The traditional way of setting up a data warehouse or data mart started in the boardroom, not at your desk.
That's because the capital costs involved meant serious oversight was necessary before the investment could get off the ground. It also meant that only a "sure thing" would get past the scrutiny of those holding the purse strings - and with that scrutiny, time would pass. What was a potentially great idea which depended on data might simply fall by the wayside.
And that goes to the heart of the great value proposition offered by cloud data warehousing. With the costs substantially reduced and the resources required to set it up consisting of little more than your mouse (and a few clicks thereof) and mere minutes to get it configured, any opportunity can be evaluated. Doesn't work? Doesn't matter, shut it down. Does work? Brilliant, go tell the boss we're onto something.
Cloud data warehousing provides the ability to set something up very quickly. What was once a matter of days, weeks or months is now mere minutes. Where once a capital budget would have to be approved and heavy-duty hardware purchased, installed and housed, it now happens in the cloud. In the on-prem world, if the initiative proved unsuccessful, that equipment doesn't just disappear; it has to be reallocated at best, or it will gather dust until a use is found for it, at worst.
The flexibility afforded by the cloud goes even further. There's no lock-in to the database type or even the hardware type. Need a relational database? No hassle. A columnar one? Go get it. There's no need to obtain specialised hardware or software or appliances with expensive licensing. Just go right ahead and spin up whatever the job calls for and get busy. No expensive planning, no big project, no need, even, for pilots or proof of concepts.
One of our forward-looking customers has jumped on the opportunity presented by cloud data warehousing. Business people have been given the ability to experiment and create data resources themselves and pay for them out of a departmental budget. IT barely needs to be involved at all. There are some controls, as is natural, but this company is now in a position to tap the great data ideas of its people.
You could drive a bus through the massive gap between how much data is being used (not much) and the expected value which data is expected to generate (a lot). That gap can be explained, at least in part, by the costs previously associated with creating the infrastructure on which to conduct analysis. Put simply, it was cripplingly expensive.
And that's precisely the problem that cloud data warehousing now solves. The technology is sufficiently mature, multiple services are available on top and the connectivity to access and use it effectively is in place.