In the scenario of using BigQuery for a multi-region dataset that includes a table updated several times daily with daily sales volumes, what is the most cost-effective method to protect the sales table against regional failures, ensuring a Recovery Point Objective (RPO) of under 24 hours?