Every list of virtual team building activities is written by someone who has to organise one. This one is written for the people who have to sit through it.
We took the fifteen activities that appear most often across the five most cited lists in the category and scored each one, 0 to 10, on the five things that employees who dread these events say they are actually dreading. The rubric is below in full. If you disagree with a score, the point of publishing the rubric is that you can change it and rescore.
First, the conflict of interest
One of the fifteen activities is ours. We make a co-op game for remote teams, so we have the strongest possible interest in it scoring well, and it does. Read the scores with that in mind. We have put the rubric out in the open so the result is checkable rather than taken on trust, and the criteria were fixed before anything was scored. If you think a criterion was chosen to flatter the shared-task format, say so: that is a legitimate critique of this piece, and we would rather hear it than not.
What this is and what it is not
This is a scoring, not a measurement. Each activity was rated against five fixed criteria by the people who run virtual team events for a living. Nobody surveyed a team after playing. The measured version, where real teams run the same fifteen activities and report whether they would do each one again, is in progress and will be published on the same page when there are at least three teams per activity. If the measured results contradict the scores below, the contradiction will be published too.
The sample is not ours. An activity is on the list if it appears in the top ten of at least two of the five most cited virtual team building roundups. We did not add favourites or remove anything that scored badly for us.
The rubric: five things people actually hate
Ask people who dread virtual team building what they dread and the list is short and consistent. It is not "spending time with colleagues". It is these five.
| Criterion | 0 | 1 | 2 | |
|---|---|---|---|---|
| A | Spotlight. One person on screen doing something while everyone watches | never | occasionally, briefly | it is the mechanic |
| B | Forced disclosure. You are required to produce personal content about yourself on demand | no | optional or trivial | mandatory |
| C | Idle time. Stretches where you have nothing to do except wait for your turn | almost none | noticeable | most of the time |
| D | No shared task. There is no common goal that actions move toward | shared task | task exists but is individual | no task, only talking |
| E | Dies in silence. It only works if a host or organiser keeps pushing energy into it | no | partly | completely |
Higher is worse. Ten is the most cringe an activity can be on this scale.
The scores
| Rank | Activity | A | B | C | D | E | Score |
|---|---|---|---|---|---|---|---|
| 1 | Two truths and a lie | 2 | 2 | 2 | 2 | 2 | 10 |
| 2 | Show and tell, desk tour | 2 | 2 | 2 | 2 | 1 | 9 |
| 3 | Virtual happy hour | 1 | 1 | 2 | 2 | 2 | 8 |
| 3 | Guess the baby photo | 1 | 2 | 2 | 2 | 1 | 8 |
| 5 | Virtual scavenger hunt | 2 | 2 | 1 | 1 | 1 | 7 |
| 6 | Pictionary, drawing games | 2 | 0 | 2 | 1 | 1 | 6 |
| 6 | Coffee roulette, random 1:1s | 1 | 1 | 0 | 2 | 2 | 6 |
| 8 | Gift box unboxing | 1 | 0 | 1 | 2 | 1 | 5 |
| 9 | Kahoot and live quizzes | 1 | 0 | 1 | 1 | 1 | 4 |
| 10 | Codenames online | 1 | 0 | 1 | 0 | 1 | 3 |
| 11 | Shared team playlist | 0 | 1 | 0 | 1 | 0 | 2 |
| 11 | Hosted murder mystery | 1 | 0 | 0 | 0 | 1 | 2 |
| 11 | Team trivia with company questions | 0 | 0 | 1 | 0 | 1 | 2 |
| 14 | Hosted online escape room | 0 | 0 | 0 | 0 | 1 | 1 |
| 15 | HeySparko co-op challenge (ours) | 0 | 0 | 0 | 0 | 0 | 0 |
What the table says
The top of the list is all talking in turns. Two truths and a lie, desk tours, happy hours and baby photos share a shape: one person produces something about themselves, everyone else watches and waits, and nothing is being built. They score high on every criterion at once because they are the same activity with different props.
The bottom of the list is all shared tasks. Escape rooms, murder mysteries, team trivia, Codenames. Nobody is on stage because everyone is busy. There is no disclosure because the subject is the puzzle, not you. Idle time is low because there is always something to do. The host matters less because the task carries the energy.
Quizzes are not cringe. They are boring, which is different. Kahoot scores 4, not 8, because answering a question is an action and nobody has to reveal anything. The complaint about quizzes is that they reward whoever knows the most trivia, which is a fairness problem, not a cringe problem, and this scale does not measure it.
Coffee roulette scores worse than people expect. A random 1:1 between two colleagues who have never spoken has no task, no structure and no host. It works beautifully for the pairs who can hold a conversation with a stranger and is a quiet fifteen minutes of pain for everyone else. The organiser never finds out which it was.
