The corporate vice president for Microsoft’s Azure Marketing, Bob Kelly, made this statement in Forbes Magazine:
Cloud changes the fundamental economics of how an enterprise thinks about IT. Large capital expenditures (capex) are a burden. Cloud flips capex into an operational expense (opex), meaning enterprises only use the resources needed when they need them. This helps enterprises to focus on the things they care about. In the past, 70% of IT was spent on maintaining its capability and 30% was bringing real value. Cloud fundamentally flips this around by disrupting the cost structure and drives cost down. Cloud allows enterprises to differentiate on who they are instead of focusing on IT.
Let’s take this statement apart and see where it fits into the geospatial technology domain. I believe that too much has been assumed about the cost savings inherent in cloud computing deployment.
By history and by design, geospatial solutions are large investments in people, software and data. Sometimes there may be an emotional investment in the vendor, as well. Counting just the software, the investments by large utilities, city governments and transportation companies, not to mention federal geospatial intelligence agencies, are big ticket items. They become large capital expenditures, the capex referred to above. Their purchase may involve lengthy contract bidding and negotiations. In the end, these investments are capitalized and depreciated as assets of the organizations.
To switch this to a cloud model, making a geospatial solution “rentable” on demand, indeed flips the expense structure and creates one that the vendor industry is simply not ready to provide. The users and buyers of technology are absolutely ready but the solution providers have just begun to explain what is delivered and how much it will cost. Companies like eSpatial have been refining their message. At the recently completed Esri UC, that company provided a clearer explanation of ArcGIS Online for those moving to its cloud solution. Solutions available today are not as robust as desktop or server editions, but that may change over time and perhaps sooner than we think … or are prepared to implement.
This is a problem from a few different perspectives. The value of software as a service (SaaS), one type of cloud-based solution, inherently lies in relinquishing the care, upgrading and support of software that will eventually be out-of-date. In the old model, the asset essentially depreciates to “zero.” My father always said, “Never buy a depreciating asset.” So, “renting” makes a lot of sense. It applies to software as much as it applies to buying a car.
Now, let’s look at cloud computing as an operational expense. Does Kelly mean that cloud computing really allows organizations “to differentiate on who they are instead of focusing on IT"? I contendthat this is the most advantageous area for cloud adoption. Geospatial has been moving from the backroom to the boardroom for some time and in many cases, the IT department has injected itself into the discussion and assumed some control over purchasing GIS. This has been disruptive to the process of acquiring geospatial software solutions, but that may not necessarily be a bad thing. The change from capex to opex would make the GIS users less vulnerable to the whims of IT managers because the support issues will decrease. Hence the objections to maintaining yet another software license will likewise decline. The focus turns to applications, analysis and location intelligence. Get me the answer fast and tell me why it benefits my organization. Apps in the cloud can offer a ray of sunshine for users.
An area of cloud disruption that is somewhat unique to geospatial technology is data. Those companies that have typically sold geospatial data are now offering plans to rent or buy their data and might even allow you to sell your data through their cloud platform. Data is an asset. It’s capex and opex. Data should be considered a depreciating asset. However, in a cloud model it could also be considered an opex that is not owned by the company and is merely used and discarded. Census demographic information is a good opex example. Client data is capex, for example, and geocoding your client data is a huge investment as well as a unique competitive advantage. Satellite data might be both capex and opex. Change detection information from multiple satellite data passes is an interesting case that could be argued either way.
But are we just making a case for cloud computing with dollars and cents? Will cloud computing instill the same loyalty in a solution provider? Are we more likely to choose an “a la carte” selection of software menu choices given certain inherent support for interoperability standards by SaaS providers? Will configuring a solution “in the cloud” offer the same technological result? That is, do the SaaS solutions today provide comparable analytical functionality to those of desktop or server solutions? These are early cloud computing days. But because of the hype cloud computing has generated around the potential for cost savings, discussions concerning the total cost of ownership enter the picture earlier in the process. Users might immediately question the virtues of cloud versus server, versus desktop, versus hybrid cloud, etc. with “cost” being a big driver this time around. Forget the IT department; everyone is more in tune these days to the options and the impact to cost structures. Cloud computing offers the opportunity for lower costs but also options to have a flexible computing architecture, extensibility, and on-demand infrastructure that would have created huge capex headaches in the past. The value play for cloud computing seems to favor opex efficiencies.