Join Alan Simon for an in-depth discussion in this video Assessing the state of your data and analytics, part of Transitioning from Data Warehousing to Big Data.
- You should begin your journey of transitioning from traditional data warehousing to Big Data by understanding how well you do different types of analytics today, and then determining what role Big Data should play in delivering those analytics, as well as, new capabilities tomorrow. And then, once you have this information, deciding how to get to where you want to be. The secret to getting on the right path though, is to spend enough time and effort to first understand where you're at today. Undoubtedly, many people all across your organization have an opinion about where you are today with analytics and data management, and it would be very easy to collect a mountain of anecdotes and sound bites.
But the best way to understand your current state of analytics and data, is to grade it. Here we see five different categories, our four types of analytics as well as our overall data architecture. What we want to do is, for every single one of them, grade what is working well today, along with what is not working so well. Within each of these categories, we see the same set of factors: how complex or how overly complex each is in terms of delivering those capabilities, quality, timeliness, and then the effort that goes into supporting each of these categories, as well as our overall data architecture.
What we want to do is gather enough input to where we can identify what we will call "hotspots", areas that are specifically problems that we need to address. You see in this example, our descriptive analytics, our classic business intelligence, is relatively free of hotspots other than we have some issues with timeliness, our data architecture has a couple good factors and a couple problem areas. But the rest of our analytics are filled with hotspots, and, therefore, areas we need to address.
The way we conduct the assessments is to use a simple scoring mechanism. I like to use the scale between the scale between one on the low end and five on the high end, but you can use any other scale that you'd like. Select where your hotspot cutoff is going to be. For example, three, or about halfway, or a lower score will signify that things aren't working as well as they should be. From that scale, you want to select the level that you'll use for your hotspot cutoff. I like to use three, about halfway through, or lower.
Make sure that you survey a large number of people from the business and IT sides of your organization, all the way from the executive ranks down to people who work with data and reports on a daily basis. And then finally, consolidate, compile, and then publish the results so this way you'll be able to start building your future state and your road map based on where you're at today. Beyond the individual findings, what you also want to do is look for patterns in where the hotsopts are and that will also help you to decide what your future state of data management and analytics should look like, as well as your road map to get there.
Are you seeing a lot of hotspots? That almost certainly means a greater need to overhaul your environment with Big Data playing a central role. Do you have fewer hotspots for descriptive analytics? Your BI capabilities are probably very effective but your other categories of analytics need to catch up. If your predictive and discovery analytics are full of hotspots, Big Data will likely help you address those shortcomings. On the other hand, if your predictive and discovery analytics actually score better than your business intelligence, your descriptive analytics, you really need to focus your attention on upgrading your fundamental reporting and dashboards, and then the underlying data architecture.
If timeliness is a consistent problem across all of your analytical categories, the velocity that comes with Big Data will help you address those issies. Finally, if you're fortunate enough to show only a handful of hotspots, you still will be able to benefit from Big Data but you can focus your initial efforts on taking advantage of new opportunities, rather than solving critical issues. Make sure that you analyze your results thoroughly, because this way you'll be able to build your future state of Big Data and your analytics grounded in the reality of where you are today and what you need to address.
- 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