Fashion & Apparel Data Scraping

Because the unit that sells is a size. A style with only size 16 remaining at full price reads as in stock and not discounted, when in fact it has sold out of every core size — the opposite conclusion.

Style-level data also makes markdown analysis misleading: marking down a style with a full size curve is discounting a genuine slow seller, while marking down one with a broken curve is clearing a tail. Same markdown depth, opposite meaning, and only size-level data distinguishes them.

Not units, but a strong proxy. We track per-size availability over time, so core sizes disappearing while price stays full is a reliable indicator of sell-through, and sizes returning to stock indicates a repeat order — which is the closest you can get to observing a buying decision from outside.

What we cannot see is unit volume, inventory depth or margin. We report availability transitions rather than sales, and we would rather be explicit about that than let a proxy be mistaken for a measurement.

Sizes are captured exactly as published, then mapped to a normalised scale per category and market with the original retained. Footwear is handled on its own scales, and extended, petite and tall ranges are treated as separate size runs.

Where alpha-to-numeric mapping is ambiguous — and it usually is, since a retailer's M is not a fixed numeric size — we keep both and flag the ambiguity rather than committing to a conversion that may be wrong for that brand.

Then we say so per retailer rather than substituting style-level stock. Some retailers expose size availability only inside an add-to-basket flow, and we do not add items to baskets or create accounts to reach it.

During scoping we tell you which retailers in your competitive set expose size-level data publicly and which do not, so you know the coverage before committing rather than discovering gaps in month two.

Yes, provided we were collecting when the style launched. We record first-seen, weeks on site, each markdown event with its date, and the size curve state at each step.

The honest constraint: for styles that launched before our collection began, first-seen is when we first observed it, not the true launch date. We report the archive start so you know what weeks-on-site is actually measuring rather than assuming it covers full lifecycle.

As published, yes — material composition, recycled content claims, certification mentions and any sustainability labelling the retailer displays. We capture the claim as made, with the wording retained.

We do not verify claims, and we are careful not to present captured claims as validated. For teams doing substantiation work, the published claim plus its wording and date is the evidential starting point, not the conclusion.

Yes, and joining them is increasingly requested. Resale listings are collected with condition and pricing, and joined to primary retail styles where identity can be established.

Matching is harder on resale because listings are user-written with inconsistent brand and style naming. We attach match confidence and flag uncertain links rather than asserting that a resale listing is definitely the same style as a primary retail record.

Daily as standard, because size availability changes daily and the signal degrades quickly if missed. Sub-daily is worth it during sale periods and around drop launches, where availability can change within hours.

Weekly collection is largely pointless for size curve work: you will see a style went from full curve to broken without seeing the sequence, which is where the information is.

We quote individually. The drivers are retailer count, category scope, and critically whether size-level collection is required — size curves multiply record volume by the number of sizes per style, which is typically six to twelve times style-level volume.

A defined category across several retailers at daily size-level refresh sits in the middle. Full-catalogue coverage across many retailers with sub-daily sale-period collection sits higher. One scoping call, a free pilot on your own competitive set within 48 hours, then a fixed monthly quote. Request a quote.