Sprint Any Mobile, Anytime: Serious Data Mining

With the release of Sprint's new "Any Mobile, Anytime," a new feature of Sprint "Everything Data" plans that allows calling to any U.S. wireless number, on any carrier's network, at any time, without additional charge, as well as the coming AT&T "A-List," which allows unlimited calling to any five to 10 numbers, it is getting even harder to figure out what a "minute of voice use" actually costs.

T-Mobile USA has had "myFaves" for some time, allowing unlimited calling to any five phone numbers, while Verizon has had a "Friends and Family" calling circle feature for years, as well. Both of those plans, like the new AT&T plan, allow the service providers to tout an enhanced feature that makes calling to some phone numbers free.

The Sprint plan moves beyond existing "friends and family" or "calling circle" plans that typically include only mobiles on a single carrier network, or only a limited number of phone numbers on any network. The issue is that the new Sprint plan, like existing plans available from AT&T and Verizon, which make all calls to other AT&T or Verizon mobile customers free calls, also introduces an element of uncertainty into user thinking about how many "out of network" minutes actually are needed on a monthly basis. That's the new element of uncertainty on the consumer end of the service.

Beyond that, large calling circle plans also have introduced a bit of uncertainty on the service provider end, at least as regards the profit margin on voice calling to "off network" numbers. One can assume that the marginal cost of terminating a domestic "on network" mobile call placed from a domestic "on network" number is quite small.

One also can assume that domestic off network calls within a purchased bucket of minutes carry a pre-determinded margin. So the really profitable scenario is when a user exceeds a bucket ("breakage") and has to pay a really-high incremental price per minute, or when users buy buckets of minutes they do not use.

Sprint's new "Any Mobile, Anytime" program is risky to a certain extent because Sprint might not fully understand how its users will behave once they have access to the new feature. Presumably quite a lot of data mining already has gone on, allowing Sprint planners to predict past termination behavior, especially the number of placed calls from Sprint mobiles that terminate off network to landlines and other mobiles.

It was just that sort of rudimentary data mining that lead all the carriers to craft programs that allow unlimited calling nights and weekends, for example, when usage is light enough that additional calling does not tax the networks.

But Sprint's data mining on behavior of its "unlimited" plan customers might have provided enough new insight into changes in user behavior that the company has confidence about the changes in user behavior that occur when users migrate from "large buckets" to "unlimited" calling plans.

Sprint might have uncovered in its data mining that the existing calling patterns of its current customers show a marked upsurge in mobile-to-mobile terminations, both on network and off network.

Sprint might also have discovered that, on balance, the cost of offering unlimited termination on rival networks can be captured within the new fee structure it has created.

In other words, unlimited mobile termination at current plan costs is at least revenue neutral when customers move to $70 a month access plans featuring 450 minutes terminated on domestic wired networks, or $90 a month when users terminate no more than 900 voice minutes on domestic wired networks.

But Sprint might also have concluded that getting more customers on data plans results in the new "Any Mobile, Any Time" plans being revenue incremental, not simply revenue neutral, after accounting for cost of service.

It's getting harder to tell what any given calling pattern actually represents for any user, at least in "cents per minute" terms. And that ultimately might be a very good thing for service providers and users. For users, the actual "cost" of calling will become less important an issue in choosing service plans.

For service providers, the actual profit margin on voice services might remain interesting, even as formal "price per minute" metrics fall.