A hospitality company's remote sales support team, about seventy-five people spread across the country, most of whom have never met in person, had a budget, a date in December and a catalogue of eighteen games to choose from. The organiser, who runs the programme with three colleagues and said plainly "we're not party planners", put the question the way every organiser puts it: "How do we choose? We are not going through eighteen games. You hear this from every single person, I know, but they are highly competitive, they like sports. How do we find the best one for us?"
The answer they arrived at was not a sports game. It was a mystery set in a snowed-in hotel. The way they got there is the most useful thing we have seen an organiser do, so this is written around it.
The filter that does not work: genre
"My team is competitive, so they want a competitive game" sounds right and leads nowhere, because almost every team game is competitive. Trivia has a winner. Escape rooms have a clock. Mysteries have a solve. The genre label tells you nothing about whether your particular team will show up, stay on camera and talk about it afterwards.
What "competitive" actually tells you is something narrower and more useful: the team needs a score and a winner at the end. A collaborative activity with no outcome, a shared playlist, a guided conversation, will feel pointless to them. Keep that requirement and drop the genre.
Filter one: a theme the team will claim as its own
The hotel company picked the hotel mystery over the adventure and the winter heist for one reason the organiser stated plainly: "It would be silly for us not to choose something that's hotel themed, because then they think it's been really tailored just for them." The team sells hotel stays all day. A game set in a hotel at midnight is a game about their world, and they walked in already knowing the rules of the place. Nothing was customised. The setting did the work that a custom build would have cost ten times as much to do.
This is the strongest filter we know, and it is rarely used because lists are organised by genre, not by who you are. The question to ask is: is there a setting the team already has expertise in? A logistics team in a heist that runs on timing. A support team in a mystery built on reading people. A product team in an escape room that is one long debugging session. When the setting matches the work, the team is good at the game from the first minute, and being good at something together is the entire point.
If nothing in a catalogue matches your world, the second-best option is a theme the team already shares outside work, which is where the sports-minded part becomes useful: not as a genre, but as a shared reference everyone can draw on.
Filter two: the tone the team will tolerate
The same organiser had a second criterion, and it is one no activity list mentions. The first thing they asked about the mystery was: "Does it use the word murder?" Then: "How does the murder occur? Is it implied?" What they wanted was a colleague who has gone missing, not a colleague who has been stabbed. Their phrase for it was "quasi-clean". A body in the plot was fine. A body on the screen was not. They also joked that if the game did use the word, they would simply not mention it to HR, which tells you where the line actually sits in most companies: not at the organiser, but one level up.
Every team has a line like this and the organiser usually knows where it is. Humour that works in a startup of thirty is a complaint in a regulated firm of three thousand. A horror theme that lands with a games studio empties the room at a hospital. Ask the question directly before you look at any game: what tone would get a complaint? Then filter on that, because a complaint after the event costs more than a slightly less exciting game.
Filter three: try it before you buy it
The hotel company's organisers did something we would recommend to anyone: four of them played three candidate games themselves before deciding. Not a demo call with a salesperson. The actual game, with the actual puzzles, as a small team.
This is the only filter that produces information instead of opinion. Thirty minutes of playing tells you whether the pace suits your people, whether the puzzles are too easy or too hard for them, whether the thing runs in your browser without IT getting involved, and whether you personally would be embarrassed to put it in front of your boss. No description can tell you any of that.
The same organiser applied a smaller version of this test before the call even happened. They had found vendors by asking an AI assistant, and said the ones they dropped "didn't show video demos, it was hard to understand what they did". The one they kept had a link that opened on a phone and was easy to follow, and they said that was, for them, "an indicator of how the actual event would go". That is a fair test. A vendor who cannot make a link work on your phone will not make a link work for eighty people.
Two more things the organiser checked that are worth copying. First, login: they had struggled to get into the call itself that morning and wanted a guarantee that the game needed no account, because "if I had a problem, other people are going to have a problem". Second, scale: looking at a screen full of players, they asked "how do eighty people jump into a game and play together? It looks messy." The answer was teams of eight, each in its own room, with the whole group coming back together at the end. Ask both questions of any vendor. If the answers are not immediate, the event will not be either.
If a vendor cannot give you a playable trial, that is itself information.
What different teams tend to pick
A caveat first: the patterns below are what we see across the events we run and what organisers tell us on calls. They are patterns, not rules, and we have not yet published the event-by-event data behind them. Your team is the data that matters, which is why the three filters above come first.
One pattern that is not department-specific and that organisers raise more often than any other: the same activity repeated kills attendance. A games studio's HR lead told us their quizzes had worked for two years and then simply stopped drawing people. The fix she found was not a better quiz. It was a yearly ritual with a changing theme, so people know the shape and the content is new. If your team has done the same thing twice, do not choose a better version of it. Choose a different shape.
Sales teams want a scoreboard, a clock and a clear winner. They tolerate public competition better than any other department and get bored fastest when there is nothing at stake. Timed mysteries and heists with team scores work. Long cooperative puzzles with no ranking do not. The hotel team is a textbook case: the first thing they said after choosing the game was "we definitely want to do a leaderboard", and they planned custom trophies for the winning team because gift cards would create a tax problem. The prize does not need to be worth anything. It needs to exist.
