Despite the notion that cloud-based computing is important for larger enterprises because it offers economies of scale, it may be other economies that allow telcos to create a business out of cloud computing. And the reason is not, as you might think, economies of scale, says Joe Weinman, AT&T (News - Alert) VP for Solutions Sales, in a post on gigaom.com.

 

In fact, large telcos might benefit more from what Weinman calls "statistics of scale."

 

It is true that large service providers can achieve economies of scale through volume discounts, experience curve effects, and consolidate operations at a scale that minimizes unit costs, says Weinman. But so can enterprises and cloud providers, who can build data centers out of the same components and similar tooling, purchased at similar price points.

 

Weinman thinks it is the ability to aggregate uncorrelated demands from multiple customers that will allow communications service providers a way to operate in ways that smooth out demand over time, thus allowing more efficient use of facilities.

 

So it is conceivable that large enterprises and cloud computing hosting services will have volume buying, experience curve benefits, marketing channel benefits and network effects that are quite similar to what telcos might offer.

 

Weinman suggests that telcos will have an edge in "statistics of scale," essentially the ability to smooth out demand peaks created by customers in different verticals and time zones.

 

AT&T has advantages. The intriguing thing is figuring out where they fit in a traditional taxonomy of business advantage.

 

As an exercise, consider the other major type of advantage economists study: economies of scope, essentially the concept of wider product range. The notion of "product bundling" is a scope economy - selling more things to a single customer - than scale economics.

 

Can a large telco offer scope economies to customers in ways that independent cloud computing providers cannot? Can a telco exploit its brand equity in ways that other competitors cannot? Can it provide a wider range of services than competitors?

 

Actually, a telco's potential advantages in cloud computing are interesting in an economist's sense, as the smoothing of demand over time and between verticals might be considered derivative from "scale" or a derivative from "scope economics."

 

One might say the ability to smooth out demand across time zones or verticals is a result of scale economies. One also might argue that is a derivative of scope: extending connectivity services by adding hosting and computing services, to create a new "bundle" or range of products sold to any single customer.

 

To the extent that the ability to average peak load is an advantage for telcos, is that a derivative of scale or scope? To the extent "smoothing" is akin to statistical multiplexing, where would you attribute the source of advantage? One might argue multiplexing in itself confers no economies of scale at any end user site (though it offers clear efficiencies in terms of bandwidth). But might multiplexing offer scale economies when aggregated at the level of a tier one carrier? That would seem to be the case.

 

So, to the extent that telcos can leverage massive customer scale as well as operation across time zones that represent another sort of traffic smoothing, perhaps the notion of economies of scale still holds, though an argument for scope economies also exists. Multiplexing customers is more logically a scale economy than a scope economy.

 

Weinman disputes the notion that all enterprise computing will "move to the cloud." Some apps will, he argues, because it is more efficient or effective. Others won't, for similar reasons. Peak demand curves and significant differences in regional power cost will play a role in determining when it makes sense to cloud source compared to local sourcing.

 

Likewise, cloud computing can substitute operating cost for capital cost. But the enterprise advantage is quite situational. For a large enterprise, it nearly always makes more sense to invest in capital (the business phone system) rather than buying a managed VoIP service, for example.

comments powered by Disqus