Cited Layer › Measurements › Skincare, UAE
Measurement
Skincare in the UAE: the label held, the winner did not
Runs 20 and 25 October 2026 · ChatGPT (two free accounts, two browsers) and Perplexity · 24 recorded answers to four questions
Four buyer questions about skincare and body care, asked on two dates five days apart. The label an engine attached to a brand repeated more reliably than the brand it named first.
"Budget" stayed on the same brand in 6 of 6 general answers. "Best overall" went to the same product in 15 of 18 answers across both dates — but on the second date, the body-lotion question produced three different first picks in three answers. And four sites built most of what the answers said: three pharmacy and retail storefronts and one newspaper.
Method
What was asked, and how
- C0 where to buy CeraVe in Dubai — control, first in every session. Had it failed to return a list of pharmacies or stores, the session would have been discarded.
- Q1 best moisturizer for dry skin in UAE heat
- Q2 best sunscreen to buy in Dubai
- Q3 best body lotion to buy in UAE
- Q4 NIVEA vs Vaseline body lotion for UAE weather — a head-to-head from the same shelf as Q3, to see whether one brand gets the same outcome in a general question and a direct comparison.
Mode, held constant: two free ChatGPT accounts in two browsers, temporary chat; Perplexity once per session, Search mode, free. Session 1 on 20 October, session 2 on 25 October. Every control passed. For each answer: brands in order, the label attached, and every source, classed as brand site, retailer, review or media, or community.
Disclosure: runs were made from a Ukrainian IP address. Every question names the country or city, which reduces the effect of location but does not remove it. One answer (Q3, account 2, session 2) is missing its Sources panel; for it only the links inside the text are counted.
The result
Who was named first, both dates
| Question | ChatGPT, 4 runs | Perplexity, 2 runs |
|---|---|---|
| Q1 moisturizer | CeraVe, "best overall" — 4 of 4 | CeraVe, "best overall" — 2 of 2 |
| Q2 sunscreen | La Roche-Posay UVMune 400 → Eucerin Sun Oil Control → ISDIN Fusion Water — same top three, same order, 4 of 4 | 20 Oct: La Roche-Posay · 25 Oct: Beauty of Joseon; La Roche-Posay not named |
| Q3 body lotion | 20 Oct: CeraVe ×2 · 25 Oct: Aveeno, CeraVe | 20 Oct: CeraVe · 25 Oct: NIVEA |
| Q4 NIVEA vs Vaseline | NIVEA described as the everyday option in 3 of 4; Vaseline for very dry or rough skin in 4 of 4 | NIVEA "better everyday", Vaseline "repair" — 2 of 2 |
Only the moisturizer question kept the same first pick across all six answers. The product behind the name still changed between runs (cream on one, lotion on another).
Finding 1
The label is steadier than the ranking
In the general body-lotion question, Vaseline was labelled "budget" in all six answers — "best budget", "budget-friendly", "budget supermarket option" — while its position in the list moved between third and fifth. In the head-to-head, the split between the two brands ("everyday" against "dry skin / repair") held in 5 of 6 answers for one side and 6 of 6 for the other.
Placement in the list was noise; NIVEA was absent, fifth, or first in the same general question depending on the run. What the engine says a brand is for repeated; where it put the brand did not.
None of the 24 answers marked a brand's own wording as a claim ("according to the manufacturer"). On the second date softer phrasing appeared — "is described as", "marketed for" — but product properties were mostly restated as facts.
Finding 2
Four sites built the answers
Crossing the source lists of both engines on Q1–Q3, over both dates, leaves a short list of sites that both engines used:
| Site | Type | Used by |
|---|---|---|
| myaster.com | pharmacy | both engines, Q1–Q3 |
| lifepharmacy.com | pharmacy | both engines, Q2–Q3; on 25 Oct its product page sat under Perplexity's first pick in Q3 |
| noon.com | marketplace | both engines, Q2–Q3 |
| gulfnews.com | newspaper | both engines, Q1, both dates — the same 2023 round-ups each time |
Two review sites appeared in one engine only: cartiq.net in 7 of 15 ChatGPT answers with sources and 0 of 8 in Perplexity; zotezo.com in three ChatGPT answers. Brand sites mattered in the head-to-head (the NIVEA Middle East site appeared in all four ChatGPT comparison answers) and almost nowhere in the "best" questions.
Reddit, Quora and LinkedIn: 0 in all 24 answers.
Finding 3
A name that rests on one source leaves with it
On 20 October Perplexity put La Roche-Posay first for sunscreen, with a single specialist site, sunscreendubai.ae, under it. On 25 October that site was not among the sources, and La Roche-Posay was not named at all. In the same five days ChatGPT, drawing on pharmacy pages, kept the identical top three.
Also observed
Two smaller observations, not yet rules
- Sources followed the account, not the date. On the moisturizer question one account used only US publications both times, with no source from the UAE; the other used Gulf News and UAE pharmacies both times. The brands named were nearly the same. Two repeats are not enough for a rule.
- A new local source entered within a week. Three articles by a Dubai clinic dated 18 October were cited on 25 October. This is one case, not a timetable.
Caution
What this does not claim
No positions
One general question returned three different first picks in three answers on the same day. A rank in an AI answer is not a stable quantity, and nothing here should be read as one.
No verdict on any brand
Brands are named because the engines named them. This page records what was said and where it was taken from, not which product is better.
A small sample
Two dates, six answers per question, one non-UAE IP. Enough to see which patterns repeat; not enough to put a percentage on the market.
Repeat it yourself
The five questions above are written out verbatim. Run them in a temporary chat on two dates and compare the labels and the sources, not the order.
The same work on your own category — which sites build the answers to your buyers' questions, measured on fixed questions across two engines — is the category source map.
alex@citedlayer.com