. It doesn't. Glycolic acid and mandelic acid are both AHAs in the same way that a pebble and a tennis ball are both objects you could drop into a bucket of water. The splash is completely different.

The variable that determines how deep an AHA penetrates skin before the surface barrier intercepts it is molecular weight. It's not brand, not 'AHA vs. BHA,' not even concentration alone. Molecular weight sets the speed of entry, and speed of entry is what determines whether a sensitized, neurovascular skin surface like rosacea-prone skin gets a controlled surface effect or a cascade it can't shut off.

The penetration-rate mechanism: why 76 Da and 152 Da are not the same conversation

Alpha hydroxy acids work by loosening corneocyte adhesion in the stratum corneum, the outermost skin layer, allowing dead cells to shed more evenly. That mechanism is real and well-established. The part the blanket-avoidance guidance skips is that 'how far the molecule travels before binding' is controlled by size.

Glycolic acid, at approximately 76 Daltons, is the smallest common AHA. That small footprint lets it move quickly through the intercellular lipid matrix of the stratum corneum and into the viable epidermis. In non-rosacea skin, fast penetration is considered a feature: faster turnaround, more visible resurfacing. In rosacea-prone skin, it means the stimulus reaches the dermal vascular and neural architecture that's already in a state of dysregulation.

Mandelic acid, derived from bitter almonds, sits at approximately 152 Daltons. That's roughly double the molecular weight of glycolic acid, and the penetration rate reflects it: mandelic acid moves more slowly through the skin barrier, which means it spends more of its contact time in the stratum corneum rather than progressing into the viable epidermis. Lactic acid lands in between, at approximately 90 Daltons, which is why its clinical behavior in sensitive skin tends to be intermediate.

The mechanism isn't unique to rosacea literature. Foundational research on AHA penetration dynamics documented the relationship between molecular weight and percutaneous absorption rate across multiple AHA subtypes, and that work has informed how cosmetic chemists and dermatologists think about acid selection ever since. The clinical consequence for inflamed or sensitive skin follows directly from that physics: slower penetration means the surface keratolytic effect happens before significant dermal contact, which reduces the vascular stimulus that triggers flushing and reactive inflammation.

What the 'avoid all exfoliants' framing gets wrong

The guidance that rosacea patients should avoid chemical exfoliants entirely didn't emerge from a controlled study comparing AHA subtypes in rosacea-specific tissue. It emerged from the reasonable observation that glycolic acid, the most-studied and most-marketed AHA, reliably irritates sensitive skin, and from the absence of any rosacea-specific AHA subtype trials to complicate that picture.

The absence of subtype trials is real. As of the clinical literature available to us, there is no published randomized controlled trial comparing mandelic acid directly against glycolic acid specifically in rosacea-diagnosed skin with flare incidence as a primary endpoint. A combination containing 20% azelaic acid and 10% mandelic acid exists, but no standalone mandelic acid RCT for rosacea is cited. ([source](https://pmc.ncbi.nlm.nih.gov/articles/PMC11512533/)) What does exist is a body of clinical evidence in broader sensitive-skin populations, small-scale tolerability studies in post-procedure skin, and dermatologist case experience that has produced something less than a verdict and more than a guess.

Some dermatologists do use mandelic acid in rosacea-prone patients. Not as a first-line, not without monitoring, but as a considered option where the clinical goal (mild exfoliation to reduce congestion or smooth texture) outweighs the risk, specifically because the slower penetration rate means the intervention can be titrated more precisely than glycolic acid permits. That's not a recommendation to go try mandelic acid. It's a reason the blanket-avoidance category contains an internal argument the label doesn't acknowledge.

The problem with 'avoid AHAs' as a rule isn't that it's wrong in practice for most patients. For most patients, especially glycolic acid in any meaningful concentration, it probably reflects accurate risk. The problem is that it's a category error dressed as chemistry. It applies the risk profile of the most reactive AHA subtype to the entire class, which leaves patients with no framework for understanding why their dermatologist might say different things in different contexts, and no way to track those contexts in their log.

Where clinical evidence does land on large-molecule AHAs in sensitive skin

The best available adjacent evidence comes from two directions: sensitive-skin tolerability studies in non-rosacea populations, and dermatologist-authored clinical protocols that include mandelic acid in rosacea-adjacent conditions.

A 2018 review by Kornhauser, Coelho, and Hearing in Dermatology Research and Practice examined the effects of AHAs on the skin's protective barrier, noting that lower-molecular-weight AHAs penetrate more readily and produce stronger irritation responses in compromised barriers. A Kornhauser authored a 2010 review on hydroxy acids classification, mechanisms, and applications, cited 257 times on PubMed (PMID 21437068). ([source](https://pubmed.ncbi.nlm.nih.gov/21437068/)) That framework directly supports the inference that higher-molecular-weight AHAs produce less barrier disruption in already-compromised skin, though the authors were writing about AHAs in general, not rosacea specifically.

