“We love games,” is how Netflix told the world big media wants you to watch less and play more. Games are good media. They’re cheap-ish to build, easy to update, and keep people engaged when they’re not watching, reading, or listening.
— — —
Two years in and Netflix games have been downloaded 82 million times. According to Netflix co-CEO, Greg Peters, the costs are, “very, very, very low…” He also said the impact is “very, very small.” Seems like their games might be “very” profitable. They’re not. They’re about keeping Netflix customers happy and busy when they’re not tuned in.
This isn’t Madden or Fortnite. Madden maker, EA, made $400 million — last quarter. Fortnite rakes in epic money for Epic. This is SpongeBob flipping Krusty™ burgers.
I’ve made my own AI… Artificial Charles’s actual intelligence. It’s trained on my unCharles posts and spews the kind gibberish I write on demand and at scale. You can call it, acAI for short.
acAI, tell me an epic story about a game that’s not epic.
One day, Josh Wardle found an old word puzzle game. When you played it at your aunt’s apartment in Brussels in the 1970s it was game of the year and called Mastermind. Josh put it online and gave it the name, Wordle. It took very little time, effort, or money to build. Ninety people played it on day one. Two months in, 300,000 people were playing his daily word puzzle. It cost Josh $100 per month to keep the online lights on. Two months later, the New York Times paid him about $1M for it. Two years in, two million people play Wordle every day. Wordle makes no money for The New York Times.
If you focus on the user growth and low costs and ignore, “Makes no money…”, it’s the kind of story that gets media people excited. “Very” excited. So, when Puzzmo launched on October 23 to reimagine classic newspaper games, Hearst just had to have it. Hearst bought Puzzmo this week. It’s been around for six weeks. Six weeks. You can’t make this up.
Clearly, Hearst couldn’t find someone to build a crossword or sudoku and add a leaderboard. They plan to license their new asset to newspaper folks everywhere who also can’t build their own. Sad really.
Let’s pivot to what’s not sad. Akinator. The Venn Diagram of people who read unCharles and know Akinator has one person in it. More people should know it.
acAI, talk to me about Akinator.
Do you remember Zoltar from the movie Big? Good. If not, it’s the animatronic fortune teller you’d find in an arcade in the 1970s. Like that time you went to Niagara Falls when you were a kid. Akinator is like that. You think of a person or character, living or dead. It can’t be your uncle Sam. It has to be someone reasonably well-known. Taylor Swift, Motahari, James Bond, or SpongeBob. Ak asks you a series of Yes/No questions until it’s ready to guess your character. Unlike Wordle, Akinator makes money. Every time it asks you a question, it shows you an ad. Based on downloads, ad rates, and the like, I’d estimate it makes $15m to $20m a year. Unlike those puzzle games, Ak is based on proto-AI from 2007. Which is to say it has a database of 100,000 characters with parameters. Everything makes money when you attached AI.
Hallucinations aside, AI doesn’t automatically add money. But, Akinator has been downloaded 100 million times for Android devices. And a too-hard-to-find but still lareg number of downloads on Apple devices. About 9000 people play it every hour. Each user sees about 20 ads per game. Each ad is worth about 1¢. The server costs run into the low thousands of dollars per month. Which means Akinator is wildly profitable.
Right now, 1% of Netflix users play games. That number will go up. Right now, it’s enough for Netflix that these games keep their customers amused. One day, these games will be profitable. Netflix is starting to figure out ads in their streaming business. That’s why it makes sense when Greg Peters said, “Mobile games business a growth engine with a ‘material impact’ on its $30 billion business.” He’s not just talking about funin games. He’s talking about real money.