How well do you manage what your business spends on cloud computing? It can be difficult to work back from a single monthly invoice to find where money could be better spent. That's why New York-based insurance giant MetLife has worked with technology business management (TBM) vendor Apptio to create a cloud reporting tool that tracks spend as it's incurred. As Rick Eckoff, Director of the TBM Office at MetLife, explains:
If we didn't have this in place, the invoice comes in every month and you try to decompose that — [whereas] with the system we have in place, all of that is tagged upfront, and all of the data is available daily.
Project managers and application owners get a daily report that details their cloud spend, which means they can keep an eye on costs without needing direct access to the cloud provider's admin console.
We've given them an easy front-end they can look at, so we don't have to give a hundred project managers access to our cloud vendor.
It's about being able to control that expense today as opposed to four to five weeks later when the invoice comes in.
'Rule of 12x5' in cloud spend
One way this helps save money is by identifying applications that are left running overnight. "There's a 'Rule of 12x5'," says Eckoff. Development servers in the cloud should only be up during working hours, except in the rare cases where work on them 'follows the sun' across multiple geographies. Shutting down these instances outside of office hours can save 60% of the cost of running them 24x7, says Eckoff.
It's direct expense, we need to show the application development team what they're spending and give them the leverage to control that expense.
The discipline of technology business management is promoted by the non-profit TBM Council, which Apptio helped set up in 2012 out of its user community. That ability to connect with a community of practitioners in other organizations is key to the value of TBM, says Eckoff:
It really helps solidify the fact you're not going down the wrong track. The biggest thing out of this is the fact that there is this network, and individuals are coming together and talking about what they're doing and how they're doing it. You're not trying to redesign the wheel at each place.
The overall objective is to bring financial transparency to enterprise IT budgets, and "bring in the discipline of business management for the CIO," as Apptio CEO Sunny Gupta explained in an interview with us last year.
How to reduce chargeback costs
MetLife's Eckoff says that everyone is "doing some form of TBM whether they know it or not," often by manually correlating information from different systems such as procurement, accounts payable and IT services management. But the systems MetLife was using before it adopted Apptio didn't join the dots quickly or clearly enough.
Really our issue was transparency — getting to the detail information and explaining that information on the chargeback to the business. The existing system worked, but we'd spend weeks just trying to figure out what drove a change or why somebody was up over budget.
Over time we were able to change the conversation from, 'Why is IT so expensive?' to explaining to the business partner, 'How can I help reduce my cost? If I pull this lever, what will it do?'
One of the roles of TBM is to help the business understand all the elements that make up a chargeback total:
From a TBM perspective, the majority of our focus is on our internal chargebacks. That is the main reason why we are using Apptio heavily.
The application owner can see what their footprint is and they can understand the cost and the expense. A lot of people take IT for granted. There's a lot that goes on in the background to enable it.
'Follow the dollar'
But it's not just a matter of making plain the full cost of IT systems. It's also making sure that those costs make sense in the context of business outcomes. As Eckoff puts it:
We talk about 'follow the dollar'.
You want to spend $1 million on something? Give us the impact of those dollars on the business.
The TBM system also measures on-premise costs, Eckoff explains.
All of our services are in the system. We're charging back for mainframe, desktops — all the dollars that go into our infrastructure budget are charged back and analyzed through the system.
Giving managers insight into their cloud spending is especially important because this is direct spend that's very responsive to changing demand. Eckoff comments:
It's an expense stream that we need to make sure that we manage, because it's large enough to be an important number that we need to track.
Cloud is one of the few levers you can pull within infrastructure that allows you variability. A lot of infrastructure spend is sunk cost. Cloud is variable. That is one of the reasons we felt it is important to track this.
Break-even point for going cloud
TBM can also help to establish which applications make sense in the cloud, and which don't.
We're able to look at and base some of our technology and layout of how we are going to fulfil a request based on those costs — internal versus external cloud costs — and we're able to make valuable decisions based on that.
The consensus among TBM practitioners is that hosting internally starts to come up cheaper than cloud once loading passes a certain level, says Eckoff: "There is a break-even point somewhere in the region of 16-18 hours a day." But the decision has to take other factors into consideration too. Having full transparency on costs helps inform that decision making, says Eckoff:
We do want to make sure we're putting the right applications on the right platform.
It seems like an obvious thing to do, but as this story illustrates, it's not an easy task to allocate IT costs to specific applications within a large enterprise. The economies of scale that come from a centralized IT buying and operations function have a tendency to obscure who's actually spending what.
This is hard enough to track when spending is relatively fixed, but as more and more IT moves to a pay-as-you-go basis, the spending landscape is constantly changing. These challenges, by the way, are faced not only in IT, but the number of moving parts in an IT infrastructure probably makes the problem especially tough to resolve.
It is one that has to be faced. Without adequate information, it's quite possible that an organization could be spending far more on certain systems than is justified by their business value — and not enough on others that could make a crucial difference to business outcomes.