Three very different data points over the last month contribute to this post
a) I was impressed to see at the Dreamforce conference last week how much custom development is going on on the force.com platform. The numbers we heard from Marc Benioff were big - 188 million lines of force.com code, 12 times as much as last year. Since June more than 15,000 companies have a signed up for the force.com free edition etc.
But the numbers are just numbers till you see what end-user organizations are custom-building – not just vendors like Coda to build their own accounting applications. You have to stop and ask - has the pendulum swung? Are people building more rather than buying packaged functionality?
b) Then I saw fellow Enterprise Advocate Ray Wang present at a SAP user event in Manchester, England and point out 5 areas SAP has invested in recent years.
- Netweaver
- Duet
- BusinessbyDesign
- Solution manager
- Enterprise support
But there has been little customer uptake. Alan Bowling, chairperson of the UK and Ireland user group is quoted: “SAP still develops stuff that only five people will ever use and then tries to drop it on us. ” No wonder SAP has had a tough job migrating customer bases. Why should they with new features they don’t really care about, and the chaos of a typical upgrade. 3-4 months of rehearsals, and the hold-your-breath long weekend execution of the upgrade. In contrast, Marc Benioff also harped on salesforce’s ability to perform“5 minute upgrades” using mirrored data centers.
c) Finally, a conversation with Zach Nelson of NetSuite brought home much more clearly how much more SaaS vendors understand their customer base. When you can see on-line which features your customer base is using extensively, and which new features get adopted in an upgrade release, that is one of the best feedback loops you can have. In on-premise world that feedback takes much longer, and the results typically based on surveys are never accurate.
Alan was further quoted as saying “We can now get in early and tell (SAP) that some things are either irrelevant or might be better if approached from a different angle. We’re gaining access to advanced roadmaps that would have been kept from us”. That is goodness, but it has taken years and will continue to take years to incorporate into future releases. How do you compete against Zach’s much better visibility in to his customer preferences?
More custom development. More tuned features in future releases. 5 minute upgrades.
The packaged software world is going through seismic changes.
This post was previously published on the Enterprise Advocates page
HP, the disruptor?
Why would Cisco move down into 25% mainstream server margin business when its networking business has yielded it 65%? That is the question many asked as Cisco announced its Unified Computing System in early 2009. Its marketing said it “..represents a radical simplification of traditional architectures, dramatically reducing the number of devices that must be purchased, cabled, configured, powered, cooled, and secured.”
Some would argue, a new class of infrastructure software – virtualization – had broken down traditional boundaries between computers, storage and networking and Cisco was merely reacting to that trend.
Events in recent weeks are confirming HP, in particular, pushed Cisco to do so.
Before Mark Hurd arrived at HP there was a strong partnership between the two with HP selling hardware, Cisco networking gear. Indeed, HP’s previous CEO, Carly Fiorina was on Cisco’s board. After Mark arrived, HP started to promote its own ProCurve networking brand while continuing the partnership. HP storage networking solutions were proposed most of the time with Cisco high-end switching directors.
Grumbling started when HP introduced its Virtual Connect product in 2006. The industry had been moving to blade servers – easy to expand by plugging in or out new blades into the chassis. But while this made easy for the server administrator, they had created problems for network and storage administration. With Virtual Connect, various network addresses were virtualized inside the module, and so are the links back to the server blades on the other side. So changes in server configuration did not affect network or storage administrators, and vice versa changes in network or storage did not affect the server administrator.
HP claimed that Virtual Connect could increase the productivity of administrators by a factor of three because of this flexibility. The foundation for next generation network virtualization was laid. Rumors started floating that HP saw an opportunity in networks and Cisco saw the HP threat
And as it introduced its Virtual Connect Flex-10 Ethernet module in 2008, HP changed its messaging from administrator efficiency to economics saying it could “cut network hardware costs up to 75%.” Flex-10 distributed the capacity of a 10-Gb Ethernet port into four connections to dramatically lower the cost and complexity of network connectivity.
And by late 2009, the split was accented as Cisco set up a joint venture to tighten relations with EMC around storage HP started more aggressively positioning its own ProCurve networking gear and announcing its own Converged Infrastructure Architecture. Now comes HP’s $ 2.7 billion acquisition of 3Com and it is all-out war.
I am not nice to HP for milking its dominant position in printers and related consumables. Its EDS unit is grossly overpriced compared to cloud competitors. Its software business gets unjustified margins.
But in reverse I need to be generous to HP for disrupting Cisco in network switches and routers.
November 16, 2009 in Industry Commentary | Permalink | Comments (1) | TrackBack (0)