I was looking at a few SaaS analytics products and was wondering as to how these companies were storing data from their customers/clients.

For e.g., Company A has PostgreSQL and Company B has MySQL as their databases. The SaaS product company is able to store data of both Company A and Company B and provide analytics. Even in the case where Company B has PostgreSQL as its database and the schemas are not the same, still the SaaS product company is able to provide analytics.

My question is - How does one (a) source (b) store and (c) analyze data coming from diverse companies/businesses having different/disparate/dissimilar databases in a scalable and reliable manner? What would be the methodology to follow?