Performance Point – the Good and the Bad – Part 1
This posting presumes the reader has at least a basic understanding of what Microsoft’s Performance Point (PP) offering is, at least at the level of wanting to assess whether it is appropriate for your requirements. Basic familiarity with Microsoft’s Analysis Services is also required to get the most out of the observations I provide. If you have that background, this two-part posting should prove helpful in making an assessment of whether PP might suit your requirements, in whole or in part.
PP is Microsoft’s principal offering for building web-based dashboards and scorecards. I say “principal” because it is not the only way to achieve these using Microsoft technology – but it is the one most directly targeted to that need and has the most out-of-the-box (OOTB) functionality for that purpose. It is a server-based offering which is fully integrated with and requires Sharepoint, as of the 2010 version (the earlier 2007 version used an independent server). You must have the Enterprise edition of Sharepoint.
PP provides a lot of functionality OOTB, but it is also disappointing, often maddeningly so, in what it doesn’t offer OOTB and, even when workarounds are possible, they typically have significant caveats – even in the 2010 release. There is a “Zen” to PP. Understand it and work within its “sweet spot”, and it can be a beautiful thing. Try to fight it, and you will lose. While this characteristic does not relate to its integration with Sharepoint, Sharepoint itself is like that too.
While PP can work with non-cube data sources, its sweet spot is definitely with using Analysis Services (AS) cube data sources. Use of PP with AS cubes will be smoothest if one also has free reign with the related cube and the skills to leverage that, as some workarounds are only possible via cube additions or changes.
This will be a two-part blog, starting, in proper performance appraisal style, with Powerpoint’s OOTB good points. Part 2 (http://blog.tallan.com/2013/03/17/performance-point-the-good-and-the-bad-part-2/) will deliver the bad news and some tips.
Performance Point OOTB Good Features:
1. With Analysis Services sources to analytic grids, cube cell actions are supported. This is in contrast to Excel Services (a Sharepoint feature), where cube actions are not supported, in my view one of Excel Services’ gravest shortcomings. Note, however, that only cell actions are supported – no other type.
2. Analytic Charts and Grids created using Dashboard Designer (DD) and non-customized query MDX provide built-in, highly functional drilldown and drill-across functionality via right-click.
3. You can optionally filter out empty rows and/or columns.
4. You can place an entire hierarchy on rows or columns, or you can drill a hierarchy to a particular member and put either it, or its children, on rows or columns. In the former case you can drilldown to children (as long as not disabled as discussed in Part 2).
5. PP remembers filter choices from session to session on a per-user basis, for a configurable length of time (server-wide), even if the data source connection uses a shared (i.e. service) account. This can be good or bad, depending on needs.
6. PP provides built-in Time Intelligence, which allows you to write expressions (in a specialized language, not MDX) to express date arithmetic based on a selected “current” date. Some fairly sophisticated dynamic date range functionality can be achieved. However, this functionality is very poorly documented and can be therefore difficult to use. Furthermore, if you have multiple, role-playing date dimensions, it loses its appeal vs. the technique of having a time-calculation dimension in the cube. However, if you lack this and cannot add it (you cannot touch the cube), then Time Intelligence may come in very handy.
7. PP KPIs can source data for each component from different data sources. This is a major differentiator of PP-based KPIs from AS-based KPIs. In the former, each component (actual, target, etc.) can come from a different data source, while in the latter they must all come from the cube containing the KPI. For instance, in a PP KPI, Actual could come from the cube and Target (say budget) from a SQL source, Excel or Sharepoint list.
8. In DD, when creating a dashboard, the pre-defined page layout templates that are presented give the impression that layout options are extremely limited. However, while not intuitive, it *is* much more flexible. You simply choose one of the pre-defined templates, and then by right-clicking on any zone you can delete it, or add a new zone on any side. Quite elaborate web part zone layouts can thus be constructed.
9. In SP 2010 SP1, cascading filters have (finally) been added.
10. PP seems to have been designed specifically with the Balanced Scorecard performance management methodology in mind. This is reflected in its support of “KPI rollup” as described in that methodology. Under the methodology, KPIs are expected to be organized in hierarchies. For instance, each department might have its own KPIs, tailored to that department’s function. KPI rollup provides a mechanism to construct KPIs at a higher level – say, for a division, or even the entire enterprise, which are derived from the KPIs at lower levels, rather than being defined specifically at that higher level. PP provides great functionality for enabling this, and in fact is the only well-known dashboarding tool that does as far as I know. Thus, PP could be extremely attractive to an enterprise using the Balanced Scorecard methodology.
11. Integration with Sharepoint. Obviously this may be good or bad, depending on whether you already deploy Sharepoint, and have the Enterprise license. Assuming you do, dashboards take the Sharepoint-familiar form of documents in document libraries, with all the built-in Sharepoint functionality that that implies for security, customization, search, and so on.
In Part 2, we will review some of the less-than-ideal OOTB limitations of Performance Point.