Facebook recently conducted an experiment to see if Indonesia's local mobile Internet service could be tweaked to make it easier for users to quickly access the company's app.

It could be. The company was able to shave seconds from the app's load time.

There's technological and historical context here, namely that mobile networks were built to support phone calls and, to a lesser extent, large video downloads. Apps like Facebook's impose a different sort of traffic on a network. Think of it as a street built to speed the way for pedestrians rather than cars. In Indonesia, Facebook is eager to have an Internet "road" across which Facebook can fly.

This comes as Facebook chief executive Mark Zuckerberg made his first visit to Indonesia to promote the company's project to spread Internet use across the world, Internet.org. Indonesia already represents Facebook's fourth largest market, and has a data-hungry young population that is placing enormous strain on its Internet networks.

For its app speed experiment, Facebook teamed up with the phone company Ericsson and the Indonesian telecommunications provider XL Axiata for a round of testing. Dummy Facebook users were set up in urban, suburban, and rural Indonesia -- Jakarta, Bintaro, and Tigaraksa -- and their demands on the mobile network were tracked. Then the partners adjusted the XL Axiata network to see if they might be able to provide a zippier Facebook experience.

The result? After jiggling some knobs and pulling levers, Facebook noted in a white paper released Monday, that it was able to hit "app coverage improvements of 40 to 70 percent," including by learning how to automatically redirect users in Indonesia's most remote regions to servers that were both closer and more robustly connected to XL Axiata's network infrastructure.

Facebook, though, isn't simply after improving the Facebook experience on one Indonesian mobile network. Its more fundamental aim is redefining the key performance indicators, or KPIs, that telecom providers use to assess the strength of their networks.

To that end, the white paper says, the Internet.org team has used the Indonesia experiment to create "a replicable model for monitoring, analyzing, measuring and improving app coverage that can be applied to any mobile network, which will help operators cost-effectively target network improvements for the most impact on user satisfaction, loyalty, and retention."

That ambition makes some intuitive sense, Facebook's own self interest aside. Much of the developing world's access to the Internet comes in the form of mobile phones -- Indonesia, the paper notes, is experiencing a glut of cheap cellphones -- and in many places, Facebook is the main way that users get online. That's especially true in places where Facebook has worked out so-called zero-rated deals with providers in which visiting its site or using its app doesn't count against customers' monthly data allotments.

But it does mean that the Internet in much of the developing world is increasingly being tuned to Facebook.