Introduction

This article describes the high-level steps and use case for building a "dashboard-like" WebI document and bursting it to users for offline consumption. We will first describe the requirements collected from our users and explain why Xcelsius 2008 cannot easily fulfill them, then explain how the Web Intelligence Rich Client and Publication capabilities of BOE can help.
Please note: this article is not about replacing Xcelsius by WebIntelligence for all dashboarding projects. It aims at documenting one specific scenario that is in the grey zone between reporting and dashboarding. Xcelsius remains the tool of choice for dashboarding within the SAP BusinessObjects Enterprise suite, especially in a connected (online) scenario. "Offline consumption" is the main constraint in this scenario; the same scenario (data volume, access restrictions, KPIs, etc) with online consumption can be implemented with Xcelsius.

Use Case Description

Here are the requirements for our dashboarding project :

• the sales organization need access to a summary view of their KPIs (revenue and margin) across multiple dimensions (geographical and product)
• users need the ability to access the dashboard offline (part of the sales force is mobile and work disconnected from the corporate network)
• the data comes from a relational data source that can be accessed via a Universe
• each user has access to a different set of data based on his responsibilities and geographical location according this org chart

What kind of dashboard ?

The sales dashboard should focus on 2 measures (Net Sales and Margin) and 9 dimensions grouped in 3 hierarchies :

• geography : Region, Country
• time : Year, Quarter, Month
• product : Line, Category, Product

Based on the org chart above, users will have access to data for their region only (dimension "Region") and the business units (dimension "Line") they cover. The dashboard will present 3 drop down lists on the top dimensions "Region", "Year" and "Line" for the user to filter values. It will also offer drill down and interactivity between the different charts. Here is what an Xcelsius mockup would look like:

Some users will only see 1 region and line, whereas others will see all lines in one region, etc. The higher in the organization chart, the more data (and filter values combination) the user has access to.

Offline consumption

In our scenario, the end users need an offline (i.e. disconnected from the corporate network) access to the dashboard from their laptop computer. We plan on sending an instance of the dashboard with personalized data by email on a regular basis. Because the dashboard instance will be disconnected from the data source, it will need to be self contained with all the data required to support all the filter values combination possible for that user. As we highlighted earlier, users at the top of the organization chart will have access to much data than other users. Let's quantify that and estimate how many cells / data points each user would require in a dashboard instance.

Our data set counts:

• 5 regions
• 16 countries
• 5 years of data (4 quarters, 12 months per year)
• 3 lines
• 13 categories
• 54 products

A regional business unit manager who oversees one product line for one region will retrieve 1,000 to 4,000 cells (depending on the region and the product line) to cover every possible combination of filter values, a regional manager (all product lines for one region) 4,000 to 10,000, a worldwide manager (all product lines for all regions) about 25,000.

Why WebI and not Xcelsius ?

Although Xcelsius is the tool of choice for dashboarding within the SAP BusinessObjects Enterprise suite, Xcelsius 2008 and BOE XI 3.1 cannot easily fulfill these requirements because:

• it is not possible to burst instances of Xcelsius dashboards with the built-in capabilities of BOE
• the number of cells required for an offline consumption are higher than what the Flash player can handle

On the other hand, WebI documents can be bursted to users with BOE Publications, and also offers a desktop client for offline consumption scenario. The WebI Rich Client provides the level of interactivity that users would expect for such "dashboard".

Implementation Steps

Here are the high level steps to implement such a solution. This series of articles provides the detailed configuration steps, each article focusing on a specific topic (security, viewer customization, bursting). Note that each step may have several variations, which we're giving examples of here.

High Level Step	Screenshot	Possible Variations	Link to article
1- Map the organization chart in BOE by creating the same hierarchy user and group accounts, and set up the security model within BOE		We used and documented Enterprise authentication, user groups could be mapped to a corporate LDAP or AD directory.
2- Configure the data access restrictions based on users responsibilities and org chart position		We used and documented Universe Overloads (Access Restrictions), Row level security may be done within the data source.
3- Build the Web Intelligence document.		We used Input Controls as filter value selectors, we could have used local filters.
4- Create and schedule a publication with email destination.		Because the WebI document relies on Universe Overloads for data security, we did not use Publication Profiles for data personalization. Global or Local profiles would be necessary if Universe Overloads were not used.
5- Customize the WebI Rich Client interface to turn it into a viewer only.		Further customization can be performed, examples are provided in the article.

References and links

This article is part of a series