Several months ago, Bob Lewis (, @ITCatalysts) recommended reading Thinking, Fast and Slow by Daniel Kahneman. This is a fascinating book about how people make decisions, and it introduced me to the psychological concept of “anchoring,” defined as “the effect of a prior judgment of an object, the anchor, on our future judgments regarding another object.” (The Anchoring Effect: How Our Prior Knowledge Affects Our Perception). Furthermore: "Cognitive psychologists tell us that human judgment is essentially relative or comparison based, even if we are not asked to make a comparison explicitly. So, in evaluating the present object or person, our minds search for an anchor. A particular anchor may be selected because it is readily accessible, because it is suggested to us, or ‘self-generated via an insufficient adjustment process.’” What does this have to do with how IT assess the benefits of cloud computing? It’s no secret that the majority of IT shops have concluded the primary benefit of cloud computing derives from one of various Infrastructure-as-a-Service (IaaS) offerings. As a member of the fledgling Windows Azure sales team in the fall of 2009, this was perplexing to me. It seemed intuitively obvious that eliminating the operating and administrative requirements of supporting application servers in virtual machines, as provided by the platform as a service (PaaS) approach, was the way to go. But I’m a developer by background. I haven’t actually worked inside a data center since college, where the largest part of the floor was occupied by an IBM 370/148 mainframe. It’s been a while. For the past three years, almost every conversation with IT about Windows Azure would quickly devolve into what Azure is not. It wasn’t uncommon to hear statements such as, “if Azure can’t do X, there is no point in continuing this conversation.” And the reality was, even if Azure could do X, it did it differently. There is tremendous value in the PaaS approach, but the IT conversation could never get there. Fundamentally, this is explained by anchoring. Taking the definition above (anchoring is the effect of a prior judgment of an object, the anchor, on our future judgments regarding another object), we have the anchor as the way things are done today, while the “other object” is the PaaS or IaaS offering. Given that we make comparisons even when uninvited to do so, it’s clear why PaaS does not fare well when evaluated by experienced IT pros: it’s simply very different from a traditional data center. In fact, it eliminates almost every artifact of a traditional data center. Conversely, IaaS will fare well in the comparison, because it really doesn’t look all that different than a traditional virtualized data center. It’s easy to overlook how IaaS is supposed to be used in a service-oriented context—and as a result, cloud services end up utilized in the same manner as always. Anchoring yields decisions with short-term benefits, but which lack any significant long-term economic benefit; the latter remains locked up in the as-yet-unexplored PaaS. Only when a new object is evaluated without the bias of old objects can the real benefits of cloud computing be unlocked. Mark Eisenberg is the Director of Business Development for Fino Consulting, a technology consulting firm, which provides cloud and mobile application development for the Enterprise.  Image: belekekin/