Business Intelligence Requirements Gathering – Start with a discussion of “capabilities” | Quisitive
Power BI Quick Start Header
Business Intelligence Requirements Gathering – Start with a discussion of “capabilities”
January 10, 2013
Here is a deep dive into business intelligence capabilities of Power BI.

When I begin discussing capabilities of a business intelligence platform with my clients, there are two words I hear from those clients more than any others.  The two words I hear most often are “reports” and “data”.  It occurred to me sometime back, that if you work for an organization, in a specific industry, and you have a specific job role – then while it sounds overly simplified, business intelligence (or BI as we like to call it) is really about the data (the data you care about) presented in the way you want to consume it (a report).

Anyone who has worked on a business intelligence solution, either from a business analyst perspective or a technology implementation perspective would be quick to tell you BI is much more complex than that.  Why?  Well it starts with understanding the main capabilities of a Business Intelligence Platform (regardless of vendor or technology), and then drilling into “sub capabilities”.  I start at the top with three overarching capabilities of a BI Platform.  They are:

Top Level Business Intelligence Platform Capabilities

1)  Data Management – All processes and technologies related to gaining insight from existing organizational data.

2)  Reporting and Analytics – All formats (or BI Assets) for users to view and/or interact with data to gain business insight.

3)  Performance Management – All formats that allow users to monitor key metrics (or goals) that drive business success.

Sounds pretty simple – only three categories.  But let’s start with Data Management and the key “sub-capabilities”:

Business Intelligence “Sub Capabilities”

1) Data Management

a)  Data Integration – All processes and technologies related to integrating data from disparate systems in the enterprise into a “single version of the truth” (See Data Warehousing).

b)  Data Warehousing/Data Marts – Central data repository (or repositories) that provides consistent and accurate data across several business domains (e.g. Accounting, Finance, Operations, Human Resources, Sales, Marketing etc.)

c)  Data Quality – All processes and technologies related to the quality control of data coming in from source systems into Data Warehouse/Data Marts.  May include exception reporting, fixing data quality issues, auditing, and logging.

d)  Master Data Management – All processes, governance, and standards required to ensure consistent Master Data.  Master data can be defined as an authoritative source for the products, accounts, and parties for which business transactions are completed.

e)  Big Data – This is a new one that has moved onto the scene.  This may not be a capability per se, but the logical place to discuss the storage and management of “Big Data” is under the BI Capability of Data Management.

“Everything is a report”.  This is often how I begin requirements discussions with a client or project team.  But there are subtle nuances that draw a line between Reports, Analytic Views, and Dashboards (as well as help you pick the right tool for the right sub-capability of the BI Platform).  Reporting and Analytics (and associated sub-capabilities) are described side by side below.

2)  Reporting and Analytics

a)  Operational Reporting – An operational report presumes to answer a question that is already known by “the business”.  A classic example is a Profit and Loss statement.  It answers the question, “What is my revenue, my expense, and my bottom line profit or loss?”.  Operational reports tend to be relatively static in nature – meaning a user may select a couple  parameters (A department and a date range for example), but the resulting format is generally consistent, and only the underlying data is changing.

b) Self-Service Reporting – Self-Service Reporting describes the empowerment of a user to build their own report, in the format they need and like.  Often self-service reporting is described as ad-hoc access to pre-defined data sources (or data models), the ability for a user to “drag and drop” data elements onto a design surface, and format the report to their specific purpose (presentation, data exploration, reference etc.).

2)  Reporting and Analytics

a)  Analytics – Different from Operational Reporting, analytics describes an activity where the question isn’t fully identified.  Often I describe this as “research on why something is happening”, drilling in and out of data to further clarify the question that the user should be asking (e.g. Profits are down –>  Why?  –> When I drill into the data it looks like one product category isn’t performing –>  Why is that?  –>  When I drill further, it appears a specific product sub-category is under performing in a specific geographical region  –> POTENTIAL RESOLUTION:  Let’s create an operational report that provides sales detail by region, by product category, and by product sub-category so we can further research on an on-going basis)

b)  Self-Service Analytics – Similar to Self Service Reporting, Self Service Analytics empowers the user to build their own analysis.  Self Service Analytics goes a step further by empowering users to not only create their own analytical view, but also to model the data to answer specific business questions.

