OBIEE Step by Step Guide

May 11, 2010

OBIEE Gathering Business Requirements

Recently, I was asked by a new OBIEE implementation project team on what are the different ways to collect business requirements for a new BI initiative. With the project objectives being identified and high level business requirements being gathered from the c-level executives, what are the next proven steps for a successful OBIEE implementation.

My experience with this would probably go back to old days where you collect the “AS IS” status of the BI reports out there. Do they feed off a OLTP system or a data warehouse? Whats the future state of this BI initiative. How do we connect the existing dots to give you a vision on whats relevant to the new vision set by the management? This certainly leads you to conduct a “gap analysis”. Gap Analysis of what? Obvious answers are

  • Existing business processes
  • Existing technical architecture
  • Existing resources

Does existing business processes support the vision set by the management? Does the existing technical architecture fit the new vision? Can the existing resources be cross-trained for this new vision.

I am sure these are dealt by the BI project managers on day to day basis.

Coming back to OBIEE, you have OBIA (Oracle Business Intelligence Analytics) that comes with pre-packaged dashboards and pre-packaged ETL mappings. Then, you have custom OBIEE development where none of the pre-packaged apps does not fit your business model. While OBIA covers 50-60% of the organizational domains, you have bolt-on apps that support your existing ERPs and other core business processes. How do I integrate these bolt-ons to the OBIA? Should I start all the way from scratch because we have invested years and years into developing a data warehouse already? Many questions linger in BI project manager’s mind before they take the next big step.

With all these premises set, lets get back to the question of how do I collect my business requirements for a new custom OBIEE initiative?
To answer this specific question, this specific list of questions have been collected and hopefully this will help you in your implementation. I am sure its not comprehensive enough, but can be a good starting point. Remember almost all these questions are relevant for any new BI reporting requirements gathering phase. Since we are talking to existing business users who are not using OBIEE yet, this is still a very good generic reporting requirements template, questionnaire.

Please add your comments to this article so others can benefit from you as you benefited from reading this article. As you know this world is all around ‘Give n Take’ mantra. When you want to take something, you have to give back to the community in other possible ways ūüôā

Questions regarding existing BI reports to Business Users

  • How do you access these existing reports?
  • Whats the existing frequency of report updates and data updates?
  • What reports go together to make them relevant?
  • What other tools do you use for analyzing the data out of these reports?
  • What are the report delivery mechanisms existing right now?
  • Are these refreshed on a batch process or based on demand from the users?
  • Any specific features in the existing report that you would like most and does not see loosing it which may cause potential business implications?
  • What are the existing bottlenecks with each of these reports?
  • Are the goals in existing reports still relevant?
  • What are the different parameters used in these reports? Whats the frequency of each parameter being used?
  • What reports do you cross check to make sure your reports are upto date and are valid?
  • What reporting metrics/KPI are being created in these reports? Are there special formulae being set in these reports as we bring data from the data warehouse?
  • How do you perform your routine analysis? How does the existing reports aid in getting your job done faster? Do you see any trends in what kinds of routine analysis are you being tasked with?
  • How complicated can these ‘one off ‘ reports being created? How long does it take to create these reports? How often do you think the data is already available in data warehouse but not in reports?
  • Who are the power users for each of these reports? Organizing these users by department, by task can give you a very good insight of what reports can be grouped together in an OBIEE dashboard.
  • Is there report usage tracking mechanism already available? Do we like to see this moving further into OBIEE?
  • Open up your c-level executive’s high level requirements and see whether they are being already covered by these existing reports.
  • Is there a backlog of users requests for enhancing the existing reports, create a new flavor of report based on the existing report?
  • Can all reports be viewed by every one? How are the users segmented so that groups can be created to access these reports?
  • Does the report format change based on the user? Is there any row level security involved? Is there any column level security involved in these reports?¬† Simply¬†ask ‘Who can see what’ for each of these reports.
  • Do you analyze information or conduct analysis quarter over quarter or year over year? How far back do you go to support your analysis?
  • Ask for access to the recent support ticket logs and see whether there is any trend of problem areas with the reports.
  • Organizing the reports by business importance i.e. mission critical every day, end of month report to be ready by 2nd fiscal week¬†of next month etc. etc..
  • What is the existing support structure for these BI reports? Ex: When a user has problem, he goes to his assigned power user, if power user cant resolve it, create a ticket for BI team.
  • Identify the existing drill down capabilities in the existing reports. Is this still valid? How often does these hierarchies change? Ex: A Product Hierarchy is created every month, we create a new IT request to include this new product hierarchy and it gets added.
  • Identify the common dimensions and hierarchies used in most of these reports. Ex: Time, Product, Customer
  • Identify the common set of filters that are being built across all these dimensions.
  • What are the known gotchas with the existing data in the data warehouse. Do we have reports where multiple versions of truth is always a problem just because the way these reports are set up?
  • Are there any decodes, if this is ‘x’ then ‘y’ kind of expressions developed as part of the reports?
  • Are we happy with the existing refresh schedules?
  • What output formats are being supported by these reports? Ex: Excel, PDF, web-based always.
  • Identify any enhancements that would really delight the customer based on an existing report.
  • Identify opportunities where previous attempts failed and the it did not make any financial sense at that time. Review these with your users and see whether that it is still the same.
  • Document on how each of these reports fits strategically to the organization’s goals and performance monitoring. What are the success factors for your organization? Are these factors quantifiable? How do you know you are on track to achieve your results? How often does the department’s/organization’s goals change?
  • Does any of the existing reports help you predict problems? Are there any alerting mechanisms built into these reports? Who gets these alerts?
  • Check the reports for the visualization techniques used in the existing reports. Does it make sense to replicate them as is or see whether you can add new web 2.0 widgets for data visualizations. Ex: There was never a goal against this KPI before. Now, we have one. We like a dial chart on the first page of our report to show how we are performing weekly, quarterly, yearly.
  • Identify users and see what level of familiarity do they have with this data. Segment these users based on criteria like influence, authority, participation levels. You always need high influencial, high authoritative and high participative users. Deal slowly and patiently with low influential and highly pessimistic users.
  • Identify the BI technology trends in their organization. Ex: We used to have these nice and fancy all excel based reports. Then the IT has changed our strategy 5 years ago and asked us to start using Micro Strategy tool. Now, we are going to go OBIEE. Check the pulse of your users to see how excited they are.¬† As the technology changed, did something worked better? Did something get lost as part of the transition. Look for those transition pain points and make sure you have/develop some strategies on how to deal with them if you encounter one.

With these said, some IT analyst will suggest creating a template for this in Excel or other tool. Well, I would like to leave it as is and ask you the readers to customize the way you want or your client likes to see.

Next time, we will discuss about the tasks of BI analyst once he collects these interviews and develop some findings.

Next week.. we will follow on the tasks to be completed by the BI analyst once he interviews the business users and collects all this useful info.

Blog at