Too Much TV: Your TV Talking Points For Thursday, May 19th, 2022
What do Santa Claus and Netflix's algorithm have in common ?
Here's everything you need to know about the world of television for Thursday, May 19th, 2022.
A MUST-READ FOR ANYONE WHO HAS EVER USED THE PHRASE 'NETFLIX ALGORITHIM'
Back in the mid-1990s, at the height of NBC's "Must See TV" success, there was an unfortunately timed piece in the Wall Street Journal that discussed the network's success with comedies. It quoted unidentified network executives who touted their ability to crank out hits, in part because they had "figured out" what worked with audiences. They had this list of of items each sitcom must include, ranging from whether or not any house needed stairs (yes) to which side the front door should be located (the audience's left). It was all incredibly misguided and predictably within a couple of years NBC was in a primetime ratings freefall.
Predictions about how a TV show or movie will fair with the audience are always doomed to failure. No matter how in touch you feel with the audience, no matter how you lean on test panels, there are too many impossible to quantify variables in play. Gut instincts or data can't account for bad acting or a lack of chemistry on screen. Or just bad timing. My favorite example of that is the unfortunate release of the movie Space Camp just a few months after the space shuttle Challenger disaster.
That's why I've always been skeptical of the idea that Netflix was somehow using data to make determinations about which shows or movies to order. The phrase "Netflix algorithm" gets thrown around a lot in Hollywood, although you'd be hard-pressed to find anyone who can accurately explain what that means. It's just some mysterious thing everyone is sure really exists, like the Loch Ness Monster or low-calorie cookies thar don't go instantly to your hips.
The Entertainment Strategy Guy has taken a look at this mysterious phenomenon, and let's just say he doesn't quite believe in it:
On top of that, the logarithmically distributed returns—the idea that most shows are flops and a few are huge hits—makes modeling success even trickier. Again, the uncertainty is huge.
This is why, at best, the “algorithm” mostly predicts which genres and types of movies will be successful, but most of these “predictions” seem fairly obvious. You don’t need Netflix’s data to glean these “insights”.
People like silly romcoms and bad holiday films? Uh, yeah, just look at the Hallmark channel’s ratings.
"Sad-coms" are out? I wrote about that two weeks ago—but I’d had the insight months before that—before someone leaked a memo to Business Insider saying the same thing.
People like comic book films? No kidding. You needed an algorithm to tell you that?
I encourage you to go read the entire piece. But the overarching point he makes is that efforts to use predictive data to select which projects should be made are almost certainly doomed to failure.
On the other hand, using metrics to measure quantifiable things such as time spent watching, completion rates and other actions can provide some real insight. Especially when you include the cost of the show and other factors. I talked about this a bit a few months ago in the aftermath of the cancellation of Netflix's The Babysitter Club:
I've written about this a bit in the past and while I've agreed not to discuss some of the specifics (unless I learn of them independently), a Netflix employee walked me through some of this last year.
A primary Netflix metric is called the "adjusted view share," which is a combination of more than 30 factors that attempt to assign an overall "value" for any piece of content. An example I was given was that it's possible to track which content was most watched by brand new subscribers last month. That content would be considered more valuable because it presumably was one of the reasons why viewers subscribed. But if those viewers exit after a month or two, that lessens the value of the content. The assumption being that some percentage of the canceled subscriptions came from people who subscribed primarily for a specific show.
It depends on where people are watching. A show that is more popular in a region such as the U.S., where the ARPU (average revenue per user) is higher has a greater value than one that tracks more in regions where the ARPU is lower. Although that indicator is weighted less than some others and whether the content is attracting subscribers in a territory where subscriber retention costs are high also factors into the equation. Netflix also tracks how many people complete a TV show within a week, the percentage of people who rewatch a series (although if the number is too high, it's discounted as possible fan manipulation). And there are many more. Each of the factors is weighted differently and the weighting can apparently change as the company's strategy evolves.
The number of Netflix employees who can access the granular data appears to be relatively small. Which is the ultimate reason for data such as the AVS. It's a way to provide guidance internally without getting anyone too much in the weeds. The general data is used for everything from decisions on additional seasons to PR efforts.
But this data is only a small part of the data harvesting that is used by Netflix. The company does extensive tracking of subscriber behavior, ranging from aggressive A/B testing of thumbnails to UX heatmaps that show how subscribers are sifting through the app looking for something to watch. And it's more granular than you might expect. It's not just determining which thumbnail is mostly likely to convince someone to click into a TV show or movie. It's determining which thumbnail gets people to click in who will then watch the program more than a minute or two. It's a fine line between teasing and trickery, with one being a lot more subscriber-friendly than the other.
My point about all of this is that a lot is made of Netflix depending on "algorithms," but it's more nuanced than that. The real data crunching comes after a project is ordered. As it was explained to me, data has limited use in determining which specific project should be ordered. There are simply too any random factors that can't be properly quantified. But it's very helpful in guiding general decisions, such as the type of content people are looking for, actors who have a built-in audience, etc. And it's extremely helpful in the post-launch period, when it is easier to track subscriber behavior and use that behavior to estimate the ultimate value of a show to Netflix.
In the end, there is no magic technological fix that can make programming decisions less risky. I've heard executives refer to the process as "one part gut instinct and three parts prayer" and I suspect that even with all of the data points available in 2022, none of them have any significant impact.
THE CW ANNOUNCES ITS 2022-2023 PRIMETIME SCHEDULE
Upfront week ended today with the CW and despite the uncertainty over its future, the network is fully embracing a seven-day-a-week schedule.
Here is this fall's primetime lineup, with the new shows in bold:
Penn & Teller: Fool Us
Whose Line Is It Anyway?
Magic With The Stars
World's Funniest Animals
The CW also has five shows on tap for the midseason: Gotham Knights, season nine of The Flash, season four of Nancy Drew, season three of Superman & Lois, Recipe For Disaster and season nine of Masters Of Illusion.
TWEET OF THE DAY
ODDS AND SODS
* HBO has renewed Barry for a fourth season.
* Second season of Bridge and Tunnel will premiere on July 10th on Epix.
* The CW has announced that two new specials will be airing on Sunday, June 19th: The Black Pack Excellence and Brandon Leake: A Family Affair.
* Season three of the Apple TV+ comedy Trying will premiere on Friday, July 22nd.
WHAT'S NEW FOR THURSDAY
Here's a quick rundown of all the new stuff premiering today on TV and streaming:
A Perfect Pairing (Netflix)
Bang Bang Baby (Amazon)
Ghost Adventures: House Calls Series Premiere (Discovery+)
Insiders Season Two Premiere (Netflix)
Interrogation Raw Series Premiere (A&E)
Kingdom Business Series Premiere (BET+)
Legendary Season Premiere (HBO Max)
Rodrigo Sant’Anna: I've Arrived (Netflix)
Swamp People Season Finale (History)
The Boss Baby: Back In The Crib (Netflix)
The G Word With Adam Conover Series Premiere (Netflix)
The Hall: Honoring The Greats Of Stand-Up (Netflix)
The Ipcress Files Series Premiere (AMC)
The Photographer: Murder In Pinamar (Netflix)
Undiscovered: Edgewood (Discovery+)
Click Here to see the list of all of the upcoming premiere dates for the next few months.
SEE YOU FRIDAY!
If you have any feedback, send it along to Rick@AllYourScreens.com and follow me on Twitter @aysrick.