Rather than building the service. It would appear they have every intention to continue providing functionality related to using/playing existing GFWL games, but not add new functionality or maintain features that don't directly relate to a gamer's ability to use a game.

So, with Steam for instance, they can offload all of the costs and services related to digital distribution. Thus, they can shutdown those services from GFWL itself. Which should also translate to less engineering (personnel) support needed from MS's end. All of their GFWL games which they put on Steam could then be directed to product managers that only need to be familiar with the backend Steam provides for their (non-gamer) customers.

From a PR/customer standpoint, it's probably also about maintaining the image that as a publisher/studio, they care about the long-term status/availability of their products. Especially with the recent bullshit that happened with GameSpy. I hope Glu dies in a fire.