Concurrent players on Steam: what CCU measures and what it tells you

Concurrent players, or CCU, is how many people are in a game at one moment. Steam's Web API returns the current count and leaves out anyone playing offline. Peak CCU is the highest reading in a day. It reflects engagement, updates and sales, but it is not a sales count, and a demo has its own App ID.

Last reviewed · SteamStats team

Definitions

  • Concurrent players (CCU): the number of people in the game, connected to Steam, at the moment of reading.
  • Peak: the highest reading in a window, such as a calendar day or the last full day. A peak is only as precise as the interval between readings.
  • Average: the mean of all readings in the window. It is steadier than the peak and better for comparing weeks.
  • Daily players: unique players over the last day. It counts each person once, however long they played, so it is always higher than the peak.
  • All-time peak: the highest reading ever recorded. Its value depends on when a tracker started reading the game.
  • Reading interval: the time between two readings of the count. It decides how close a recorded peak gets to the real one, and it is the first thing to check when two trackers disagree.

What Steam says

Gets the total number of players currently active in the specified app on Steam.
Steamworks Web API: ISteamUserStats
  • The public Web API method takes only the App ID and does not count players who play while not connected to Steam.
  • Steam's Most Played chart ranks the top 100 games by players in the game right now, with each game's recent peak, and has a view ranked by daily players, the unique players over the past day.
  • A demo is a separate App ID linked to the full game, so its players are counted separately. Demo players can write reviews, but a demo has no Community Hub of its own.
  • Valve's visibility documentation lists players and purchases among the signals that matter: Steam looks at whether people are buying and playing the game.

Worked example: Pebblewick's first week

Illustrative inputs for Pebblewick, the fictional demo game by Hollow Fern Games. On launch evening, readings once an hour show 1,180 players at 20:00 and 1,205 at 21:00. Readings every 15 min catch the real peak of 1,240 at 20:15.

Pebblewick: what the numbers say
QuestionCalculationAnswer
How much does an hourly reading miss?(1,240 − 1,205) ÷ 1,2402.82% below the real peak
How much of launch day is left on day 7?668 ÷ 1,24053.87%
How does the peak compare with daily players?1,240 ÷ 4,96025.00%

The day-7 figure is the one most worth watching: it shows how many launch players returned once the novelty passed. The ratio of peak to daily players shows how spread out the play is; a low ratio means people play at different hours, as happens when buyers live in many time zones.

Concurrent players do not convert into sales on any fixed ratio. A free weekend, a streamer or a sale can lift CCU without the same lift in revenue, and a long game keeps more players online per copy sold than a short one.

Reading CCU day to day

  • Compare like with like. Saturday evening and Tuesday morning are different days for any game, so a reading means most next to the same weekday and hour in earlier weeks, not next to yesterday.
  • Use the peak for news, the average for trends. A patch, a sale or a streamer shows up as a peak within hours; whether the game is holding its audience shows in the weekly average.
  • Watch the first evenings closely. Steam refunds a purchase within two weeks with under two hours played, so players who drop off in their first sessions are still inside the refund window.
  • Allow for a delay after sales. A discount brings new owners, and they may start playing over the following days, so the CCU effect of a sale can outlast its first-day spike.
  • Know what it cannot show. The public count is one number for the whole world: it does not say how many people own the game, which countries they play from or how long they stay. Those come from Steamworks data, not from CCU.
  • Read demos separately. A demo's players are counted under the demo's App ID and say how many people are trying the game, not how many own it.

Why CCU numbers differ between trackers

  • Reading interval. A tracker that reads once an hour misses short peaks; one that reads more often reports a higher peak for the same day, as the example shows.
  • Where the day ends. A daily peak depends on the time zone that defines the day. The evening peak of one country can fall into the next calendar day in another time zone.
  • Start of tracking. All-time peaks are the highest value a tracker has seen, so a tracker that began after launch reports a lower one.
  • Separate apps. Demo players count under the demo's App ID, not the full game's.
  • Offline play. Players not connected to Steam are not counted anywhere.

Where to see this in your own data

The Players view records concurrent players every 15 min on Indie and every 5 min on Studio and Publisher and compares each reading with the usual range for the same weekday and hour. It shows today's curve, the peak of the last day and the all-time peak, the busiest hours of the week and daily peaks over 7, 14 or 30 days. With Steamworks data connected, it adds daily active players and typical playtime. Players are read from launch day on.

The Players feature page linked below shows the view with the demo game.

Sources