Organizations continue to invest in data warehousing and business intelligence tools despite the wobbly economy...
-- or perhaps because of it -- as companies try to remain competitive.
Yet for companies developing an SAP data warehousing strategy, there are several areas to take into consideration. First and foremost, they need to look at where their data lives now. Whether to turn to SAP for a data warehouse or to a data warehousing specialist depends largely on the company's existing structure.
"It largely depends on your data sources. If you're a wall-to-wall SAP shop, it's a no-brainer; but of course, that's not always the case," said Madan Sheina, principal analyst at Ovum Research.
As more companies use information like a strategic asset, effective access to that information is key. So, if roughly 80% of the data is coming from an SAP source, there's little reason to look elsewhere.
"But if it's less than 60%, then there's little reason to go with an SAP data warehouse," said Boris Evelson, principal analyst at Forrester Research. "You'd be trying to put square pegs into round holes."
SAP's data warehouse -- called BW -- is a big database that stores transactional historical information to help business users make decisions.
"The SAP data warehouse is not typically architected and wasn't invented to hold non-SAP data," Evelson said.
Bringing non-SAP data into BW introduces complexity.
"When you have to build new structures, non-SAP structures, it can be tricky and problematic because you have to build the routines from scratch and then integrate these different formats within the SAP format," Sheina said. "It's all about making BW fly in a heterogeneous environment."
The top two reasons customers cited for staying with BW (according to a May 2009 report by Forrester, "To BW Or Not to BW") are the large proportion of data sourced from SAP applications, including R2, R3, MySAP and xApps; and tight integration with SAP applications and SAP BI products. Out of the 40 enterprises surveyed for the report, 47.5% stated that more than 75% of the data that feeds BW comes from SAP applications.
SAP's acquisition of Business Objects may help companies turn to BW because Business Objects has its own ETL tool called Data Integrator XI -- which is part of a bigger data management suite within SAP Business Objects.
"This strengthens [SAP's] BW position as a bona fide data warehousing platform," Sheina said. "It could also signal a change in emphasis, as it will be much easier to consolidate data from non-SAP sources into BW, which has always been a source of contention for any SAP system."
SAP lagging behind other vendors in scalability
Although SAP is one of the largest data warehousing vendors, some users have been dissatisfied that the company has fallen a little behind in terms of scalability.
"Competitors such as Teradata, Oracle and IBM can now scale into the petabytes (1,000 terabytes)," James Kobielus, senior analyst at Forrester, said. "SAP is making some strides to evolve with BW, by allowing you to accelerate queries using a front-end platform called BW accelerator." He also points out that SAP is going to continue to evolve BW to handle hundreds of terabytes. "So they're making great strides. However, their users are concerned that they haven't been innovating as fast as the competition, and that's a justified complaint."
In addition, SAP's competitors are expanding into more complex data mining functions like scoring and database analytics. SAP doesn't have a strong focus on data mining and advanced analytics efforts, according to Kobielus.
"BW customers are frustrated by the fact that it doesn't scale, and it's more complex to administer and design integration code," he said. "SAP BW is a good data warehouse, but it's not intuitive enough. But they know that, they're evolving that. SAP realizes they need to continue to beef up that product, because customers are looking for ways to migrate away from BW if they can."
In-house BI specialist key to success with SAP BW
The next consideration when it comes to strategy concerns systems integrators. "BI and data warehousing is not plug and play," Evelson said. "It requires lots of customization, it's all about best practices, it's not like you can buy a textbook and do it yourself."
More than just implementing it, SAP BW requires some specific skills to keep it up and running.
"The biggest shortcoming of SAP BW is not on the implementation side -- although that can be complex -- but in keeping it running," Kobielus said. "First of all, BW shares a lot of code with its parent R/3 product; it uses the same kernel and ABAP languages. It isn't necessarily the functionality of the product but the cost involved in implementing and keeping it running."
As a result, companies have to make sure that they have a BI specialist in house. "Finding those resources can be difficult and costly," he added. "They also have to make sure that they understand some of the processes in R/3. They have to understand how the data is structured and the process models are upgraded in BW to analyze it."
According to Sheina, Business Objects helps by broadening BW and extending it on the front end. "Before, the front of BW was Business Explorer, which was a complicated tool. Now, with Business Objects, SAP extends its breadth and its value at the back end and the front end as well," he said. "The complexity is still there, but it does give you more leeway to source a complete end-to-end BI data warehousing solution from a single vendor."
"The company needs to continue to understand that the world is not SAP-centric," Evelson said. "They need to make tools more open. But, on the other hand, by making them more open, you lose some precision, so it's a fine line: better integration vs. being more open."
There will be an ongoing role for BW as a big database to do predictive models, according to Kobielus. "It's a valuable investment that customers don't want to squander, but they need to be able to leverage it for its full potential."