Most gemstone shoppers start with a colour. Blue, green, red, purple. The species follows the colour, not the other way around. A blue sapphire is sapphire, but so is yellow sapphire and pink sapphire; a green tourmaline shares a name with a black tourmaline but they are very different gems for a buyer. A library that indexes the corpus by gem-and-colour pairs ranks for the queries shoppers actually run, around 100 gems crossed with their colour ranges.

SleekRank reads the gem-colour sheet as the schema. Columns for slug, gem, color, variety_name, mohs_hardness, refractive_index, specific_gravity, treatments, price_band, and image_url feed the base page at /gemstones-by-color/{slug}/. Selector mappings fill the gemmological spec table, list mappings render the treatment notes, and a meta mapping wires the OG card.

Cluster blocks at the bottom of every page surface other gems in the same colour family (other blues, other reds) and other colours of the same gem (other sapphires). Both are list mappings against the same sheet, returning six related entries per page deterministically.

Shoppers don't search for sapphire. They search for blue sapphire, or for a green gemstone, or for an alternative to emerald. The mental model is colour first, species second, and the sites that index their content the same way rank for the queries shoppers actually run.

A species-indexed library buries blue sapphire, yellow sapphire, and padparadscha on one big sapphire page, and the colour-specific queries land somewhere in the middle of a long scroll. A gem-and-colour indexed library has each variety on its own page, with the gemmological spec, the treatment disclosure, and the price band laid out cleanly. Building 300 to 500 such pages by hand is realistic but slow, and the values drift between pages as authors change.

Building them from a sheet keeps the corpus consistent because hardness, refractive index, and treatment vocabulary all flow from controlled columns. The other reason gem-and-colour pages benefit from data-driven generation is the cross-linking. A buyer reading the blue sapphire page wants links to other blue gems (tanzanite, blue spinel, paraiba tourmaline) and to other sapphire colours (yellow, pink, padparadscha).

Both lists are list mappings against the same sheet, generated automatically and kept current as new rows are added. The buyer browses sideways through the corpus, and every related-page block stays fresh.