World Brand Library

The company is the row. Click one to open it and see every mark it owns, with the marks' own logos. Adidas AG is one line holding 20 brands rather than 20 lines repeating the same company, and every figure on the line — country, revenue, headcount — describes the company, because that is the only level at which those exist.

The name search matches the company, its other names and its brands, so searching Nivea finds Beiersdorf. That is the point of looking at it this way round: the brand list cannot answer "who owns this" without already knowing the answer. Sorting by revenue or headcount puts the unknowns last rather than treating a missing figure as zero, since most of the library has no financials and they would otherwise bury the rows that do.

A company with no logo of its own shows its flagship brand's, marked from brand, because for most companies that is the mark people actually recognise: nobody has seen the Beiersdorf wordmark and everybody has seen the Nivea roundel. Nothing is copied onto the company — switch to Brands to edit a single mark.

The library is filled by a cron on this host: a nightly pass that re-reads the rows nobody has confirmed in longest, a daily pass over what changed at the source in the last week, and a weekly walk that keeps growing the library from where it last stopped. Wikidata is the only source running today; EUIPO TMview and WIPO need credentials that have not been issued, so they are wired up but refuse to run. Starting a scrape by hand does the same thing the cron does, immediately, and only one may run at a time.