A few weeks ago, I finished a speech at Netflix and got into a long discussion with a business information manager from a major technology company. This fellow asked me the best way to get value from Tableau with his data cubes. I told him to get rid of the data cubes and buy a good columnar analytics database, specifically Vertica.
I have this conversation at least every other week.
Above: Me speaking to the Bay Area TUG at Netflix HQ.
Here are three good reasons you should replace your data cubes with a high-performance database, especially if you plan to use Tableau Software to analyze your data.
1. You’re Wasting Money on 1980’s Technology
That’s right. I said cubes are 1980’s technology. That was the last time I remember thinking data cubes were cool. When I was sitting in a hotel on my 2400 Baud Modem, it was amazing how fast data cubes were. But this is the 2010’s. You can do better.
Above: These things were also cool the last time data cubes were cool.
2. You’re Building Data Structures that Eliminate Data
Cubes aggregate data, which by definition means that they remove potentially useful granular detail from view. As someone who loves ad hoc performance analysis on large data sets with Tableau Software, there is nothing I hate more than dealing with a data cube. The MDX coder decided how much granularity that I’m able to see. Also, the data hierarchies defined are an automatic downgrade versus what you can do when Tableau is connected to a relational data source.
3. You’re Wasting Time by Building Cube Structures
You don’t have to pay your persnickety and high-turnover MDX coders to spend weeks building a soon-to-be obsolete structure, making your slow database technology faster by eliminating data. Instead, invest in a high-performance column-store database. My preference is Vertica, but pick any one of them. You’ll be glad you did.
Learn More, Contact Us!
If you have any questions about what, why, when, how or who, contact the BI Services Team at InterWorks. We can help.