Engineering teams want the puzzle to be fair and the logic to be sound. A clue that only works if you guess what the designer meant will be found and resented. Escape rooms with real logic and a visible system behind them land well. Anything that depends on social performance, charades, improv, lands badly. Give them something to debug together.
Support and customer success teams are good at reading people and keeping several threads going at once, which is exactly the skill set of a detective mystery with interviews and contradictory statements. They also tend to be the most dispersed across time zones, which pushes toward formats that can run in two sessions or asynchronously.
Marketing and creative teams want something with a story and are the most forgiving of weird. They are also the most likely to produce shareable moments afterwards, so if the recap matters to you, this is the team it comes from.
Finance, legal and operations are the teams organisers worry about most and the ones that most often surprise them. They like precision and dislike being made to perform. A structured heist or escape room with clear roles suits them better than anything open-ended. Keep the tone conservative and the rules explicit.
Mixed departments that do not work together are a different problem from any single department. The best brief we have had for one came from a bank that wanted four departments to understand each other's jobs. The game was built so that each team had to do, in miniature, the work of another department to get through. The organiser's phrase for it was "walk in my shoes". If the goal is to connect departments rather than to entertain one, choose a game where the puzzle requires the other team's knowledge, and assign the teams so that knowledge is spread.
Leadership offsites are a special case: the question is usually not what the executives will enjoy but what they will not find beneath them. A mystery with real deduction, a time limit and no silliness in the framing is the safe shape.
Why games cross cultures when conversation does not
The hotel team had people who had worked together for five years and never met. That is now normal. A distributed team is also, very often, a multicultural one, and that is where choosing well matters most, because the wrong choice does not just bore people. It excludes them.
An HR director at a fintech company with 180 people across twenty-one countries told us about the year she ran a New Year quiz on films and music. One group complained it was in the wrong language. Another complained the films were all American and none were the ones they grew up with. Attendance was low, the feedback was angry, and she did not run anything for a year. A quiz is built on shared cultural knowledge, and a team from twenty-one countries does not have one. Any activity that rewards knowing the same things will split such a team along exactly the lines it was supposed to close.
What she agreed worked, when we asked, was the opposite kind of game: the arcade mechanics everyone on earth has met. Dodge the thing, collect the thing, beat the clock. "Does everyone know those?" we asked. "Everyone," she said. The same logic runs through every shared-task format. The puzzle is the same for everyone, the goal is explicit, the rules are written down, and the thing to talk about is in front of you rather than inside you. Conversation formats depend on conventions that differ by culture: how much to disclose, when to joke, how long a silence can be, whether interrupting is rude. A happy hour between a team in Germany, a team in Brazil and a team in India is three different events happening in one call. A heist is one event.
Two smaller things from the same conversations. First, the title of the game matters in a multicultural team: the same HR director hesitated over a game called "Save Christmas" because of colleagues in Israel, Georgia and Kazakhstan, and chose a game with a neutral name instead. Second, let teams pick their own working language inside the game while the game and the host run in the company language. She described the alternative, a game forced into one language, as something her executives would object to on sight.
We run events for teams across more than fifty countries, and this is the most consistent thing we see: the game is the common ground, and the conversation happens because of it, not instead of it. So if your team is distributed across cultures, weight filter one (a shared setting, or a setting that needs no cultural background at all) and filter three (try it) over anything that relies on people already having things in common.
The short version
- Ignore genre. Write down what "competitive" or "quiet" or "creative" actually requires: a winner, no spotlight, a story.
- Look for a setting the team already knows, from their work or from what they share outside it.
- Decide the tone that would get a complaint, and filter on it before looking at anything.
- Play two or three candidates yourselves before booking. If you cannot, ask why not.
- For a distributed or multicultural team, prefer a shared task over a shared conversation.
What is the best online game for a competitive team?
One with a score, a clock and a winner announced at the end, in a setting the team already knows something about. The format matters less than those three things. Timed mysteries and heists with team rankings are the usual fit.
Do sales teams and engineering teams really want different games?
In our experience, yes, but less by genre than by what they will not forgive. Sales teams will not forgive a game with nothing at stake. Engineers will not forgive a puzzle whose logic does not hold. Pick for the thing they will not forgive.
How do I choose a team game for people who have never met in person?
Choose a shared task over a conversation format, and run a short low-stakes warm-up a week or two before, something people can do in their own time, so faces are familiar by the time the main event starts. The hotel team asked for two warm-ups in the four weeks before the event and was specific about the shape: individual, not team-based, because they did not want to reveal the teams yet, and built on photos, because "our team really likes showing photos". A recent holiday photo, a filtered photo, a photo of the view from the desk. Low effort, no performance, and everyone sees forty faces before the game starts.
How many people should be on a team inside the game?
Four to eight. Below four, one person can carry it and the others watch. Above eight, the quiet ones have nothing to do. If you want people who do not normally work together to mix, assign the teams yourself rather than letting them self-select.
Can I try a game before I commit?
You should, and with us you can. The free co-op challenge runs for up to five people in the browser with no signup, and the full catalogue games can be trialled by organisers before booking. If a vendor will not let you play before paying, treat that as a reason to look elsewhere.
Does the game need to match the occasion, like a holiday party?
Less than organisers think. The hotel team's December event is a hotel mystery, not a winter game. A theme that matches the team beats a theme that matches the calendar. Add the occasion in the framing and the prizes, not the game.