c)  Data Exploration –  This sub-capability is relatively new to my list, but has been added as a distinction to draw based on some of the new products out there that allow rapid visualization of large data sets (Microsoft SQL Server Reporting Services PowerView is a great example of this).  In the activity of data exploration, the question may not even be conceived.  The user is simply exploring known data for any key outliers or notable trends.  Let’s say we were exploring historical data on patient outcomes at a fertility clinic over a period of 3 years.  Think of an animated bubble chart mapping favorable fertility outcomes at this fertility clinic.  If there was a period of time where a specific doctor was having great success with a specific in vitro fertilization technique, data exploration would be able to spot it, and allow the clinic to inquire about the best practice that doctor was employing during that time period.  Also see the screenshot of data exploration on pollution below.

d)  Predictive Analytics – I like to introduce the concept of predictive analytics to clients and project teams by asking the question, “If you had a crystal ball, and could ask any question about the future of your business, what would you ask?”.  Predictive analytics (often termed data mining) applies statistical models to known data (historical data, populations of data etc.) to provide insight into “what may happen?”.  Determining the odds of who will win the super bowl at the beginning of a season is a form of predictive analytics.

Data Exploration:  Below is an example of visually exploring trends and outliers related to the source of air pollution in the United States over a 10 year period (source of data:

3)  Performance Management

a)  Monitoring – The activity of monitoring helps you answer the question of, “What is happening right now in my business?”.  A help desk, for example is constantly monitoring the number of outstanding tickets in the queue.  The help desk may have a “Target Time To Resolution” metric with a stated goal of resolving priority 1 tickets in under 3 hours.  If this goal supports the help desk’s mission of providing accurate response time, the team will want to monitor this metric (maybe in an aggregated scorecard with other metrics), and instantly know if the team is missing or exceeding this mark.

b)  Analytics – The same definition above applies here, but it is called out again because if you are monitoring a key metric, and it is not performing as expected, the natural analytical question is, “Why is this metric not performing as expected?”.  See Analytics described above under “Reporting and Analytics”.

c) Planning, Budgeting and Forecasting (PBF) – PBF closes the loop of performance management.  Planning is the activity of communicating what you would like to happen in the future (e.g. you set a personal financial plan (or goal) for how much money you may need for your kids’ college tuition).  Budgeting is the activity of agreeing to (or interlocking) on what needs to happen to support your overall plan.  (e.g. so much money needs to be put in a bank account each month to ensure you meet the savings goals of your financial plan for your kids’ college tuition).  Forecasting is an ongoing activity of “course correcting” based on actual outcomes that are happening (e.g. you get a bonus in a given month that you add to your bank account that changes your forecast on how much money you need to set aside each month (budget for) to achieve your plan for college tuition).  Financial PBF is an easy one to get your head around, but PBF can apply to anything – inventory, scrap, employee or customer churn etc.

Once you gain a collective understanding among the BI project team on the capabilities of an overall BI Platform, you’ll be honed in on which requirements can be supported by which platform capability.  Once the requirement and capability is matched, this makes the tool selection and technical aspects of the solution to be delivered more clear.  It is also important to note that some vendor tools focus squarely on one single “sub capability”.  Caution should be taken to ensure the vendor technologies and overall solution can support existing requirements as well as future requirements that will bleed into other “sub capabilities”.

BI Platform Capabilities – Cheat Sheet

CapabilitySub Capability
Data ManagementData Integration
Data Warehousing/Data Marts
Data Quality
Master Data Management
Big Data
Reporting and AnalyticsOperational Reporting
Self Service Reporting
Self Service Analytics
Data Exploration
Predictive Analytics
Performance ManagementMonitoring
Planning, Budgeting and Forecasting (PBF)

Technorati Tags: Business Intelligence,Requirements,Data Management,Performance Management,Reporting,Analytics,Dashboards,Planning,Budgeting,Forecasting,Monitoring,Data Warehouse,Data Mart,Data Integration