Can vertical integration solve more enterprise problems than virtualisation
When he was running HP, Mark Hurd talked about virtualisation and automation as being at the heart of the enterprise IT future. When he spoke at the Future in Review Global conference in Seattle recently, his view as president of Oracle was somewhat different. He still has the same focus on knowing the numbers that matter (we once saw him speak at a joint HP Microsoft event with Steve Ballmer and he had the Microsoft numbers on the tip of his tongue as well), and on getting things done. But with his Oracle hat on, he's now pitching a rather different solution.
Automation and virtualisation will happen, he says, but they won’t be enough. “I think companies want to minimise their investment in IT that does the simple, mundane tasks of keeping the business running; all the tactics you use – server consolidation, data centre management, automation of menial tasks that can be done with software. I think you’ll see a continued move towards that in a continued effort to get IT budgets for maintenance down. I think IT organisations need to shift their budget from maintenance to innovation and innovation means how do I align with the business, how do I get my money into the business and help grow revenue, help automate a task that makes us more efficient… I think you’ll see the IT industry do more vertical integration to try to lower that budget. I think you’ll see vendor consolidation and I think you’ll see capabilities that fewer companies can bring. I don’t think you can get IT budgets aligned just by doing technology tricks. Most of IT’s inefficiency comes from the lack of integration of business processes to the application suite that runs the company and the lack of understanding and magic between the two causes the inefficiency.”
Hurd blames home-grown, custom apps for the fact that many organisations spend up to 80% of their IT budget on maintenance of existing systems rather than developing new ones. “The company I just came from, HP, at our lunatic worse had 7,000 apps to run the company – and it wasn’t because we loved applications! The way those happen to you is not just through acquisition; they happen when you take a business leader and you say ‘tell me about the processes you have to run the business’ and the business leader explains to you the processes without understanding the sprawl in IT that has been created. We dispatched parts to fix computers; we had a hundred different ways to dispatch parts – it was different in Germany than it was in France. And I’d ask why was it different in Germany form in France and I'd get the feedback ‘you just don’t understand’ – exactly! I don’t understand; not only don’t I understand, I don’t agree with you and we’re going to take those hundred processes and we’re going to fuse them into one. The hundred processes spawned incredible complexity because you had a hundred different applications therefore you had a whole suite of servers, you had a whole suite of infrastructure tools and the number of humans that have to be in place to support a hundred apps is huge. Change it to one, you lower headcount and you get one or two code wraps on one application rather than a hundred and everything moves faster. Getting those processes right is huge.”
And Hurd dismisses the solutions other vendors are pushing - the solutions he was enthusing about at HP - as stopgaps for when the IT department doesn’t get the authority to do that. “If the CIO can’t get to the business table and get that outcome, then the CIO is left to do the tricks; OK, I'll get the budget down, I’ll virtualise some stuff, I'll get more powerful servers, I’ll do a better job procuring… that will get you part of the answer but to get the whole answer you've got to get the whole business fused together.”
One throat to choke It's interesting that Hurd uses the word 'fuse' - because it sounds like a clear reference to Oracle’s fusion middleware, which will tie together the vertical solutions he thinks Oracle will use to dominate the market. They're solutions that sound more like the past of the IT industry.
Hurd maintains that integrating systems from multiple suppliers has many disadvantages for enterprises and we wouldn't disagree - but he doesn't mention the advantages, of course. “The industry, if you went back 30 years, was vertically integrated; you received everything from one supplier – you get the hardware, you get the support, you get a database if that was involved, you get an application and operating system. And then – if I can use the word – the industry dis-integrated over the last ten years. Today if you've got an app, my guess would be you’ve got a bladed server from – well, pick a vendor – you've got an operating system, it could be a Linux from somebody else, you’ve got a database maybe from us, you could have middleware from yet another vendor, applications from yet another vendor. As a result of that, when you have an issue, who troubleshoots that, who takes accountability for that? The answer is ‘you do’. As a result I think there will be a move back towards vertical integration and it could be a significant one.”
So industry-specific apps (the SAP approach) but on Oracle hardware (the IBM approach - and Hurd called out IBM as both competitors and partners); that's the way he thinks enterprises get away from the 'lunacy' of all their apps. How soon? That bit he's not sure about. “The amount of IP that's custom in the marketplace, the amount that has been unique; I believe all of that transforms over the next - I don’t know if it's ten years…" But he is sure of one thing; "We're going to be a player that helps that transform.”
Part 2: Mark Hurd on why apps matter and why the cloud is not the answer Mary Branscombe