OBI SE One Installation Issues

It turns out that OBI SE One is a great low-cost solution for an initial BI project in an organization…..but it has a few “unique” installation issues.  We were having a difficult time at a client getting the software to install correctly on a Windows machine for about a week.  Dealing with the folks at Metalink3 was not the greatest as there seems to be very limited experience in installing the OBI SE One software.  The install kept failing out when attempting to install and configure the Oracle Personal Edition database that is included in the package.  We figured out a work-around and came up with a hypothesis regarding the error.

We created a new virtual machine to be our OBI SE One server and then the install went super-smooth.  That further led credence to our hypothesis that because the installation is very “Wizard” driven and not very configurable, the install didn’t know what to do when it found an existing Oracle database and listener on the same box.  It just did a “I give up” and failed out…….  OBIEE should not have the same problems as it does not require the Personal Edition DB to be installed.

Also, in the OBI SE One documentation there is a very small section that basically says that OBI SE One should be installed on a machine that is not running any other Oracle software.  Seems to be that little bit of info should be given much more importance in the documentation and displayed very prominently during the install…but what can you do.

Focus on Solution, not report specs

Business Intelligence tools these days allow for full-featured analysis to be accomplished by all levels of users with minimal (if any) training. Having this foundation in place, projects should focus on delivering the components a user needs to explore and analyze data; not twenty slightly different reports that all display the same results.

The great majority of BI projects I have been involved with had some component of “legacy work” where we had to re-create a certain number of reports that were produced in a legacy system or tool. This usually comes from a push within the client to keep some users or groups of users looking at the same content in the same way…..the “don’t rock the boat” mentality.

When implementing a new tool, I completely understand the need to deliver similar formatted content to key users in order to allow a smooth transition and no interruption of decisions, but I always look to do more. I ask probing questions about not only the technical construct of a legacy report but how that report is used. What are the key numbers a user is looking at? What decisions are made based on those numbers? What other data is required based on report results? Who do they call if a number changes drastically? I look to deliver reports, charts, graphs, alerts, dashboards or any other format of data to answer all of these questions.

Think about it…..there must be some reason that a new tool is being bought and implemented…and reasons to absorb the cost of those activities. Use this opportunity to exploit the functionality of the tool. Don’t just deliver the same reports in a new font or interface.

Use Text on Dashboards

Today’s BI Dashboard tools, including OBIEE, allow for text or other descriptive data to be added a dashboard page to help all users understand the meaning and purpose of a particular page or frame. The chart and visualization options that OBIEE offers allow for easy to understand charts and graphs to be displayed to the users, but a number of times there may be a particular point of emphasis or background information that a subject area expert or administrator can only express via a sentence or two of text. Dashboards don’t always need to be speedometers and traffic lights only.

Design for Users not Admins

OBIEE Answers allows users to choose data elements and measures that are grouped and named in ways that make sense to them. Don’t lose sight of that. Too often I see that the names of objects and folders that are displayed to the user in OBIEE Answers is just the same name as the underlying data column(s). If the column on the database is “FACT.Units” why not use the functionality of OBIEE to display to the business users something like “Daily Units” or “Units net Returns” depending on what the FACT table is actual storing? These minor tweaks to names reduce any possible confusion by the users and increase the confidence in the information they create.

Just as with the names for the data elements, I see the grouping in OBIEE Answers is most often one folder per database table. Admins create one folder for every table in the database and all columns that appear in that table appear in said folder. This is a good first step when developing the repository file, but it should be looked at as just that – a first step. Users do not always mentally group objects in the ways that the database tables do…and that is where the beauty of OBIEE allows for those groupings make business sense. If users think of a certain 5 or 10 objects as being something they call “Demographic Info”, even if it comes from 3 or 4 different database tables, why not group them in a Demographic Info folder?

“KISS” Ad Hoc

Most everyone is familiar with the “KISS” acronym used to be a guide when building or developing virtually anything. “KISS” stands for Keep It Simple Stupid and that makes perfect sense to me. KISS makes its way onto a BI-focused blog by being the best way I can communicate how the Column Selector view in OBIEE Answers allows virtually any user to perform simple ad-hoc analysis of data.

What the Column Selector does is allow an Admin or Power-user to set up a set of different data values that can be placed in the same physical spot of a data table. Think of data in a spreadsheet, 3 columns of data that represent City, Units and Dollars in columns A through C. The column selector allows the user to take an intuitive drop down menu and replace City with Product and have the data in the table change accordingly.

To most users this will easily suffice for the “ad-hoc” or “slice and dice” ability that business users undoubtedly demand from their BI tools. The Column Selector allows a layer of simplicity and ease of use to be applied to OBIEE Answers. Granted, it is not the solution for power users who crave true ad-hoc analysis, but for the majority of users out there, this is a great option.