Self Service Business Intelligence (SSBI) refer to a new set of business intelligence applications and tools that empower decision-makers to satisfy their information needs with no technical skills or intervention of IT professionals.

When compared to traditional business intelligence, SSBI is expected to provide faster results, improved decision-maker agility and easy navigation inside strategic information with a reduced investment.

SSBI make available  information, not reports and can be constructed incrementally mapping available data to business information.

SSBI strategy is to let available information “elements” from which they assemble reports and build charts the way they need, as shown by this PISA screncast (click on “Next” button, when visible, to advance).

Many SSBI tools are based on spreadsheets that are familiar even to non-power users. On the other side spreadsheet formulas can quickly become impracticable when data analysis computations become too complex. In such cases an hybrid approach is probably best alternative.

Future SSBI software will possibly provide “smart” capabilities, such as decide best chart type to plot data, automatically identify situations that need attention or select applicable forecasting algorithm.

SSBI success heavily depends on a flexible and documented data model that is intuitive and shared with no ambiguity by decision makers.

But software is also a fundamental factor in SSBI success. You can't expect decision makers to write complex expressions of nicely format reports. It is not a question of being or not a power user: it is just a question of time. Decision makers will not spend their precious time in such clerical tasks.

Frankly speaking most report cosmetics is dispensable, at least from a decision-maker point of view. What actually count are quantitative values and exception highlighting. By the way many “bells and whistles” (like gauges) widely used in dashboard are considered poor communication examples by visualization experts.

On the other side complex calculations are unavoidable. Take a look at MDX generated by PISA for the above example (and this is an high level language, still to be mapped by OLAP engine into SQL queries):

WITH SET [Top 5 members of Store by Amount ] AS 'TopCount([Store.By region].[Store].Members, 5.0, [Measures].[Amount])'

MEMBER [Measures].[Improvement] AS '([Measures].[Profitability Oct/09] / [Measures].[Profitability Oct/08] - 1)'

MEMBER [Measures].[Profitability Oct/08] AS '([Measures].[Profitability], [Time.By month].[Any month].[2008].[4th qt. 2008].[Oct/08])'

MEMBER [Measures].[Profitability Oct/09] AS '([Measures].[Profitability], [Time.By month].[Any month].[2009].[4th qt. 2009].[Oct/09])'

MEMBER [Measures].[Amount Oct/08] AS '([Measures].[Amount], [Time.By month].[Any month].[2008].[4th qt. 2008].[Oct/08])'

SELECT

NON EMPTY {[Measures].[Amount Oct/08], [Measures].[Profitability Oct/08], [Measures].[Profitability Oct/09], [Measures].[Improvement]} ON COLUMNS,

NON EMPTY [Top 5 members of Store by Amount ] ON ROWS

FROM [Sales]

Can you expect a decision-maker to code such functions (even in simplified format) inside an Excel spreadsheet?