OliveYoung Data Scraping

Launch cadence is the metric, and most feeds are built for price

K-beauty product cycles run far shorter than Western beauty. A brand can launch, peak and be replaced within a period where a Western equivalent would still be in its first campaign.

That changes what a useful dataset looks like.

• New arrivals are the leading indicator, not price movement. What appeared this month predicts what travels to Western retail in the next year.
• Assortment turnover — how much of the range changed — is a stronger competitive signal than average category price.
• Delisting is fast and normal, so a product disappearing is routine rather than notable, unless it was performing.

We record first_seen and last_seen per product and per shade, plus a derived days_since_first_seen, so new-arrival tracking and turnover are queryable rather than reconstructed.

One honest caveat: first_seen means first seen by us, not launch date. On a feed started last month everything looks new. The archive_limited flag marks records where that applies, so lifecycle analysis does not silently understate product ages.

Shade level, and where the gaps actually are

A product-level availability figure on a beauty retailer is close to useless, because a foundation or cushion available in six shades and out in four is not simply "in stock".

• Shade-level availability is the field that matters, and it is where range gaps show.
• Deep and light extremes go out first and are restocked last, so a product-level figure systematically hides the shades most likely to be missing.
• Not ranged is different from out of stock. A shade never offered in Korea is a range decision, not a supply problem.

We deliver one record per shade with shade_offered and shade_in_stock as separate states, and aggregate shades_available_ratio so both views exist. This is the same discipline our Nykaa service applies in India.

Claims are captured, never verified

K-beauty listings carry dense ingredient lists and efficacy claims — brightening, barrier repair, percentage concentrations of active ingredients.

We capture all of it exactly as published, and set verified_by_actowiz to false on every claim field.

Why we will not verify

• Ingredient verification requires testing, which is a laboratory exercise rather than an extraction one.
• Efficacy claims are regulated differently in Korea, the EU and the US, so the same claim can be compliant in one market and not another.
• A verified flag we cannot stand behind would sit in your data looking authoritative.

Capturing claims is genuinely valuable — tracking how a brand's claim language changes as it prepares for Western distribution is a real finding, and it is exactly the kind of signal a brand entering the US or EU wants. We just do not assert that any claim is true.

Reviews

Review content, ratings and timestamps, with no reviewer names, profiles or purchase histories. Korea's PIPA is strict and the boundary is the same as everywhere else.