Twitter Weekly Updates for 2009-12-20
- I just became a Vokleite at http://vokle.com #Vokle #
What is Business Intelligence?
A quest to define the broadest term in business…
The nature of Visual change is you probably didn’t notice
An interesting study was done by Harvard to quantify how often we notice visual change. A quick summary is that they had a series of people enter a room and ask for a form to sign. The man that they asked for the form knelt down behind the counter to get the form, and a new person stood up to hand the form to the requester. The second person had on a different color shirt, totally different hairstyle, and was just obviously different. The amazing part is that only 25% of the time, when asked, did the subjects notice that the person changed.
That means that 3 out of 4 people would not notice if a person was replaced by a different person in a split second. Besides being incredibly humorous, it leads us to some interesting conclusions on the subject of visualizing data.
By nature we tend to scan reports, charts, etc. I wonder how often that change in one of these data devices might be totally missed by the observer. The common solution to this type of scenario is exception reporting in the form of alerts, notifications, events..whatever work you choose to use, the point is that instead of showing me everything, only show me what’s wrong.
That might be great if you only follow a few things. I am not sure how your job works, but I have a thousand balls in the air at any time. If I had an alert for each ball, I would go from trying to find a data point in a sea of numbers on a report, to trying to find a data point in a sea of alert notifications. The solution now mimics the problem.
This is where you might expect me to offer some form of solution or start my sales pitch for a product that fixes this. I don’t do that here, I ask questions. The questions that I would like answered are:
- What do you feel the most useful format to display observable change is?
- Where in your daily work life would you expect would you expect to see this display?
- When you notice this change, what activity would be your first reaction? (email, im, phone call, analysis tool, etc.)
Analysis, the beginning or the point?
If I were to define Analysis, it would be something like:
Analysis is the process of breaking a complex topic or substance into smaller parts to gain a better understanding of it.
That’s a pretty good definition and we should all thank Wikipedia for it. In the world of business and more specifically business intelligence, that definition gets a bit more interesting. There are at least two distinct ways that Analysis is used in business and I would like to break both of those down for a moment.
Responsive Analysis
This is pretty obvious. Your business is underperforming in some key area, and it’s not entirely obvious why. This type of analysis is used to dive into the data that is generated as your business runs, to find the anomaly that caused the overall heath of the business not to achieve what you assumed it would. Imagine following something as simple as sales, and identifying that it’s underperforming, but not knowing obviously why. You would most likely take your army of analysts, give them as much context as you can, and have them dig into the data to try and identify the root cause of the failure. It makes perfect sense.
Predictive Analysis
Predictive is a little different. In this case a person (or potentially an intelligent system) is monitoring an aspect of your business that they have particular knowledge of. In the name of always improving performance, the analyst is actively looking for problems anywhere he or she can to help tune the system that existing metrics assume is health. This is the type of analysis that takes a business and squeezes more performance out of it.
The Scenario
Both of these types of analysis are very valid scenarios. They differ mostly on the initiator and the intent of the analysis. In the sense of Responsive Analysis, it’s important to be able to react to events and adjust in-flight with the least amount of disruption and while keeping the maximum amount of context you can.
The world of predictive analysis is much more interesting to me. This is where artificial intelligence, pattern recognition and self-tuning software have domain to add value.
Imagine that your business is a hospital. For altruistic reasons, the primary metric that determines your success is the number of days from diagnosis to release.
When your average diagnosis to release goes over 5 days, the people on the top are warned and everyone swarms to try and improve it from many different directions. That’s kind of how BI works today.
Now imagine if the nurses, doctors, specialists and even the people that run the cafeteria understand how the smallest movement that they make effects the business as a whole. The lunch lady installs an antibacterial goo station at the lunch line and somehow ties that to the diagnosis of flu symptoms of staff.
The nurses all decide to hold an internal clinic to teach advanced medical techniques that can decrease the length of hospital stays by some percentage.
