Just to point this out, traffic always drove shaping of journalism. That's exactly how beats are established and budgeted. For example, typical metro desk stories are always popular but not always overly staffed because the news pretty much writes itself. Education takes more work, but it's not that important to audience so you pick a person or two wisely to run it, maybe throw a general assignment reporter on a few stories. Investigative pieces, for those you throw your best people and if it's newsworthy - that is, in the matter of public interest - you wastefully throw resources at it to get the story out. You can't do that for every story though, so editorial meetings are harsh.

The only difference is that for traditional print, the metrics lagged behind by months to a year (these were done by circulation analysts or media analysis companies that do this for a living, with lots of phone calls and even more interpolation). With interpolation, you can imagine how you can under- or overestimate a situation. With print interpolation, it's even worse: they would ask people if they didn't read a newspaper this week, did they at least glance at one that someone held, or left on a bench? If they said yes, they got a checkmark for a partial reach. Based on these delayed and inaccurate numbers, people sometimes got or lost their jobs. More often, they got reassigned to do something else. It's always been like that.

Now, with near-realtime feedback, you better know how a piece or a series or a section front is doing and can react to it in a more timely manner and make adjustments (ex: throw more people on the project, or pull people away from it and have them try something else). For example, ebola as a news topic is not doing so well now because it's pretty much contained, so lets get that reporter onto something else.

Don't assume that this modernization change is going to be a rowship with a whip and a piper. Besos has a decent track record on enabling journalists to do their job better, seen both from the outside and the inside. And yeah, metrics can be misread, misused and misinterpreted, but that's always been a possibility.