There are different types of fact tables in a data warehouse, such as Transaction, Periodic Snapshot, and Accumulating Transactions. However, a Factless fact table is distinct from other fact tables because it does not store numeric measures or facts. Factless tables can be used to understand the occurrence of events. Businesses use them to analyse events that did not occur, such as identifying products that remained unsold during promotional dates or pinpointing customers who did not visit the store.
Businesses often want to know ‘What didn’t happen?’ as it enables them to plan better strategies to improve their sales and, eventually, profits. The data warehouse must have provisions to provide statistics for such scenarios.
In the simplified design of the ‘CustomerVisit’ Factless table mentioned above, only dimension columns serve as foreign keys, referencing Date, Customer, Store, and Promo dimension tables. Unlike other tables, it doesn’t store numeric measures such as ‘sale amount.’ Despite this absence, business analysts can gain meaningful insights, for example, they can identify products that had no sales during promotional dates or within a specific range. Gaining insights into this question allows business owners to plan promotions for those particular products, such as through advertising or offering more discounts. Another valuable insight is knowing the stores that had no sales during the promotional dates or within the specified range. By understanding these patterns, businesses can strategize and optimize their sales efforts.