A large number of games designed for years of operation have closed within months. The failures follow a recognisable pattern.
The model
A game intended to be played continuously, updated with content, monetised over time rather than at purchase.
Which requires a team maintaining and extending it indefinitely, infrastructure to run it, and a population large enough to sustain matchmaking and social activity.
The economics are attractive when it works — revenue per player over years far exceeds a single purchase.
The attention constraint
The fundamental problem.
These games ask for sustained time rather than a single purchase, and players have a limited number of such commitments.
Which means the market is not for purchases but for slots in players' routines, and those slots are already occupied by established titles.
Displacing an incumbent requires being substantially better rather than comparably good, because switching costs include abandoning accumulated progress.
The content treadmill
Players consume content faster than teams produce it, consistently and by a wide margin.
Which means a launch that goes well creates immediate pressure for the next update, and gaps in the schedule cause population decline that is difficult to reverse.
The content pipeline must therefore be running before launch, which is a substantial investment made before knowing whether the game will succeed.
The population threshold
Below a certain population, matchmaking degrades, social features feel empty and the experience deteriorates.
Which produces further decline, and the process accelerates rather than stabilising.
The threshold varies with the design — a game requiring large matches needs far more players than one requiring pairs.
Designs with high population requirements are correspondingly fragile.
The launch window
These games get essentially one chance.
Players who try a game at launch and leave rarely return, even after substantial improvement, because the recommendation networks have moved on.
Several games that were transformed after poor launches eventually found audiences, which took years and is the exception rather than the pattern.
Differentiation
Entering a genre against an established leader with a similar proposition has failed repeatedly.
Which is obvious in retrospect and was not obvious during greenlighting, when the market size looked like an opportunity rather than an occupied position.
The successful entrants have generally differed structurally rather than incrementally.
The shutdown problem
When a live game closes, it becomes unplayable, including for people who paid.
Which has prompted consumer campaigns and legislative interest in several jurisdictions, arguing that sale of a product that can be unilaterally disabled requires disclosure or an end-of-life plan.
Industry responses have cited licensing, server costs and the difficulty of releasing server software.
Some studios have released offline modes or server tools at shutdown, which demonstrates it is possible where the will exists.
What has worked
Starting smaller and growing, rather than launching at scale against an incumbent.
Designs tolerant of lower populations.
And treating the game as a platform for a community rather than as a content delivery schedule, which is a different organisational proposition entirely.
The sunk cost problem
Studios continue funding underperforming live games longer than the economics justify, because shutting down writes off the investment publicly.
Which has produced extended declines rather than clean decisions, and the eventual closure arrives with a smaller population that suffers a worse outcome.
Announcing a defined support period, rather than an open-ended commitment, sets expectations that closure does not violate.
Player investment
Time and money spent create attachment that makes closure genuinely costly to players.
Which is why refund practices at shutdown attract attention, and why offering refunds for recent purchases has become a common gesture.
Transferring progress or cosmetics to another title from the same publisher has been used occasionally and requires foresight in how accounts are structured.
Seasonal structure
Most live games now operate on seasons with defined content and a battle pass.
Which creates a predictable cadence and a treadmill, since missing a season means missing content permanently in most implementations.
Making past content obtainable later reduces urgency and increases accessibility, and studios weigh those against each other differently.
Testing at scale
Problems that only appear with large populations cannot be found before launch.
Which is why beta periods and staged rollouts exist, and why they are frequently too short to reveal the economy and progression problems that appear over weeks.
Games that ran extended technical tests have generally launched more smoothly, at the cost of exposing the game earlier.
Infrastructure cost
Servers scale with population, which means a successful launch produces immediate large costs.
Under-provisioning produces queues and failures at exactly the moment first impressions form.
Over-provisioning wastes money on capacity that will not be needed a month later, which is why elastic capacity has become standard.
The pivot
Several games have relaunched with substantially changed models — going free to play, changing genre elements, or rebuilding progression.
Which has worked occasionally and requires a marketing effort comparable to a new launch, since the audience must be told the game has changed.