Xcelsius is one of the most innovative business products of the last decade and in combination with the BusinessObjects platform (both XI 3 and BI 4), it provides a world-class, market-leading, enterprise dashboarding solution which is delivering enormous value to thousands of organizations around the world. Given this, it is not surprising that on Twitter and in the blogs, there is a excitement building in Xcelsius circles about SAP BusinessObjects BI 4.0 and Xcelsius 2011.
The focus is mainly on three new capabilities:
- The new embedded query creation capability (replacing QaaWS)
- The ability to bind these queries directly to components (rather than going through the spreadsheet)
- Built in “Query prompt Selector” controls.
As stated in the “What’s New” documentation for BI 4.0 these new features are “focused on improving productivity and direct enterprise data connectivity”. The $64,000 question is “are the productivity improvements they bring marginal or breakthrough?”
As anyone who has been involved in developing Xcelsius dashboards connected to server side data will know, the two key things which hamper productivity are:
- The need to manage multiple connections to the server
- The need to use many formulas to manage the data from these queries into the components based on end-user selection
These are also key drivers of total cost of ownership because the complexities they introduce to the initial development also, even more dramatically, affect the cost of ongoing support and maintenance.
Multiple Data Connections
The main reason for having multiple data connections is to get data from different hierarchical levels with each level (indeed each combination of levels from different dimensions) requiring a different query. So a dataset with just two hierarchal dimensions, each with three levels of hierarchy, would require 9 (3×3) different connections to get all possible levels of data. However, it goes beyond that, the need to also have “select distinct” queries to populate lists of values and a number of other auxiliary requirements, mean that most non-trivial connected dashboards typically require between 10 and 20 queries to serve the end user needs, even on a single dataset.
With multiple datasets comes the need to manage which dataset drives which component in response to the end user selection. The formula engine of the embedded spreadsheet is an incredibly powerful and flexible method for doing this, however, as has been commented on so many times in the past, spreadsheet formulas are incredibly difficult to debug, document, and maintain. This is even more true when the use of complex formulas (e.g. ones using data manipulation functions such as VLOOKUP) is required.
If we combine these two together into an index of Xcelsius complexity it would look something like:
Xcelsius model data complexity = (100 * number of queries) + (10 * number of unique “complex” formulas) + (3 * number of unique simple formulas) + number or repeated formulas
(I have just pulled the factors here out of the air to illustrate the point, but would love to engage with others in the Xcelsius community on a standardization of a measure like the one above)
So the question is: “do the new query and binding capabilities of BI 4.0 reduce this (or any similar) complexity index ?”
Unfortunately, based on what I have seen so far, the answer is “no”. Although the new query capability “replaces” QaaWS, it looks to have pretty much the same level of query sophistication, i.e. fairly static, parameterized queries (although the new capability does add the ability to add static sorting), with no dynamic, hierarchy awareness. This means that it is unlikely that Xcelsius models ported to BI 4.0 would be able to rescue the key complexity drivers of “number of connections” and “number of data manipulation formulas”.
This in turn minimizes the value of the other new features, because directly binding data from a query to a component and managing that query directly with a prompt selector is only really useful if the query returns all the data you need. Otherwise, you are still left with needing to manage the data through the spreadsheet which disqualifies the use of these new features.
That is not to say that there are no benefits to the new capabilities. They do make the whole process of initial query creation simpler, because you can do it all from within the Xcelsius environment and for simpler dashboards that have static (albeit parameterized) data requirements the new direct binding and prompt selectors will prove to be a productivity boost. Additionally, they also make demos of data connectivity in Xcelsius much simpler!
If anyone feels differently, or even better, has an example of a real-world Xcelsius dashboard which is much simpler using the new features, I would love to hear about them. I would also love to hear any thoughts people have on the creation of a complexity measurement for Xcelsius dashboards.
Donald MacCormick is a board member and strategic advisor to Antivia, an SAP® software solution partner, and creator of XWIS the Xcelsius-to-SAP BusinessObjects connectivity solution. In past years, he was a long term member of BusinessObjects and Crystal, and part of the team that brought Xcelsius into the BusinessObjects portfolio.
Current Highest Rated Post
- Video: Show Chart Values without Hovering (5.00 out of 5)
- Pictures from the Xcelsius Community Event at GBN 2009 (5.00 out of 5)