Ask An Expert

Can I really replace a data warehouse with…nothing? | US-Analytics

David Kebo
January 24, 2020

I've been talking to a number of clients these days that are curious about the common QlikView claim that no data warehouse is needed to use it right out of the box, even if you're wanting to look at information pulled from multiple transactional systems, web services, flat files, Excel spreadsheets, etc. They're mostly curious because they're used to hearing "oh, you want Business Intelligence, so first you should build a data warehouse." Or they're curious (and skeptical) because they know how untidy data usually is coming from these types of systems, and one of the main functions of a data warehouse is to tidy data up (and then there's aggregation, data marts, indexing, cube-building, defining drill-down paths, etc.).

Here's the interesting reality. You don't need a data warehouse. And you do. And sometimes something in the middle works just fine. How can this be? Well, simply depends.

Scores of companies have avoided building a data warehouse (for the purpose of supporting their BI initiative) by implementing QlikView. Scores of companies have used QlikView as the information delivery part of their existing data warehouse architecture. Scores of companies have decided that even with QlikView they need to stage data in a relational database and build a dimensional data model. Each of these scenarios and resulting architectures is the result of that situation's business and technological requirements.

So no answer is universal - your mileage can and usually will vary. This is partly because of the unexpected (for many) ways that QlikView eliminates a lot of the complexity in the underlying data structure normally necessary to build BI. But QlikView simplicity can't always negate the needs of Master Data Management or data governance, for example, which may in the end dictate an architecture which addresses those needs. But sometimes it can.

You may find that you can avoid building a data warehouse with QlikView. Or you may find you can replace yours with QlikView. Or you might find that while there's no avoiding a data warehouse, there are ways for it to be constructed in a fraction of the time and deliver value very quickly to the business.

The answer for you and your situation is in the subtlety between what you need and what is possible. Even in the IT world, there's a lot more grey out there in a world you might assume is black and white, especially when new paradigms collide with old ones.

All Analytics

Want something amazing in your inbox?

Subscribe to our newsletter below