Beware upgrading enterprise software: total cost can be many times the licence fee

As Oracle and SAP vie for unsettled PeopleSoft users, what does the future hold? asks Bruce Richardson

New Asset  

As Oracle and SAP vie for unsettled PeopleSoft users, what does the future hold? asks Bruce Richardson




There is a lot more to buying software than just paying for the licences. The real catch for many long-time enterprise resource planning customers is the continuing cost of upgrades.

One of AMR's clients is looking at upgrading its ERP system. The initial assessment is that the upgrade will provide a 10-times return on the cost of licensing new software. That sounds great, until you consider that the cost of doing the upgrade is more than 75 times the software licence cost.

Another client said the high cost of ongoing ERP upgrades has persuaded the executive team to switch to business process outsourcing. In this case, the client lets the offshore supplier worry about the upgrades.

A third client complained about the size of the ERP software and the complexities of upgrading. He said it takes his ERP supplier a year to make the new release stable, another year to plan the implementation of the new release, and a third year to install and go live. Although there may be some hyperbole here, this ends up as a two to three-year cycle, which limits innovation.

So, if you are working with a hybrid SAP/PeopleSoft or SAP/JD Edwards system, do not let emotion make your decision. Do the maths and factor in the lifecycle costs.

I would wager that SAP's acquisition in January of TomorrowNow, the support provider for firms using JDEdwards software, will have minimal effect on revenue in 2005. (SAP will be somewhere in the £4.7bn to £5.2bn revenue mark this year.)

I recently ran into someone who works for a professional services firm with an existing investment in PeopleSoft. In his first meeting with his new Oracle sales team, he was informed that his maintenance would be going up and that the increase covered support and "innovation". When he asked about innovation, he was told this would cover new actuarial software being developed. His firm had no need for that function.

He would be a logical candidate for TomorrowNow's services. In some ways, software maintenance becomes analogous to car insurance. You get a bill with a hefty increase and the logical response is to shop for comparable coverage at a lower price.

Will SAP really be cheaper? It will depend on the individual account. But no doubt SAP will be capturing all the data on every prospect calling for a quote. It will want to know what applications and release levels they are running, when they bought their software, the number of current and planned users and their level of satisfaction with the software and service. You cannot put a price on that data.

Some time later this decade, it is probable that service-oriented architectures will move from Powerpoint and statements of direction to reality. On one side we have Oracle and Project Fusion. Across the ring stands SAP and enterprise services architecture. To be successful, both heavyweights will have to recruit partners for their respective corners. The first two that come to mind are IBM and BEA.

IBM is an obvious choice. If you are competing for future hearts and minds, you want to leverage IBM's consulting and infrastructure expertise and customer base. Support from Hewlett-Packard and Sun is important too, but IBM remains the first pick.

But BEA will probably have a working service-oriented architecture product set ahead of Oracle and SAP. It does extremely well in financial services, one of the few vertical sectors not dependent on an ERP backbone. Many of the early adopters of service-oriented architectures will come from this sector.

At the Netweaver launch a few years ago, a SAP executive acknowledged that one of the intentions of Netweaver was to function as "the BEA-killer".

Oracle chief executive Larry Ellison has long expressed interest in either acquiring BEA or eliminating it. Now, BEA's products and growing developer community are critical weapons.

The reality is that BEA assumes a more strategic position in the next generation of web service-based, heterogeneous computing. This is especially true if users follow through on threats to reconsider best-of-breed as an alternative to three upgrade cycles and costly complexity.

Microsoft's role in the enterprise service-oriented architecture market is unclear. It is hard to imagine Oracle embracing Windows as an alternative user interface to its plans for dynamic HTML, but this may be worth hundreds of billions of dollars.

It is also too early to say what role Accenture and the offshoring companies will play in the service-oriented architecture market. If I were running Oracle or SAP, I would want to make sure my executives had been to India by the end of this year.

Bruce Richardson is an analyst at AMR Research

Read more on Database software