Join Alan Simon for an in-depth discussion in this video Taking a critical look at BI, part of Transitioning from Data Warehousing to Big Data.
- A good place to start to understand the limitations of our traditional technologies in data architecture, is to explore what they actually do well, and then we'll take a look at where difficulties and complications start to set in. Let's start with Business Intelligence, which gives us five significant advantages over the old-style techniques from before the 1990s, that we looked at earlier. First, Business Intelligence provides us with a slicing and dicing of our data. We can look at our data a number of different ways, where we see certain key measures, and then look at it by various dimensions.
Time, by product, by customer, or however anybody want to see that particular data, as long as the data warehouse has been set up that way. Second, Business Intelligence has a number of different very mature tools from leading vendors, and many people have a lot of experience with those tools, and those tools have been evolving for a number of years, and they're very powerful. All of the technology associated with Business Intelligence, not just the tools and products, but the surrounding ecosystem and supporting technologies are incredibly powerful and mature as well, and have been around for a number of years.
Fourth, we also have many years of best practices to draw on, what works and what doesn't work when it comes to successful Business Intelligence. And finally, we have many skilled Business Intelligence and data warehousing professionals to draw on whenever our organization needs to build new capabilities. Let's take a look at what a Vice President of Sales would need when it comes to data-driven insights for business intelligence and analytics. Remember earlier, we mentioned that traditional business intelligence today is widely known as Descriptive analytics, in which we take a look into the past or into the present, and questions such as these that you see here, such as the top and bottom account managers of last quarter and these other insights would be good examples, of Descriptive analytics, that a sales vice president would want and need to know.
At the same time, our sales vice president also needs to have a look into the future. Predictive analytics. Again, we don't have a crystal ball, but by using our data, and very sophisticated quantitative techniques, we can help figure out what's happening with our products, with our salespeople, with our pricing, and make accurate predictions as to what is likely to happen, and then act accordingly. Our sales vice president also has a strong need for Discovery analytics, where again, we aren't asking specific questions, but we are using the power of our technology to comb through mountains of data, and look for patterns and indicators that we may not necessarily be aware of, such as those that you see on this slide.
Despite the fact that our sales vice president needs descriptive, predictive, and discovery analytics, what typically happens, though, is that many organizations stop short of building out their analytical capabilities and never get much further past descriptive analytics, because the descriptive analytics tend to be the reports that have run the organization for many years. When we deal with predictive and discovery analytics, our data warehousing technology is typically not architected to support those capabilities.
Other organizations, though, do find a way. What many organizations do when they build data warehouses, they will bring in the source data like they would for any other data warehouse, and out of that environment, they will produce the descriptive analytics, the business intelligence, such as the reports, and the online analytical processing, the dashboards, and the visualizations, and everything else they need. At the same time, though, what the intention is, they will pull some of that data from the warehouse into some sort of a separate Analytical Data Store, and that's where they will run the predictive and the discovery analytics.
So what they intend to do, then, is build a well-architected environment where data flows into the warehouse and then downstream, for the specific uses. What often happens, though, is that once the Analytical Data Store is built, the Data Warehouse itself is actually bypassed, and data flows directly into the Analytical Data Store, which on the surface may not necessarily seem to be that much of a problem, and this is where we start to see the inconsistencies and fragmentation in our overall data architecture.
The end result is that despite our sales vice president, or any executive or manager, needing a broad range of business intelligence and analytical capabilities, the traditional approaches to BI only deliver some of those critical insights, and therefore, we need to look at other techniques to get access to the predictive and the descriptive analytics that we need.
- Exploring big data, Hadoop, and analytics
- Examining the shortcomings of traditional data warehousing
- Comparing big data architectures for next-generation data warehousing
- Understanding alternatives
- Building a roadmap
- Managing big data-driven projects
- Monitoring and measuring success