The commoditization of technology has reached its pinnacle with the advent of the recent paradigm of Cloud Computing. Infosys Cloud Computing blog is a platform to exchange thoughts, ideas and opinions with Infosys experts on Cloud Computing

« How enterprises have benefited from the cloud ? | Main | Hadoop – the Open source BI+Data warehouse solution »

The Cloud ROI Framework

Last week, I had an opportunity to discuss the cloud computing ROI model with a large banking major.  We did some illustration using the simplistic ROI framework and figured-out the savings to the scale of 90% on CAPEX and 50% OPEX cost Y-o-Y.

The numbers were attractive enough to get attention of “C” profiles, he was bit positively surprised. The first question he asked was: what’s the trap here? He understands the concerns of security and outsourcing in banking business better than me, so I need not attempt raising technical obstacles. I simplified the answer: It is paradigm-shift; IT needs to manage this change! To my surprise, the wave-length matched just here, and we had constructive meeting.

We discussed the future of IT is in cloud, it is a big wave like moving from DOS to Windows; Desktops to Internet and On-premise IT to Cloud; through iterative lean IT transformation. We were on the same-page throughout, we concluded meeting with a note saying that – enterprises will adopt cloud tomorrow if not today; but if they fail to do it, they will be forced to adopt.

The Cloud ROI Framework

I will present an illustration of cloud based ROI model, I am not MBA, I used a very simple framework, maybe it is over simplistic, please bare-with my expertise on this. You can help me with the flaws in the model including depreciation, tax benefits, Y-o-Y savings and NPV – I don’t understand this well enough.

Excluding migration cost, it releases the 90% of the CAPEX investments and reduces OPEX cost by 50% every year. I know, there are multiple assumptions in this model, also not every piece of data might be accurate, however the range of benefit is still attractive, we can make refinement & correction in given customer context.

The Financial Model

I also tried to work-out some sort of financial model for Y-o-Y savings and what-if analysis around cost of migration.

You may want to read my previous blog entry on The Economics of Cloud Computing for some more horizontal business cases. You may also want to know How enterprises have benefited from cloud in my earlier blog entry.
 

Comments

Great initial work. Lots depends on the assumptions regarding computing power, storage, and bandwidth. It would be helpful to see these discussed at greater length. Also, your software maintenance and infrastructure maintenance estimates for on-premises seem quite low.

Thanks a lot Peter, it is my pleasure having a industry veteran leave a comment on my blog.

Yes, as I mentioned, it was an illustration - the intent was to present the initial framework. I agree with your point the assumptions play a key role - I tried apple-to-apple on both the sides w.r.t. computing hours and storage. It can turn the tables, specifically on Y-o-Y savings % cost of cloud HPC is high.

I would also extend your comment - it will make more sense if we can "contextualize" the ROI framework to the customer's existing IT infrastructure. If we can present ROI, feasibility analysis and transition roadmap - it can earn a multi-million dollar project.

I don't think I was clear enough: a spreadsheet, even if just an initial framework and meant to be illustrative, needs to have its assumptions and definitions crystal clear. How, for example, are you determining "units" for computing power, storage, and bandwidth in the cloud? It's not at all clear to me that it's apples-to-apples -- how would I tell? Also, you're assuming that what takes 8 people to administer today will be administered by just 3 when moving to the cloud -- can you discuss and justify?

By the way, your total sum is missing the initial labor component in the cloud column (minimal, but still there). Some will quarrel with your categorization of labor costs under CapEx, but my understanding is that this is for implementation, and thus capitalizable. In any case, you need to explain a spreadsheet like this, more than just present it and hope that it's self-evident.

In short, this framework, and discussion around it, would benefit from a clear list of assumptions and definitions.

Thank you Peter for clarifying, I understand the issue now.

The bandwidth, power, etc are pure guess and no scientific estimates or assumptions for the same; similarly for storage also to an extent some guess work - as we will not be using 100% of the back-up capacity; for computing hours, I assumed 4 CPU machines on-premises.

We can always revise the numbers & calculations to see the new ROI in this framework. I do also agree that not always cloud will give positive ROI y-o-y.

Also, my intent was to present a framework where-in we should be able to put "actual or estimated" customer data instead of assumptions to calculate ROI - definitely with accurate calcuations. :-)

Lastly, thank you for spending time on identifying other calculation mistakes and providing constructive feedback. You can blame my mathematics and/or excel skills. Similarly, the per unit costs can also be different in various cases / (assumptions).

There is a different school of thought for the Capex to Opex model of Cloud computing

http://gevaperry.typepad.com/main/2009/01/accounting-for-clouds-stop-saying-capex-vs-opex.html

Thank you Sankar for sharing this - It is definitely an interesting read & perspective.

The thing is - "people are thinking that cloud can give cost benefits" - and that is the whole point. Refer to my earlier post on "how enterprises have benefited from cloud ?"

"How much ?" and "really ?" can always be answered in more qualified & realistic manner in the context of client situation.

It may or may not have positive ROI, and we can always have better accounting and/or technology people than me :-)

Post a comment

(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)

Please key in the two words you see in the box to validate your identity as an authentic user and reduce spam.