A company running such a car portal website could store the HTTP access log in a database table. This can be used to analyze users' activity. For example, to measure the performance of the application, identify the patterns in the users' behavior, or simply to collect statistics about which car models are of the most interest. This data would be inserted into the table and never deleted or changed, or maybe deleted only when it is too old. However, the amount of data would be much bigger than the actual business data in the car portal database, but the data would be accessed only from time to time by internal users for performing analysis and creating reports.

These users are not expected to execute many queries per second, rather the opposite, but those queries will be big and complex, therefore the time that each query can take matters.

Another thing about OLAP databases is that they are not always up to date. As the analysis is performed on a basis of weeks and months (for example, comparing the number of requests in the current month with the previous month), it is not worth it to invest the effort to make the data of the very last second available in the data warehouse in real time. The data can be loaded into the data warehouse periodically; for example, daily or several times per day.