Prior to being an industry analyst, I was in enterprise IT for 15 years holding a variety of positions. I've hung tapes on mainframes, punched JCL decks, developed transactional systems, worked on IVR systems, played the role of lead architect for client-server systems, been a project leader, application architect, emerging technologist and product manager for content and collaboration-related technologies. I've worked in IT groups that reported directly to a business group and I've also worked in centralized IT structures. As an industry analyst for over 10 years I've had the opportunity to talk to literally thousands of companies. The people I talked to also had a diverse background. Some were from the business side, some where from the IT side. I've talked to CXO's as well as IT architects, developers and infrastructure managers. I have a tremendous amount of respect for what IT organizations need to accomplish in support of their business units and enterprise overall. So I just cannot help but respond to this rant on enterprise IT.
Most enterprise software is not built to sell to users but rather to IT and CIOs who are likely never to use the applications that they are selecting for their employees.
Effective IT organizations spend a lot of time working with business representatives on business and system requirements. Some organizations even have groups focused solely on adoption issues. So the implied insult here to IT organizations is somewhat audacious. Sure, there are IT groups that are dysfunctional, but many are not. Sometimes the business area itself is not as effective as it could be when it comes to explaining and rationalizing what they want in terms of a solution, the amount of investment resources available, the metrics needs to assess business impact, and how they want the system supported (e.g., service levels).
Enterprises talk about “shared services” environments all the time, as if the notion exists that their is a services marketplace behind the firewall. How many global 2000 companies allow their employees to select what HR system best serves their needs? How about email or mobile devices?
The notion behind shared services is often driven by a desire to improve operational efficiencies through economies of scale, policy enforcement, lowers costs and so on. But through reuse and shared know-how, the hope is that IT organizations can be more agile by exploiting existing capability to support business requirements that are often similar in terms of infrastructure needs. The important point not brought out is that the degree of what constitutes shared services can easily vary from organization to organization. Some enterprises can have a very strong center, others elect to have a more weak center. This gets to decision rights between business units and "corporate". The degree to which organizations decide to let business units differentiate when it comes to technology services is a governance issue. IT groups are a reflection of the social contract business leaders arrive at concerning shared services. Business management ultimately decides how much flexibility workers have when it comes to personal devices. IT groups do play the role of a trusted advisor. A poor relationship between business and IT can lead to IT "telling" business what is doable or not - but a healthy relationship is more bi-directional where exceptions are handled as needed. However, I would agree though, that over time, people will come into a workplace with a computing lifestyle and IT organizations need to abstract more and more from platforms and certify what capabilities need to exist on a machine before it becomes connected to the enterprise network.
People in large companies are looking at Google’s Premier Apps and saying “jeez, $50 a year for 10gb of storage, that alone is worth it” because they well understand that they can buy a 100gb drive at any Best Buy for $100, so why pay EMC $10,000 for a 4×500 storage array? Enterprise technology vendors are hoping against all odds that we’re not aware that there is this thing called the internet where we can see what things really cost. You want to tell me that you need management software and failover capability Mr. IT guy? Well that’s your problem not mine so don’t try to shove $1,000 of “shared service cost” on me just so I can have a bigger email inbox.
I thought several of the ZD blogs did a very good job at the "small print" when it came to Google Apps and the various disclaimers on service levels and security ("hey, we're not responsible if anything goes wrong"). Again, business management funds IT and defines to IT groups what requirements need to be satisfied in terms of compliance, discovery, business continuity and so on. Business leadership today can simply say that they don't want to pay for archival and records management. They can also elect to pursue a SaaS direction.
Niel makes a good argument about how IT can no longer use the hostage strategy to force me to pay for their expensive infrastructure, as in “your box is in our datacenter and if you don’t pay we’ll send you the SATA cable cut up into a thousand pieces” so they are now using the fear strategy as a fallback. Hosted services aren’t safe and you don’t know who is looking on Google’s servers so you can’t trust them. What about compliance?This argument has evolved from “you mean your going to trust your mission critical data to some third party who might have a datacenter powered by hamsters running on wheels?”, because who is really going to question Google’s chops when it comes to running a world class datacenter?
Depending on your industry, there are legal implications with use of outside services for sharing of certain confidential information and data. Google has made it clear by the small print around Google Apps that they are ducking when it comes to service levels and accountability when "bad things happen". For some organizations, a SaaS model is the right solution. For others, it is not. The truth is that many large organizations will have a blend of software and SaaS. Ultimately however, business leadership needs to take ownership of IT costs. For instance, I've talked to organizations that complain about high e-mail costs. When the discussion moves deeper into "current state" I sometimes find out that he organization is running on an old version of software that prevents consolidation of servers and uses much higher bandwidth than current software versions. This also usually means a very old topology design. Now maybe a SaaS option is the right move forward. Or, if there are requirements for on-premise (security, discovery, records management) then investing in an software upgrade could save the organization money over time. If business units tell IT that they are not going to pay more than what Google offers then IT groups need to respond by delivering a service at that price and at that service level. This will reveal what capabilities are needed by the business that Google does not support or will help the business eliminate waste and overhead in their e-mail infrastructure. The challenge is to have the conversation and to go through that journey of what capabilities are critical, optional, etc so that a true total cost of ownership is identified that can be compared to a SaaS option.
From where I sit the best thing that could happen to enterprise IT, at least for users, is that IT becomes an infrastructure services provider inside the enterprise. In other words, they keep the power on and a dial tone on the line, while business units take greater responsibility for selecting and managing their hosted applications. IT should also be responsible for legacy systems, after all they are responsible for building them in the first place. If business units are responsible for the cost of providing the applications they use, the actual cost and not the “shared service” cost that is used to pay for all the software that is bought and not used, then they will make better decisions and be more accountable for bad decisions because, after all, it’s coming out of their budget and they have direct control over the decision making process.
When there is a lack of trust, cooperation, shared ownership, sense of community and governance spanning business and IT, bad things happen. Enterprise 2.0 does not change that situation. Indeed, it could easily exacerbate it. Giving business units rights to do whatever they want is a governance issue. Some organizations that I have talked to have principles that promote independence between its lines-of-business. I've talked to other companies where it is absolutely critical that they are act in a unified and coordinated fashion. Shifting all the blame to IT organizations is simply ludicrous. IT organizations plan, build, run systems based on business requirements and within a policy framework that is jointly determined at a management level. Sometimes that management level is decentralized, distributed or centralized - there are many variants - but it all comes back to business/IT alignment and the continual engagement involving all parties that is necessary so we avoid the "us" vs. "them" mentality that really doesn't solve anything in the long run.