Cost Savings as the only driver for the move to the cloud?
Is it not what we all get drilled into us each and every day that cloud computing is just cheaper and saves a whole lot of money. Some providers even drive this to the extreme and advise customer to drop functionality (one size fits all) for the sake of the least cost in market. The question now is not whether the individual cloud service is cheaper but whether you should take that at face value and take your decisions just based on this.
Sounds like outsourcing, doesn’t it?
It surely does to me. I remember when I stepped into the outsourcing business the major line of reasoning has been cost savings on all fronts. Over time customers have learned either through own experience or by watching other projects fail that cost saving, while it has an impact into the overall business case, should not be the only driver towards a decision. There are other bits and pieces of equal importance and it is about expectation management as well as clear contracting on top of these.
Why do I bring outsourcing to a discussion about cloud computing? The reason for that is that cloud computing services do fall into the same category. The services with the label cloud are an additional sourcing option for a CIO and should be evaluated with the same methods. Have a look here for some details on evaluating SaaS proposals: https://clouddiscussions.wordpress.com/2011/09/23/how-to-evaluate-saas-proposals/
Why cost comparisons fail
First of all, the biggest roadblock prior to a deal is the cost base to compare with. Often the internal cost base is unclear. The people creating the cost analysis might be biased in fear of losing their jobs or just for the fact that they believe a substantially higher cost on internal delivery would make them look bad. The cloud provider on the other hand is inclined to paint a best case picture in order to win a deal. But reality will bite back and that best case picture will not come true. And actually I believe nobody is to blame for that as some things are just not clear at the time the contracts are signed. It is though up to the customer to think ahead, evaluate the risks about missing the best case scenario and create some reserve funds (money and time).
The second thing that comes into play are secondary cost saving expectations like easy access to skilled resources in outsourcing or a reduced effort with regards to service provider management. A cloud provider needs to be managed pretty much in the same way as a hosting or outsourcing vendor. And this stands even in case of online information instead of service manager meetings. In addition you might want to consider even to plan slightly more for this and go for an extra offering that will make the cloud service look slightly more like an outsourcing. Have a look here for some ideas: https://clouddiscussions.wordpress.com/2011/09/12/where-the-cloud-needs-to-learn-from-classic-outsourcing/
The next area of uncertainty is the migration as well as updates. While you can get a good grip on potential migration cost you need to make an effort to do so. Thorough testing and detailed planning is needed before you can really get a good estimate on the costs. And do not forget to include the cost for these tests into the business case. Updates are a little harder to plan for. While the cloud service provider will tell you that they take care of all the updates and that it will be without extra cost for you this is only half the truth. Surely updates will be cheaper per se but you cannot determine the timing any more. The update takes place when the provider says so. The deeper your integration is, you might be forced to upgrade your internal environment as well, You might not have done that in an on premise environment. Examples are updates needed to your directory services as it is federated with the cloud or certain versions needed to use the service outside of the web interface. So you might face the need to update client software as an example.
Another factor to weigh in is the network element. With a cloud solution even traffic that was formerly internal only now is routed over your internet connection. Does this connection in speed, bandwidth and availability stack up to your expectations and the cloud service? Remember in SLA chains it is the weakest link that counts. So what are the additional costs for the network you need to calculate?
What are the reasons for going into the cloud then?
What is the difference between a good date and a bad one? It is your expectations. Rushing to the cloud for cost reasons only is a little bit like going out in expectation of a hot date and your imagination just setting the completely wrong expectations.
There are many good reasons to move to the cloud from added flexibility to simplicity and not forget cost. Rather for the best cost you should look for the service that fits the requirements best and be quite detailed on your cost expectations. Actually even though I have painted a rather black picture I do believe that cloud computing delivers cost savings. But you need to use it in an area that makes sense. Imagine you decide to go for cloud services, do a migration and rather than have a happy bunch of users and watch the savings stack up you need to revert back. That would involve additional cost for another migration and rather than savings you have created additional cost. Search for LA police department and their failed attempt to move to the cloud!
Bottom line is that it is really about a clear management of expectations from the beginning on the customer side and a need to plan for a little bit worse that the best picture scenario on the provider side.