Wednesday, August 31, 2011
As a former principal consultant at PeopleSoft/Oracle, I am proud to discuss this year at OpenWorld something we do as consultants for large organizations. This presentation was originally given during a state government implementation. State government reporting is like doing an implementation for 50 different major companies simultaneously. They all had complex statutory obligations, different legacy tools, and different user communities. The leadership of the implementation team wanted to better understand how OBIEE would be able to schedule reports, handle security for such a massive audience, integrate with their portal and ERP system, and what were the benefits of moving off the old tools such as SQR and Crystal. As our team prepared the presentation and were doing rehearsals, we kept switching from one persona to another, saying that the administrator would do this, and then the end user would see this, and these features would be like what an operational user would see, but a manager would want to see something else. My manager said, "why don't we split up this presentation into segments by four different personas." When we did this, it changed everything and made the whole thing so much clearer.
Who are the four personas (four user types)?
The four user types are operational, manager, power developer, and system administrator. There are actually a lot of user types left out of this presentation, including the executive, the performance troubleshooter, the RPD developer. Also I always say in the beginning of this presentation (except in the 5 minute version) that the four user types almost always overlap - rarely is someone purely one user type without having a connection to other types. And in addition to that, the new interface makes the operational user more managerial and analytic naturally.
What is this presentation about?
The presentation is centered around showing the feature/functionality of OBIEE in the context of these four user types. OBIEE needs to be understood as a system that will be simultaneously in production for all of them. How can massive operational reporting be achieved so that when business opens up early in the morning, all the operational reports are available and quickly retrievable so that operations clerks can do their jobs? For management reporting, can we see a true menu of aggregative analytic reports that perform data mining and actionable navigation that managers can use as a jumping off point for doing all the analytical work they need to do their jobs? What about development of reports, how is it done? In the large multi-unit organization, how can shared folders be organized so that power developers can share their work with the teams, organizations? Finally the systems administrator, who is probably used to managing a large system of reports bursting and distribution, what will his/her life be like after OBIEE is implemented? OBIEE's robust features for role-based and data-based security at the object level, scheduling, output of large result sets to cache that persist throughout the day, publishing to roles, notifications, and FTP and e-mail notifications are all covered.
How can attending benefit me and my organization?
One of my favorite descriptions of user conferences is that it is a place where we who have implemented these systems come together and share expensive mistakes and how to fix them.
The failure to communicate important context about how the OBIEE implementation scales as it is developed for a large organization is a huge problem. Maybe it's the biggest problem with large scale adoption of OBIEE reporting; I can't think of anything bigger. We found that once we presented OBIEE in the context of these four user types, it made it possible to discuss and understand the points of this presentation in an hour, in a few minutes, or over the course of weeks depending on who was listening and what we were trying to do.
Please join me at this presentation at Moscone 2000 on Tuesday, October 4th, 2011 at 10:15 a.m.
Sunday, August 14, 2011
Wednesday, January 5, 2011
...playing and experimenting with analytics (reports) in On Demand... I helped develop training for it. I designed a hands-on workshop on reports development. I have built hundreds of reports for my customers. I found out that I am one of only a hand-full who really seem to know the tool well. I also found that there is not a lot of good documentation on some of the more detailed and intricate elements of report development in On Demand. That's when the idea struck me. I started by documenting things for myself, and decided that there is enough here for an actual book.
Wednesday, May 20, 2009
We had a nice turnout at the Visual Dashboard Design presentation at Collaborate! It was nice to see that Mark and Borkur came from Rittman Mead and Mark Thompson from Vlamis Software.
It would be a great idea for the Delphi Methodology (currently being defined by Jennifer Albu of Rittman-Mead to incorporate Visual into some of the methodology steps. In my presentation I list some Critical Success Factors that might play nicely into an outline for a methodology.
Thursday, August 7, 2008
Wednesday, July 16, 2008
The presentation begins with a discussion about how cognition works... I will also point out different uses of variables in dashboards... in this case the slightly larger font is directing more attention to the most important information.
The next visual here is a little video with some discussion about what a dashboard is...
The discussion continues:
Here we are discussing some great painters.... these painters used a "spotlight" effect, using light and dark. Rembrandt did not listen to the contemporary school of thought on art and did what he thought was right... as a result, he is considered the greatest painter in an era of greatness. So too, it takes courage to do what's right visually in dashboard design, even though it is not necessarily what the "glitzy" dashboards look like.
Avinash Kaushik is a brilliant author and blogger, and he discusses his issues with dashboard design... we try to solve these issues here
Here is a framework for a solution.
This slide will be discussed at some length, but it has the elements of a value driven strategic thrust in a business area, in this case Enterprise Risk Management, that we want to improve with the help of Business Intelligence. We come to understand what visual displays will be effective in the analysis.
Then we will show how Crystal Ball, Data Mining, and Microsoft Visual Analyzer, et al., in the OBIEE suite, present the visualizations we need. These can be stored in OBIEE for Information Lifecycle Management as well.
We then put the dashboards up, with good visual design. We will discuss how this supports a tremendously valuable area where organizations miss value, which is in the area of aligning tolerance down into the operations and workers.
Oracle Data Mining results are useful to the line managers trying to improve operations.
Using data mining, rules analysis will help get to the root of the problem at a fine grain.
Scatter charts in data mining help find outliers quickly.