One of the main drivers for cloud is cost. The technology is seen by many enterprises to be a way to run services more cheaply, partly because of the pay-as-you-go model and partly because of the shift from a capital expenditure cost model to one of operating expenses. There are, however, instances where organisations have got the calculations wrong and ended up paying more, examples where savings have not been as high as expected and there are the large numbers who will have wondered whether they’ve made the right decision.
A major part of the problem is that it’s not easy to ascertain what the return of investment (ROI) is. The point of moving to cloud is to transform the business but that very transformation means it’s not always easy to see where savings can be made – or what additional costs there are.
As Alistair Maughan, partner at law firm Morrison & Foerster LLP, points out: “ROI is hard to measure in the cloud area because it’s difficult to accurately assess and assign a value to the extra ‘hidden’ costs of cloud adoption. It’s easy to compare a monthly AWS cost to what the previous datacenter spend might have been. But what about the migration and transition costs? Or the extra training costs? Or the investment in additional IT workforce needed to police compliance with privacy regulations? Or the cost of running extra back-ups that may not have been needed before? Or the possible cost of downtime with the cloud provider?”
There aren’t just the actual costs to bear in mind – there are some intangibles to consider as well. “How do you price user or customer satisfaction that might come from the move to cloud or the extra business flexibility that accrues to the business?” asks Maughan. “Businesses need to think laterally and consider the hidden costs and benefits, not just the obvious ones. “
Daniel Steeves, a business consultant at Beyond Solutions, says companies often underestimate the extent of disruption and don’t look at the far-reaching effects. “One of the key things missed in looking at cloud ROI is to look at what systems / solutions or investments are being displaced as cloud replaces them. Often businesses look only at the new and, unless the existing systems are taken into consideration they are at risk of misstating ROI, since there can be a resultant reduction or loss of return on those existing investments.”
The effect of the accounting change from capex to opex should not be underestimated either. According to James Munson, director of digital services and technology at the Driver and Vehicle Standards Agency, this plays a vital part in assessing the effectiveness of cloud.
“From me the ROI for cloud should be looked at like any other investment. One difference is the cost is often monthly rather than a high capital expenditure which suits some businesses but not all. Sometimes I see pre-payment discounts available which would be more like capex. I would look at a total cost of ownership of the investment over the life of the application, perhaps over three or even up to five years. This can be compared between providers or against an on-premise or dedicated hosting solution.”
He stresses the period is important. “Some other factors to take into account will be growth over time which will increase cloud costs if more hardware is needed. But this may be able to be offset by planning in elastic load balancing - only pay for what you need. If done well this will have a major impact on running costs but careful modelling will be needed to allow for peaks and lows in demand to give numbers for a business case,” says Munson.Looking for instant savings that quarter – or even the next one – could be too soon. “Firms can only assess the ROI of their cloud investments after a) a period of time and b) a period of change. Time is needed to assess whether the longer term costs associated with IT investments, maintenance and licensing have reduced,” says Adrian Bridgwater, IT blogger at Forbes
He adds that cloud is extremely good at ‘vertical’ flexibility, getting more power for similar tasks whenever you need it. “If a firm can use this power through a period of change and reap efficiencies, then the ROI should appear positive. If the period of change required new application types and new data processing that required more ‘horizontal’ expansion in the IT shop, then it is less likely to have been a profitable strategic move in terms of ROI,” says Bridgwater.
There’s a knowledge gap that needs to be overcome too. Max Cooter, IT journalist and founding editor of Cloud Pro explains why CIOs need to learn some new skills. “One of the big problems for assessing how financially effective a move to cloud is that the decision to go with cloud often goes hand-in-hand with other changes: maybe there’s a departmental reorganisation, maybe there’s a new project under way, maybe there’s been a shift in working hours and practices. The question then is to separate those changes from the infrastructure upheaval.”
There are other difficulties too. “There’s no firmly established methodology that CIOs can employ because they’re often kept in the dark. Many of them don’t know the power costs of their datacenters, so how can they know what’s been saved? If you close down a datacenter to move to cloud, how do you assess the savings on real estate? CIOs have to become accountants rather than technologists,” says Cooter.
As many businesses are finding out, data is everything. Companies have invested huge amounts in finding out more about their customers but there’s more to analytics than helping customers, CIOs need to implement a methodology that can help them make accurate assessments.
Richard Absalom, senior analyst with Ovum's enterprise mobility & productivity practice, explains the significance. “With any cloud or mobility investment, usage analytics is key. It’s hugely important to know how workers are using the tools and applications provided: what the rate of uptake is, how often they are used, at what time, for how long, etc. Individual metrics must be applied too, for instance working out the number of transactions per user going through on a CRM system. But the main point is that understanding precise uptake should allow for solid numbers, show how the investment is improving processes and making people more productive – or otherwise – and calculating ROI based on that analysis.”
One of the other factors to be considered is making a move from cloud provider. What happens if you want to move from one to another, or to take services in-premise again?
Alan Mather, director at business consultancy Ardea Enterprises, says you should take this account.“Build in potential transitions in the future: maybe you like “Huddle" today but you will want to use something else in the future - maybe Huddle will fall behind, or be taken over or you will decide it isn’t giving you what you need and want to try something else out – and that’s even more true if you are using IaaS.”
And he says CIOs should also think of the software they have installed to manage the cloud implementation. “Consider the tools that you will need to manage your cloud investment - if you are using IaaS you may end up with tools that are specific to the provider you have chosen, making it harder for you to move in the future. That means the investment in those tools could lock you in (exposing you to price or service risks in the future).
And what if the company is new; how can you assess the impact that cloud has? Phil Hochmuth, director, mobile workforce strategies research at Strategy Analytics, agrees that it’s not an easy question to answer.“With newer companies, there is no history or baseline for IT costs. Peers and networking are crucial here, and the simple question, ‘what are you paying for X?’ goes a long way in terms shaping cloud decision making. Older firms of course have histories of IT spending and trends. Create scorecards of previous IT on-premise hardware/software initiatives and analyze what went well, what didn’t, and where costs were unnecessary. Then take this model when shopping for cloud/SaaS platforms. The differences will likely be startling.”
On the other hand, cloud is an ideal vehicle for start-ups. “One good thing about cloud is the ease at which you can generally get started. However I would caution against just building though unless it is only some prototyping. The use of a secure, scalable, flexible cloud architecture is invaluable. Ideally being able to maintain consistency between environments, auto build environments - stamping them out will be important and if needed the ability to orchestrate across providers. This will pay dividends further down the line when the service and therefore cloud infrastructure has become more substantial. And without it, a costly re-build may be required as it becomes harder to keep environments consistent for frequency releases,” says James Munson.
And Steeves says there are other factors to consider.”Moving to cloud often entails new relationships (and contracts) with new providers, which need to be managed (so those costs need to be considered as well). Finally, there exists the risk when moving to multiple providers that the integration / handling of issues between those providers may become the responsibility of the business / IT department rather than a common single point of contact provider, which can also have an impact on costs.”