I recently visited California to meet with one of our clients the Golden State Warriors to talk about the analytics we ran for them. Dennis is at the top of the field. He drew the diagram above out for the head of analytics at the Warriors while I tagged along.
Their head Analyst was talking about wanting to set up a “data warehouse”. After a long discussion on this, like really long, I learned that data warehouses may seem like a good idea, but for most companies it is not really needed.
Let’s look at the diagram. Dennis drew this out because he could see that their discussion about whether a data warehouse was necessary or not wasn’t hitting through. So we went into a conference room and he drew it out.
There are three columns, and although they’re labeled in the drawing, they’re kind of hard to see so I have labeled them down below the drawing. These three steps are MAA: Metrics, analysis, and action.
In the left column is a visual of how information flows in a business with a data warehouse. A data warehouse is put in place so that when someone needs information (such as the number of tickets sold last year, how many per game. Or the number of ice-cream cones bought. Whatever the business and the need is) they can pull it from a location or “data warehouse”.
This information is our Metrics, or the numbers. This way they don’t have to go to different sources for information but can pull it out from one spot. This can be helpful to keep data reliable. If you have no data warehouse the risk is that in people will come up to the source for information and with a number of people asking for this information, numbers can be skewed.
This is the Metrics of this process. With a data warehouse Information is more dependable and more easily accessible.One can take this information and do some analysis on the numbers received (middle column). Then one can make special offers based on analytics ran for example to receive greater business.
For example, we can see what Golden State Warrior fans are commenting on and what their similar interests are in order to know which market to target. Targeted marketing done right can lead to greater conversion or greater sales and results in lower acquisition rates. When this special offer is made or something is done based off the numbers evaluated, this is action.
So why shouldn’t every business pay to have a data warehouse? Doesn’t it sound great? Well if the objective is to boost sales then the more powerful route would be to take the information, analyze, and do something with the change in numbers that will lead to conversion (someone acts on the offer), a data warehouse may not really be needed.
After this log discussion I could see that Brandon, the head Analyst was really stressing over this. His bosses wanted the warehouse of data, but he was starting to see that it was expensive and not really needed. The companies that seek data warehouses usually don’t need one. Perhaps large companies dealing with huge amount of information might direly need one but in most cases- it probably isn’t necessary.
It costs a lot to set one up and can be a lot more trouble than its worth. If one can take the information and immediately convert it to sales through measuring, analyzing, and acting then that is the ideal process.