A Case of SAP "Picking up the Scraps"?

A tremendous amount of news coverage has been given to the acquisition yesterday of Business Objects by SAP. While this was probably the Valley's biggest unkept secret this past week it came as some surprise the 21% premium that SAP was willing to pay on Business Objects shares, already trading at their highest levels in many months.

I will not go into the financial details of the transaction, as they can be found elsewhere and in greater detail than the intended scope of this blog. I will, however focus on the broader technology impact and what this may have in store for current BO customers and future SAP customers.

Although the BI market leader, BO has a fractured architecture and one where a confusing array of options is sadly indicative of the BI market (Cognos being another notable example of this lack of architectural elegance, although 8.0 has made huge strides in the right direction). BO XI - the flagship product has many admirers. It gets the job done for many analysts and data junkies, often tasked with the job of making sense of data residing in rigid schemas not particularly well designed for long-term business growth. Hence, the flexibility of XI to enable analysts to build out metrics and reports directly off OLTP or analytical tables while ensuring consistency, and no chasm or fan traps has always been a strength. As soon as you move off the thick client however, is where things start to look ugly.

I've never been a big fan of BO's web architecture and trying to support a large number of web users (particularly analysts with requirements for ad-hoc analysis) is something that BO has never done well. Web Intelligence is neither well suited to the masses (it's overly complex and difficult to use) nor embeddable (the breadth of the APIs are weak and BO's OEM sales model is not particularly customer-friendly). With SAP placing a large emphasis on Netweaver as the future technical direction of the company I'd be interested to see how SAP will pull off trying to get Netweaver customers to integrate Web Intelligence into their environment. A far better choice would have been something more seamlessly embeddable and more open to the pro-development stance of Netweaver - such as Actuate (strength in embeddability), MicroStrategy (APIs) or QlikView (data discovery for business analysts). Cognos in this area is also marginally better than BO.

However, it's pretty clear to me that SAP decided to go with BO for one key metric, and that's customers. If Henning wants to achieve his goal of 100K customers getting 40K or so of them from BO certainly helps!

The question is how many of those 40K will remain loyal to BO once the SAP reps start talking-up Netweaver and tools such as Visual Composer. It's hard to see the complimentary nature of these fundamentally different approaches. Netweaver+VC was always a pitch to developers and CIOs. That's territory that BO hates. BO reps always aimed at the CFO (with their EPM products), CMO (with XI) or Sales (with Crystal) - but essentially the business side. Netweaver is an IT sell. So who wins? Will Netweaver be pushed aside by BO reps as they continue the charge into the BI space armed with fresh SAP customers contacts and lots of new Euros to spend on marketing? Or will the Netweaver push eventually force BO to become more open, embeddable and standards-based?

If one looks at the traditional BI customer base for SAP they are pretty much all die-hard SAP BW users. Amongst these folks a BO acquisition is hard to swallow for two reasons (1) BO's architecture will require significant investment to move from BWs InfoCubes to Universes (2) BO doesn't have native support for BW data structures the same way Cognos has. This is why from a BW perspective, Cognos would have been a much better buy.

At SAPPHIRE in Atlanta this past summer Nimish Mehta's group sung the praises of the SAP BW accelerator - a strange foray into hardware for SAP (albeit with partnerships with Intel and HP). Yet as many solutions architects will tell you, throwing more hardware at a badly designed software application is only a short-term remedy. Yes, the Accelerator may have improved performance for BW but the longer term future for BW was bleak. The BW dev team being in Europe kept them too far from the important US customer base and SAP's clear lack of desire for good UI design meant that the product was a laggard in terms of features. Arguably it's customer base would have died out sooner if it wasn't for SAP reps giving it away with R/3. As BO reps become dominant within the company for every BI sale, BW will become relegated to the SAP museum. If it does, the Accelerator will soon follow.

Finally, what happens with xApps? The thought of them being deployed in BO with the maintenance nightmare of managing those Universes, is painful enough. BO's rigid Universe structure would make enterprise wide analytics (which SAP touts) very difficult especially with the array of customization choices that SAP customers have become accustomed to. Again, Cognos would have been the better choice here.

The question now is what happens to the other players? Cognos will almost certainly get acquired in the next few months, with IBM or HP being a likely suitor. With Mike Saylor still at the helm MicroStrategy seems destined to remain independent, as does Informatica. I see these companies competing more and more against the open source vendors (Pentaho, Jaspersoft for BI and Talend and Kettle for ETL) who will steal significant market share.

As always, we'll continue to watch the space to see how this plays out.