You can think of tokens as place-holders, which will be filled with actual data from the right context (shop, product, variant…).
There are 2 types of token, static and dynamic.
Static tokens have fixed name, e.g {domain}, which will be filled with the current domain of the shop, or {product_title}, which will be filled with the product title in context.
Dynamic tokens include product attribute tokens ({pa_} tokens) and product metafield tokens.
Product attributes tokens are named after the product attribute (option) that it take value from. For example {pa_color} or {pa_size} will take the value of the option Color and option Size of the variant in context, respectively. That also means if your product stores the color of a variant under a different name, e.g. Options, Colour or 顏色 (color in Chinese), you need to change the token name accordingly, e.g. {pa_options}, {pa_colour} or {pa_顏色} accordingly.
Product metafield tokens are used to pull in data stored in product metafields. These are used to get SEO title/description of the products (global.title_tag and global.description_tag). Metafield tokens are also helpful if you want to make use of product Google Shopping data stored as metafields (e.g. Shopify Google Channel app stores product category as mm-google-shopping.google_product_category, age group as mm-google-shopping.age_group). Some de-facto metafields can be selected from the dropdowns at the end of the inputs, but any metafields can be included using the format {namespace.key}.
Token name are always in lower case.