The Questions
The question is simple. In this scenario does one benefit from the use of analysis at all levels? Is there a point when this type of analysis becomes important? Does it ever cease being important and you begin simply a monitoring activity?
What does this kind of BI look like? How does one understand his or her contribution to the machine as a whole?
What other advantages does this knowledge give the users at all levels?
Beat up this scenario. Think outside of the box and tell me what you really think.
Want a job? IBM says Analytics is where you want to be
This article I ran across from BusinessWeek is a very interesting introduction to the topic we are discussing. The historical perspective is that there are these number ninjas we like to call Information Workers, and among that subset of users is yet another slice of power users. These users we traditionally think of as being super savvy, have complex skills like the ability to construct complex queries and understand the nature of database systems, and techie enough to use traditionally complex tools used to get the information out of these systems and do something with it. (notice I said Information and not data, that is intentional)
The article goes on to say that the field is growing so fast that there are not enough qualified candidates to fill the positions. The solution IBM seems to be taking is to get involved in education and help new prospects understand the nature of this analytical thinking to create it’s future pool of prospective employees. Of course helping the educational system is never not a noble goal, but it is a little unique from the point of view of a software vendor in this case.
So let’s use this article as a jumping off point to another set of questions that have been rolling around my mind for awhile. Lets take on the persona of someone in middle management. We are trying to find employees to help us make use of our mountains of data and better understand our business. I wonder…
- What are some of the minimum skill sets that prospective employees need to possess?
- What tools should they come equipped to use?
- What makes them special as opposed to someone who is just designing and generating reports?
What is Business Intelligence?
While I will admit to creating this blog on a whim, it does have a very good purpose and I hope you agree and become involved as it grows.
In my two years on the Business Intelligence team at Microsoft, responsible for the PerformancePoint Server/Services product, I have learned quite a bit about BI, but the thing I have learned above all, is that in the world of the user, I know very little about Business Intelligence.
The problem with defining and explaining business intelligence is that so many people do BI-type activities, but never equate them to being BI activities. Business Intelligence is seen as the playground of Cognos, SAP and of course Microsoft, but don’t feel that it’s also the world of Excel, Access, Quicken or Mint.
Wikipedia says Business Intelligence (BI) refers to skills, processes, technologies, applications and practices used to support decision making.
The purpose of this blog is to seriously engage with people and find out how they work, how they think and what leads them to make decisions.
The questions I would like to tackle are things like:
- What activities constitute Business Intelligence to you?
- What do you do today that would benefit from the services that BI tools provide?
- Where do you expect to be able to do Business Intelligence type activities that you cannot today?
Other things I would like to know more about are:
- What do you want to be able to do that you can’t do today with existing tools?
- How do you want to share the results of your work?
- What tools do you use today, and what are their strengths and shortcomings?
I’ll fill in the gaps with some conjecture on how I feel about the above topics. I’ll reprint responses I get and solicit more feedback on my thoughts.
Let me place one last constraint on this blog. While the topic may come up from time to time, this is not a Microsoft blog. The purpose is to talk about Business Intelligence as a concept in abstract. I will not blog about Microsoft product features, product futures or even competitor products unless it is an analysis of a particular type of feature or activity in general.
I hope you enjoy this journey, get actively involved in the debate, and maybe by the time blogs are replaced with some other hot new thing, we’ll all know what the heck BI is…
A Little Context
That's all we want right? And that's part of what BI means to me...context. But that's just part of the picture. What does it mean to you? What is Business Intelligence and when do you do it? That is what I hope to answer by starting a dialog on this blog.
What I'm Doing...
- I just became a Vokleite at http://vokle.com #Vokle 2009-12-19
- PerformancePoint, Cognos, Informatica, Proclarity, MicroStrategy, Hyperion, Panorama, SAP BI.. Is BI expensive? Is BI complex? Define BI... 2009-12-10
- If you had to choose one single activity that encompassed the meaning of Business Intelligence, what would that one activity be? 2009-12-10
- More updates...