We read the public product feed of 311 independent beauty brands that sell direct through their own Shopify store, every variant, once a day, and keep the record. Not Sephora, not Amazon, not the department store shelf: the brands themselves, at the source. Yesterday's price is gone the moment it changes; ours stays.
39.67 percent of all tracked variants sit in the largest store, skinnydiplondon.com: 67,784 of the 170,849 variants tracked on 8 September 2026. That share is measured across every store observed that day, exactly the same set the counts above are drawn from. Every global number on this page carries that weight inside it, and we say so rather than let an average hide it.
You know your own prices. You do not know that a rival cut a hero product by 30 percent for nine days and then put it back, or that their bestseller has been out of stock since Tuesday. We do, because we wrote it down every day, from the same public pages your customers see.
| Every Monday | One page on the skincare category: which stores moved, how deep the discounts went, what went out of stock and what came back, with the observation date on every figure. Skincare Pulse, 99 USD per month. |
| Once | Your store against five measured rivals: every recorded price and stock change since 24 August 2026, row by row, with thirty rows re-checked against the live stores on the day of delivery. First report, 300 USD. |
| Always | We record what happened on public pages. We do not infer why a price moved, and we do not guess sales. If a figure is not in the archive, it is not in the report. |
Anyone can open a store and read the price. Nobody can open a store and read what the price was three weeks ago, how long the discount ran, or which products quietly went out of stock. That record only exists if somebody was writing it down. We were.
| Changes recorded | 40,523 | Price cuts, price rises, out of stock, back in stock, discount starts and ends, new variants, across 311 stores |
| Busiest day | 5,312 | 3 September 2026, changes in 111 stores on a single day |
| Quietest day | 869 | 30 August 2026, a Sunday, changes in 66 stores |
| Held in quarantine | 18,964 | Raw change records excluded from every number on this page, kept in the archive, never published |
Prices are published because we would rather you decide than negotiate. Payment is by card through Stripe. Invoices are issued by SmartikaGlobal OÜ, Estonia. Every price below is a founding price, locked for twelve months from the day you start.
We publish the method because a data archive is worth exactly as much as its weakest claim. If you are going to make a decision on our numbers, you should know how they were made.
| Source | Public product catalogues only. No login, no account, no private endpoint. |
| Frequency | One full pass per store per day, every day, since 24 August 2026. The latest pass read 311 of 311 stores. |
| Granularity | Variant level: price, comparison price, availability, product type, vendor and SKU as the store publishes them. |
| Quality control | Fifty units re-checked against the live store every morning. Latest audit, 8 September 2026: 144 of 144 manually audited fields passed, across 48 of 50 sampled units; the two units outside the denominator were pulled from their store between snapshot and check. |
| Quarantine | Two fixed rules, published above, hold back price changes outside a market range and records with no real transition. Held back records are counted and kept, never published and never deleted: 18,964 at the latest count. |
| What we do not claim | A variant that vanishes from a catalogue entirely does not yet produce an event. We record observed changes; we do not infer the reason for a price move. Both statements appear in every report we deliver. |
| Robots and removal | We respect robots.txt. Any retailer can ask to be excluded and is removed within 48 hours, permanently. |
If your store is in the archive, you can see your own numbers, ask for the underlying data, or ask to be removed entirely. All three take one email, and removal is free and permanent.