Application intelligence.

The promise of 'a single version of the truth' has been one of the driving forces behind the growth of business intelligence (BI) since its inception. But increasingly, the competitive landscape of the BI market is forcing users to consider who should provide the source of that truth.

The debate has been sparked by the arrival of the enterprise resource planning (ERP) software vendors on the BI scene. And while traditional BI software vendors stress that only in rare circumstances would an organisation be able to satisfy its analytical requirements by drawing on data from a single ERP system, the ERP vendors insist that the data generated by their systems provides the best basis for getting at a true picture of what is going on in the business.

Fierce competition among the 'pure-play' BI vendors is nothing new; the likes of Cognos, Business Objects, Hyperion and MicroStrategy have battled hard for a decade to establish market leadership. But they now all face a fight on a completely separate front, this time against adversaries such as SAP, Oracle and Microsoft with vast resources.

The relationship between the vendors of BI software and ERP applications, at one time symbiotic, is now increasingly uneasy. BI systems have always drawn on the data created by ERP production systems as a basis for analysis; in doing so BI has added considerably to the value that users have derived from their large investments in ERP.

But with most large organisations having now completed their ERP implementations, the vendors of such systems are looking at other opportunities within their customer bases - and that means expanding into adjacent territory, with BI viewed as a particularly lucrative area.

The ERP heavyweights certainly have some advantages. Above all, is their ability to exploit a captive customer base. Companies from SAP and Oracle to Coda and Sage are targeting existing customers with BI products, often encouraging take-up by bundling the analytical capabilities with core ERP components.

Second, the ERP vendors argue that the tight coupling of BI tools to their applications means that analytical capabilities can be seamlessly integrated into the business processes that are driven by ERP.

The upshot is that BI vendors have more convincing to do, says Jacqueline Coolidge, an analyst with IT advisor AMR Research. "Smaller [BI] vendors will need to work much harder to distinguish a unique value proposition to survive," she says.

They also face cost arguments. Users are having to decide whether they need separate BI capabilities or can make do with those delivered by their ERP provider either for free or at a hefty discount.

That is an easy decision, says Bernard Liautaud, CEO of BI tools vendor Business Objects. While different ERP vendors may have optimised their applications to share data with their own BI products, whenever users need to deal with data from outside that ERP suite they are going to encounter performance and integration issues. The idea of one ERP vendor's BI tools working with rival ERP vendor's data sources is a stretch, says Liautaud. "I just don't think you're ever going to see anyone using a Siebel BI front-end to access SAP or an Oracle BI front-end to access DB2 [IBM's database]," he says.

He is joined by others from the BI market who highlight the shortcomings of the ERP approach. "They have not delivered a lot [of functionality] yet; they've promised it, but we haven't seen it in the market," says Nigel Youell, UK marketing manager at Hyperion.

The reality is that ERP vendors still trail their BI rivals in terms of the sophistication and breadth of their products - with one or two exceptions such as Microsoft's online analytical processing (OLAP) product. But they believe they can gain ground fast.

That shows in their efforts to offer access to multiple data sources. Using data from non-SAP systems is already no longer an issue, claims Roman Bukary, marketing vice president for analytic applications at SAP. "We now have the integration layer, and we can provide the context. The idea that we can't do this isn't an honest argument."

There is also a question of data quality. Says Pete Snelling, product manager at BI vendor SAS UK: "Operational systems, by their nature, suffer from data quality issues. These often don't affect the operational business dramatically but their impact is magnified once the data is used in analytics."

ERP vendors say they are addressing that. And although still very much in its infancy, their focus on service-oriented architecture (SOA) may help iron out the issue of developing a semantically consistent data model across multiple platforms. SAP, for example, has built its data integrity engine, Master Data Management (MDM), into the core of its SOA platform. MDM cleans data extracted from multiple sources.

The shift to a componentised, service-oriented architecture, where applications are built from a pool of functionality, will have a profound impact on analysis and reporting. Potentially data could be coming from any number of sources, making context vital.

Not to be outdone, the BI vendors are also embracing the opportunity that SOA can bring to their tools. In January 2005, Business Objects released its latest BI suite, BusinessObjects XI, built using web services. Similarly, Information Builders' iWay division has bolstered its integration capabilities with its Adaptive SOA Framework.

And the BI camp believes the arrival of a componentised architecture will make it easier for users to embed analytic functionality directly into business processes. That, they say, will allow them to challenge the perception that ERP is the only 'home' of BI business processes.

That convergence is likely to lead to ERP vendors making acquisitions in the BI arena, raising the stakes further for users as they pursue a single version of the truth. n

Consider ERP when:

The business needs a single-vendor platform with out-of-the-box, pre-packaged, integrated analytics;

BI applications are centred around ERP systems.

Consider BI vendors when:

The business needs an enterprise-wide BI tool to support the managed generation and distribution of highly formatted reports, and ad hoc query and online analytical processing capabilities for a broad range of users;

The business needs heavily customised analytic applications and data models.

Source: Gartner