Some of this week’s most interesting news came out of Amazon, by way of the estimable Reuters reporter Jeffrey Dastin, who surfaced previously secret documents on the audience sizes of Amazon Prime shows, and the very different way the e-commerce giant measures success with its multi-billion-dollar original programming investment. Most importantly, the story shows how beside the point it is that Nielsen, that metrics behemoth undergirding TV’s Old Guard, would measure audiences in the New Order, on OTT channels such as Netflix and Amazon Prime that don't run ads. Yes, it’s interesting that Nielsen believes “Bright,” with Will Smith, was seen by 11 million people when it debuted over the holidays, while “Cloverfield Paradox,” launched with a single $5 million Super Bowl ad, attracted only 5 million in its first three days. But really, no one’s buying ads based on those data points. And yes, sooner or later Amazon may put ads into its original programming, though for now, it seems happy to do nudge-nudge-wink-wink product placements for the Echo and its Alexa digital assistant in shows such as "The Tick" ("Alexa, shuffle ominous music playlist").The original rationale for Nielsen to track viewership on Amazon and Netflix was putatively to help its main clients, the networks and studios, better know how their licensed shows were doing there. Even that had some big holes, like Nielsen's inability to track mobile viewership, by all accounts a huge share of Netflix business and presumably Amazon's also.  But in the three-plus years since Nielsen jumped into that effort, the entire industry has changed radically.  Netflix said it will spend  up to $8 billion this year on original programming, precisely because the networks and studios aren't licensing their own shows as much. Amazon plans to spend around $5 billion. Meanwhile, the networks are busily building their own streaming channels, which almost certainly will have minimal advertising, to compete with the Netflix experience. And make no mistake: everyone's coming to the OTT party. This week, The Diffusion Group issued a report predicting every "major" TV network will have its own streaming channel by 2022. The non-"major" channels presumably are zombies just waiting to be killed off the skinny bundles of the future. Regardless, those networks  almost certainly won't be licensing material to competitors, as Disney's sunsetting Netflix deal suggests (and, as The Diffusion Group pointed out, also may cause channel conflicts with traditional distribution partners). And given that all those channels will know exactly who's watching what show for how long on their direct-to-consumer offerings, I’m still wondering how Nielsen monetizes this OTT-tracking investment in the long-term. Who's buying the data? But back to Amazon. Some 26 million U.S. customers have watched original Prime shows, a stat the company itself has never released. That's a nice, big number, though considerably less than half even the estimated number of U.S. Prime subscribers (Amazon has never released that number either). That in turns suggests there's room for growth if Amazon can just give its Prime subscribers good reason to tune in, which brings us to the next interesting tidbit from the Reuters report.Like Netflix, Amazon doesn’t care much for traditional ratings. They just don't track any useful part of Amazon's business. What Amazon does care about, however, is how well its shows get people to sign up for Prime subscriptions, and at what cost per new subscriber.  The company's important metrics include which show is the “first stream” for a new Prime subscriber, i.e., the first one they watch after signing up for Prime, presumably because that was what helped motivate the customer to join. Amazon then divides the "first stream" number  by the show’s production and marketing costs to get a customer-acquisition cost.For instance, “The Man in the High Castle,” Amazon’s award-winning alternative-history series based on a celebrated Philip K. Dick novel, cost a pricey $72 million to make and market in its first season. In turn, however, the show was credited with attracting 1.15 million new subscribers, at an average cost of $63 per sub. For comparison, though I’m not completely sure it’s a good one, Prime costs $99 a year, and provides Amazon a sticky way to sell even more goods than usual with free two-day shipping. (Prime has since larded on an remarkable array of other free services, such as Prime Now two-hour delivery, discounts, online books, magazines, music, game content and audiobooks.)Amazon's three best-performing shows (of 19 in the Reuters data) were “High Castle;” the auto/travel show “The Grand Tour,” built around former “Top Gear” stars; and “Bosch,”  based on Michael Connelly’s best-selling series of detective novels. “Top Gear,” the most efficient subscription converter, cost $49 per subscriber.All three of those shows are connected to existing, high-profile intellectual property, and brought with them built-in audiences. Their relative cost-efficiency in converting fans into Prime subscribers suggests why Jeff Bezos visited Amazon Studios last summer and declared he was tired of small shows, and wanted the next “Game of Thrones.” The result: $250 million for rights to make a prequel series set in “The Lord of the Rings” universe (which may cost another $250 million to actually produce). Soon after came another pricey deal, for a “Conan the Barbarian” show. Despite their costs, both could be catnip for new Prime subscribers, especially internationally, where Prime is now widely available after a global expansion in late 2016. Action fantasy shows such as "Conan" and "Rings" should travel internationally a lot better than the chamber dramas, quirky comedies and awards fodder of the Roy Price regime, which drew critical attention for shows such as “Transparent,” “Alpha House,” “I Love Dick,” “Mozart in the Jungle” and “Sneaky Pete.”It’s likely we’ll see a lot fewer of those kinds of prestige plays from Amazon going forward, and more broad-appeal beasts with international reach. In that regard, the decision to hire Jennifer Salke away from broadcaster NBC to succeed Price as head of Amazon Studios, rather than Nancy Dubuc of prestige-focused cabler A&E, makes more and more sense.