On the mandelic acid side specifically: a small pilot study in acne-prone and post-inflammatory-hyperpigmentation patients with skin of color A comparative study examined 35% glycolic acid, 20% salicylic–10% mandelic acid, and phytic acid combination peels for treating active acne and post-acne hyperpigmentation. ([source](https://pmc.ncbi.nlm.nih.gov/articles/PMC11514325/)) reported better tolerability than glycolic acid controls at equivalent concentrations, attributed in part to the slower penetration rate. Skin-of-color patients in that context share one relevant variable with rosacea-prone patients: a surface that responds to deep-reaching irritants more intensely than the visible immediate reaction suggests.

None of this constitutes a clearance for using mandelic acid in a rosacea routine. It constitutes a basis for the observation that the penetration-rate mechanism is real, its clinical consequences are plausible, and the blanket category rule misrepresents the internal variation within the AHA class. Patients who ask their dermatologist 'what about mandelic acid specifically?' are asking a smarter question than the guidance they've been given trains them to ask.

What a useful log entry actually needs to capture

Here's where the diary problem becomes concrete. A typical rosacea flare log entry for an exfoliant application looks like this:

*Tuesday. Used exfoliant. Skin looked fine Wednesday. Thursday morning: flushing, papule on left cheek.*

Three days of data, one useful signal, zero variables that would let anyone, including the patient or their dermatologist, reconstruct what actually happened. Which AHA? What concentration? What pH did the formula carry? Was it the first use of the week or the fourth?

Those four variables, molecular size, concentration, pH, and application frequency, interact. A 5% mandelic acid at pH 3.5 applied four times in a week is a meaningfully different intervention than a 2% mandelic acid at pH 4.5 applied once. The molecular-size advantage of mandelic over glycolic collapses under high concentration, low pH, and high frequency. This is why some patients try mandelic acid on a dermatologist's tentative suggestion, flare badly, and conclude 'AHAs don't work for rosacea' when the more accurate reading is 'that concentration at that pH applied that often was too much, and I don't know which of those variables was the problem.'

The patient leaving the dermatologist's office with 'try mandelic acid if you must' has no framework for what to capture in their log to find out whether it worked. The log entries that accumulate are 'exfoliant yes/no' with flare yes/no, and the pattern, if one exists, stays invisible.

The four variables a rosacea exfoliant log needs to distinguish

If the penetration-rate argument is right, 'did you use an exfoliant?' is roughly as useful as 'did you eat something?' The log entry that actually generates a usable flare-correlation pattern captures:

Molecular subtype. Glycolic, lactic, or mandelic are not synonyms. Log the AHA by name. If the product lists 'alpha hydroxy acid' without specifying the type, treat that as missing data.

Concentration. The same mandelic acid at 2% versus 10% behaves differently in compromised barrier skin. The percentage is on the label; log it.

Estimated pH. This is harder because most consumer products don't print their pH. Some dermatology-focused brands do; for those that don't, the product category gives a rough guide, and a pH strip test on the formula at home is more reliable than you'd expect. Log what you have. 'Unknown' is more accurate than not capturing the field.

Application frequency that week. A once-weekly application at any concentration is a different cumulative stimulus than daily application. The flare that arrives Thursday may be responding to Wednesday's fourth application of the week, not Wednesday's application alone.

None of this is onerous if the log is built to accept it. The difficulty is that 'exfoliant yes/no' is a default that most diaries, paper or app-based, were built around because it's the simplest version of the question. Simplest is fine for capturing that something happened. It can't tell you which part of the something matters.

How Skinframe approaches exfoliant logging

We built Skinframe's product log around the observation that a 'used exfoliant' checkbox is a dead end for pattern analysis. The topical product entry captures the product type, the AHA subtype where the product lists it, the concentration, and the application count within the rolling seven-day window, not just the day's binary.

The reason is exactly the mechanism above: whether a flare follows an AHA application is determined by cumulative dermal load across several variables, not by a single Tuesday entry. A log that can separate 'mandelic acid 2%, pH 4.5, first use this week' from 'glycolic acid 8%, pH 3.0, fourth use this week' has the structural information needed to surface a correlation. One that can't, can't.

We're also not going to claim Skinframe will tell you which AHA to use or whether mandelic acid is safe for your skin. That's a dermatologist conversation, and a structured log is most useful when you bring it to that conversation with something more than 'I think exfoliants bother me sometimes.' Your dermatologist can read a pattern; they can't read a series of yes/no checkboxes.

Pre-launch, Skinframe is available on iOS. Talk to your dermatologist about your exfoliant routine first. Then build the log that can actually track it.