Earlier this week, data from Axiom Capital Management showed that the meteoric popularity of the Pokémon Go mobile game plateaued in mid-July—only two weeks after its earth-shattering debut—and has declined steadily since then. Daily active users and user engagement are both down by roughly 30% from their peak.
A deep analysis by Ars Technica found that the game's initial retention was worse than that of comparable mobile games. Analysts went so far as to tell Bloomberg that downward trends cast doubt on the long-term viability of the entire augmented reality category.
Get Data Sheet, Fortune’s technology newsletter.
But that’s jumping the gun. Huge initial interest in Pokémon Go was clearly motivated in part by a fleeting nostalgia among twentysomethings, but also by a fascination with the novelty of playing a digital game by moving around in the real world. If Pokémon Go has failed to capitalize on that fascination, it’s not because augmented reality isn’t viable. It’s because, based on what I've experienced in several weeks with the game, Niantic and Nintendo’s joint foray is fundamentally half-baked.
The flaws in Pokémon Go are multiple, and range from small details to all-encompassing philosophical shortcomings (and I'm just talking about design, not the game's various technical problems). For new players, there's no introduction to basic principles and gameplay. Some have argued that this has actually encouraged players to work together, and anyone who has seen crews of college students or young families gallivanting around with their phones out can attest that the sense of group effort was key to the game’s initial appeal.
But to what end? The big, glaring problem highlighted by Pokémon Go's declining engagement is that, while the actual hunting of pocket monsters is fun, there’s very, very little to do with them. The gameplay proper centers on capturing gyms scattered around the world, but to even try their hand at that, players have to have leveled their monsters quite a bit—most gyms I’ve checked out are held by monsters of over 1,000 combat power, which I haven’t managed to build up over more than a month of casual but fairly consistent play.
And leveling can only be accomplished by collecting monsters, since there are no quests or other content. Players can’t even directly fight each other, which seems like a massive missed opportunity.
Even if you do find a gym you can contest, the game’s combat mechanics are as simultaneously basic and opaque as all of its other elements. You try to dodge attacks by sliding left and right, and I think you attack by tapping on your opponent. But my few tries were over so quickly, and the controls so unclear and unresponsive, that I'm not entirely sure.
For more on Pokémon Go watch our video:
In short, Pokémon Go feels like a game made as a homework assignment by a Niantic team who would rather have been doing something cooler. Taking a look at Niantic's Ingress, the technical precursor to Pokémon Go, only drives this sense home—with its black and neon colors and cyber-punk storyline, it’s like the raw and energetic basement tape to Pokémon Go’s washed-out compromise of a major label debut album.
But the shortcomings of one game don't mean that augmented reality itself is a flop—if anything, Pokémon Go has whetted audience hunger for a more developed take on gaming that integrates real-world movement and social interaction. Let me take this opportunity to personally lobby developers to make an AR dungeon crawler or detective game.
Just be sure to make it an actual game.