ROI and ROA
by David Merrill on February 5, 2010
Lately, people have asked me about the difference of ROI and ROA, both techniques that we use in defining storage economics. Here is my IT-econ perspective.
ROI is a mechanism used by CFO and financial people to view the viability of a given investment. Sometimes when capital is tight and there are several projects competing for limited funds they may compare the ROI of the projects to see which is best. ROI is used before an investment is made, to check that the payback and projected savings ‘clear the bar’ a company has set relative to savings, business improvement etc. When business is great, and lots of capital spending is going on, ROI is a very popular metric to track priorities and quality of new investments.
ROA tends to be a mechanism to review after the investment is made. Many CTOs and CIOs are concerned about ROA of the existing IT infrastructure. Poor utilization, limited useful life (for example long time to migrate on or off an array shortens the useful life of that asset), or poor management reduces the ROA of the asset. ROA is measured in terms of business impact. Again, it has already been bought, so ROA is a measurement of the effective usage of the asset.
Hu Yoshida (as well as many analysts) talks a lot about ROA because people want to maximize the assets they already have. With a bad economy, there is limited new CAPEX so ROI is not always as relevant. In order to reduce new CAPEX people need to better use what they already have, hence the interest and discussion on ROA. There is a need to improve the return on assets already in place. ROA is a very, very strong story for HDS right now because we can help improve, through heterogeneous storage virtualization, what a data center may already have in a sunk (storage) investment.
A local small town barber noticed a discount/chain $6 hair cutting salon going in across the street from his shop. He started losing business so he went to a local banker friend for advice. After a few hours of dialog with the friend, he returned and put a new sign in his store window. ‘We fix $6 haircuts.’
ROA is not about ripping up and looking for a new ROI, it is about improving what you have. Many IT storage are running on an inferior ($6 haircut) storage architecture.
- Migration costs are high, mobility within dynamic tiers is complicated
- DB copies reside on the tier in which they were created
- White space and stranded space rates are very high, and these levels of waste are expanding as new capacity is added to the old architectures
- Storage management is stretched beyond reasonable limits, current architectures and the growth is un-sustainable in many environments
- Storage keeps increasing with capacity, but the IO rate per drive is not changing with these higher capacities
- Higher densities and lower prices presents the false sense that total storage costs are going down, when in fact they are flat year-on-year or rising
- Power, cooling and floor space costs tend to exceed the initial purchase costs, so they become more important in the TCO measurement and the carbon footprint needs to be an integral part of the storage architecture decision
Our four-year old, proven virtualization solutions can help fix the existing infrastructure. This transformation is not free, but it does improve on the sunk costs our customers have with existing kit. Don’t be fooled by the $6 haircut…
Comments (1)
David Merrill’s Blog >> Blog Archive >> What was of most interest in storage economics in 2010? on 27 Jul 2011 at 4:33 pm
[...] 2. ROI and ROA [...]



