As more businesses consider turning to the cloud for their storage and backup needs, calculating ROI becomes increasingly important. In this expert podcast, Andrew Reichman, Principal Analyst at Forrester Research, offers advice to users who are comparing cloud offerings. Levels of service, transactional charges, and the cost of building your own cloud are among the many items to consider. At the end of the day, Reichman says, the perfect cloud storage ROI equation may be out of reach. But, he says, you can get close.
Download for later:
- Internet Explorer: Right Click > Save Target As
- Firefox: Right Click > Save Link As
SearchCloudStorage.com: Based on your recent research, how difficult is it to calculate ROI when a company is considering switching to the cloud?
Reichman: It isn’t that difficult, I think, if you isolate the cost driver and what you might spend in the cloud – that part isn’t that hard. The hard part is that a lot of these cloud offerings are fairly immature, so it can be pretty difficult to know what you’re getting and to determine whether a cloud offering that you switch to, from internal infrastructure, is an apples-to-apple comparison.
A real pitfall with ROI is that people try to get perfection. They end up with a model that’s too complicated or they don’t make any comparison because they’re forever looking for the right numbers.
Andrew Reichman, Principal Analyst at Forrester Research
Which numbers must users have a handle on to get started calculating?
I just published a document that measures the cost of cloud storage for a year -- compared with internal file storage for a year. The way I went about it was to take all the costs associated with doing it internally, building a storage environment, buying a storage capacity, hiring the staff, any data migration, power and cooling, and maintenance. All of those things are what goes into your baseline comparison. Then the cloud side of the equation: what would it cost to store that same amount of data in a cloud repository?
That math is actually easier – most cloud providers are giving you a dollar-per-gig-per- month rate. There are a couple differences here and there – you may have increased prices because you’re sending more data across the WAN. A lot of the cloud providers do have transactional charges, the amount of data that you send to and from the repository for every given month. And that can be a bit of an unknown or unpredictable charge. But it’s not huge.
If you don’t know the answer to all of the questions in your calculations, put in your best guess. You’re never going to get it perfect, but if you start and do your best and refine your assumptions over time, you’ll get closer to an accurate ROI comparison.
What are the most difficult factors in calculating cloud ROI?
Any sort of financial comparison is an art and not a science. A real pitfall with ROI is that people try to get perfection. They end up with a model that’s too complicated or they don’t make any comparison because they’re forever looking for the right numbers. Use your best guess and make this a pragmatic approach is one thing I recommend.
Another thing to avoid is leaving things out – when you think about what you’re comparing cloud to, you get facilities, you get people and you get data redundancy. Really, what you need to compare these to is the total cost that you would spend on doing this yourself.
The third factor: is there a reasonable cloud alternative to what you’re doing internally? People might look at Amazon S3 and say ‘Okay, that’s cloud storage. I’ll compare it to my tier 1 primary storage and it’ll look much cheaper,’ but it might not be an apples-to-apples comparison. You may not get the same level of service, and it may not be a valid replacement for what you’ve done internally over the past several years.
What ROI results have you seen among users with users who have switched to the cloud? Are they accurately calculating their ROI?
I don’t think I’ve talked to a lot of companies that have done an an accurate cloud ROI comparison. Not that many enterprise companies have moved active primary data to the cloud, it’s still pretty early on in this process. My impression is that they have some sort of mandate or they are excited at the prospect -- and want to be an early adopter. But it’s not necessarily a scientific comparison that led them to do so.
For more mainstream adoption, that’s one thing that has got to come together. It has got to be more of a clear comparison of scenarios – continuing to build your own infrastructure or sending your data to somebody else. Making that scientific comparison is the way that more companies are going to make this decision.
Are there any ways or situations in which ROI could drop or disappoint?
In my comparison, cloud storage is very attractive. The biggest driver is that when you buy internal storage, you’ve got to buy at raw capacity and enough of it to support data redundancy. That becomes a very expensive proposition, especially because of inefficiency.
When you go to the cloud, that’s part of the service. All of that excess raw storage capacity comes with it and you only pay for the primary data that you’re sending to the service. They absorb any inefficiency and redundancy that’s needed to make that service viable. That’s why it looks so attractive.
A way that the ROI case could be less rosy than you think it is? It might be that the service that you’re comparing it to isn’t enterprise class and doesn’t have a high enough quality of service. Another way would be that you have to pay more for a higher quality service, and that could become a hidden cost. Those transactional costs could be hidden, and that could diminish your rate of return.
There are also add-ons. I’ve been looking at the cloud gateway vendors who offer products and services that can make the cloud storage experience better – more enterprise class, more reliable, more local data caching, API translations, security features . They can make the cloud experience closer to what you do internally, but it does make it more expensive. You add an ongoing or an upfront gateway charge, and that can diminish your ROI.
The final factor is how much storage you consume. If you put a cloud storage model in place and your users realize you have easy access to unlimited capacity, you could expect to see more storage capacity growth than what you had internally. That’s something we’ve seen with virtual servers. When it’s really easy to get a virtual server set up, people use more of them. I think the same thing could happen. What if, because it is easier and cheaper to access, the 100 TBs you’re storing become 300 TB? That could chip away at the ROI, but it’s not necessarily a bad thing.
This was first published in August 2011