Demand management has been the unicorn of enterprise IT – something frequently talked about but rarely seen and never captured. Every centralized IT organization would love the ability to accurately manage user demand. It would provide tremendous return if it were possible; but to date it has been largely or completely thwarted in large enterprise IT organizations. But there’s good news, thanks to the as-a-service model.
The reason demand management has been thwarted is that IT is organized on a functional basis; the data center, servers, network, purchasing, and security app development and maintenance are all defined functionally. IT leaders are held responsible for driving out cost and building capability that is shared by multiple departments. The problem is there is no relationship between demand and supply capability because business users don’t understand how to measure their usage/demand.
When IT asks business users how many servers they you use, their response is often “How many did we use last time?” Or when IT asks how many programmers will you use and why, business user typically respond with “We need 10 percent more than we had last time.” There is no relationship between actual demand and the actual demand drivers and the estimates they must provide to IT. This is a hopeless and fruitless exercise. It’s like a broken clock that is right only twice a day. This demand estimate is doomed to be wrong every day.
So what’s the answer? We need a service construct where IT is organized into service models. This construct gives business users a way to understand their usage or demand. For example, a healthcare payer understands how many people it expects to enroll. This is a metric the business can use to predict usage and the time frame in which they will need the service. IT can then manage the demand for the service it provides to the payer based on the number of enrollments.
When companies organize IT along service lines, they can translate business activities into technical consumption. The as-a-service construct attempts to make as much of its service chain or supply chain as elastic as possible. It adjusts each part of the supply chain to the usage demand. So unlike the traditional functional IT structure, business users only pay for what they use.
There are three mechanisms to make a technology or service elastic:
- Share it (such as AWS); when you’re not using it, someone else is
- Automate it; spin it up, do the work, and shut it down
- Buy it on a consumption basis
Typically, as-a-service providers use all three of these techniques to allow them to use their full service stack with the business metrics that the technique serves.
The as-a-service model achieves one of the Holy Grails of centralized IT – it provides a realistic demand management vehicle where the business can make accurate estimates. It also provides paying for the services only when they are used; this is the consumption-based model that the services industry is moving to.
Demand management to date has been completely illusive to centralized IT because of the take-or-pay nature of IT. This method for building capability – and business users sharing the cost to be able to use it – has no connection to business metrics that the business can control and understand how to estimate their technology capacity/demand. But the good news is the as-a-service model puts a rope around the unicorn. It creates the ultimate answer to demand management.