Scavenger hunts are a surprise seven. They look like action, and they are, but the action is "run around your home on camera and show it", which is spotlight plus disclosure in one move. For people who do not want colleagues to see where they live, it is one of the worst formats on the list.
What this looks like with a real team
The rubric is abstract. The thing it is scoring is not, so here is the same table as it plays out on an ordinary Thursday, assembled from what we see on calls with the people who organise these events and from the largest body of employee opinion on the subject we know of: a corpus of 4,772 Reddit posts and comments about workplace team building that we analysed in September 2026. Nobody is quoted by name from it. Aggregates only, in our own words.
The warm-up. Forty people on a call. The organiser, who has told us privately that they do not believe in this either but have been asked to run it, opens with two truths and a lie. Person one goes. Person two goes. By person six the chat has gone quiet in the way that means a second, private chat has opened. The people with cameras off are not coming back on. This is a 10 on the table and everyone on the call can feel it. In the corpus, "cringe" and its relatives appear in 38 threads, and in most of them the organiser is the one using the word, in the title, while asking for something that is not.
The quiz. A company with staff in twenty-one countries ran a New Year quiz on films and music. Half the room complained it was in the wrong language. The other half complained it only had American films. The organiser told us she put the whole idea down for a year afterwards. A quiz scores 4 on cringe and that is accurate: nobody was embarrassed. They were excluded, which is a different failure that this scale does not measure and that any activity built on shared cultural knowledge will produce in a team that does not share one.
The escape room that went wrong. The most upvoted account of a virtual escape room in the corpus is a disaster: half the group could not get the link to work, one microphone was broken, twenty minutes were spent arguing about a virtual chest, and the facilitator called time. Then HR sent a survey asking how connected everyone felt. The activity scored 1 on our rubric and the event was still terrible, because the rubric scores the design, and that event failed on the door. Low cringe is necessary. It is not sufficient.
The one that worked. The praise in the corpus for the activities that did work is consistent to the point of monotony, and none of it is about theme. It is "it needed you to think and work together", "just enough of a competitive touch", "an hour was fun, six days would be hell". Shared task, a bit of a score, a fixed and short length. That is the bottom third of the table, described by people who do not know the table exists.
What people hate more than cringe. The thing the corpus is most angry about is not any activity. It is being required to attend one, especially outside working hours. "Mandatory" appears in 255 threads, "cringe" in 38. The seven most upvoted posts in the whole corpus are all about obligation and personal time, not about the game being bad. So the first design decision is not which activity. It is: during work hours, short, and genuinely optional. Get that wrong and the best-scoring activity on this table still loses.
How to use this on your own team
You do not need our scores. You need the rubric.
Take whatever you are about to book and score it on A to E. Anything with a 2 in both A and B is going to lose the quiet half of your team before it starts. Anything with a 2 in D has no floor: it is as good as the mood on the day and no better. Anything scoring 3 or under is safe to put in front of sceptics.
If you want to go further than scoring and measure it, we have published the procedure we are using: one real team, four to eight people, each activity on a separate day, and three anonymous questions afterwards, of which the only one that matters is "would you do this again". Three teams per activity is the minimum before you can say anything. One team is an anecdote.
Where do the employee quotes and numbers come from?
From a corpus of 4,772 public Reddit posts and comments about workplace team building, collected and coded in September 2026, and from discovery calls with the people who organise these events for their companies. Reddit material is reported as aggregates and paraphrase only: no individual is quoted, linked or named, because the people writing those posts did not write them for a vendor's article.
Who scored these, and how?
The HeySparko team, who run hosted virtual events for remote companies, scored each activity against the rubric above. The criteria were fixed before scoring began, and the full rubric is published so that the scores can be checked and redone by anyone, including people with no reason to like our answer.
Why is your own activity on the list, and why does it score zero?
It is on the list because the measured version needs a shared-task activity with no host to test against the hosted ones, and ours is the one we can run with any team for free. It scores zero because it is a ten-minute co-op task with no talking required, which is exactly the shape the rubric rewards. That is also the conflict of interest stated at the top. The measured version will tell us whether real teams agree.
Does a low score mean people will enjoy it?
No. The scale measures how much an activity exposes people to the five things they say they hate. An activity can be low-cringe and dull. Enjoyment is what the measured version asks about.
Why isn't my favourite activity here?
The sample is the activities that appear in the top ten of at least two of the five most cited lists. That is a popularity filter, not a quality one. Score your favourite on the rubric and you have the same information.
Can I reuse this table?
Yes. The rubric and scores are published under CC BY 4.0. Name the source and the date, and link the rubric so your readers can rescore.
When will the measured results be out?
When every activity has been run with at least three real teams. We will update this page rather than publish a new one, and the date of the last update will be at the top.