Most people choose team building backwards.

They start with the activity. What's available, what looks fun, what a mate did once and wouldn't shut up about. They book it, and they hope it fits. Sometimes it does. Often it doesn't, and you get a perfectly pleasant day that changes precisely nothing by the following Tuesday.

The better way round is to start with the problem. Work out what's genuinely not working in your team first, then pick the activity that fixes that specific thing. Sounds obvious when you say it out loud. Almost nobody actually does it.

So this is a guide to doing it the right way round. Diagnose first, book second.

Start with the problem, not the activity

The most useful idea I've come across on this is a simple one: the closer an activity resembles the real thing your team struggles with, the more likely any of it sticks. A day out can be collaborative, and fun, and still transfer nothing, if it has no bearing on what your people actually find hard. Fun on its own is lovely. Fun aimed at the right problem is what changes things.

There's a related trap worth naming, and I owe this one to a sharp exchange I had on a previous post. Trust built while you're all relaxed and having a laugh is not automatically the same trust that holds up when there's a real disagreement or a blown deadline. Comfort and conflict are different weather. Being able to rely on someone during a treasure hunt is a start, but it's not the same as trusting them when the pressure's genuinely on. The best team building gives people at least a taste of the second kind: a shared goal, a bit of pressure, real decisions to make together, not just a nice lunch and a name badge.

So before you book anything, ask one question. What's the actual problem here? The rest of this guide is built around the most common honest answers to that.

The brand new (or newly merged) team

You'll know it's you if: people are polite but stiff. Nobody's quite sure who does what, or who to go to for what. Meetings are a bit careful.

What's really going on: there's no shared history yet. People simply can't read each other, and everything feels slower because of it.

What suits it: low-pressure activities that get people talking, and learning each other's names, quirks and strengths, fast. Mixed small groups, a shared goal, and lots of natural chatter. With a new team you want sheer volume of interaction more than anything clever. Klein's meta-analysis on team building found most of its effect lands on the human stuff, trust and confidence, and a brand new team has none of that banked. Job one is just generating some.

What to skip: anything that spotlights individuals or creates early winners and losers. Far too soon for that. You'll cement first impressions you didn't mean to.

The team that doesn't talk to each other

You'll know it's you if: information sits in silos. People find things out late. The same mistake happens twice because nobody mentioned it the first time.

What's really going on: either there's no habit of sharing, or, more often, people don't feel safe enough to speak up.

What suits it: activities where the team literally cannot win without passing information around. Where the quiet one is holding the clue everyone else needs. Something that forces communication rather than politely encouraging it. This is the finding from Google's Project Aristotle, their big study of what made teams work: the top predictor wasn't talent or seniority, it was whether people felt safe to speak up. You can't lecture that into a team. You build it by handing people lots of small, low-risk wins that come from opening their mouth.

What to skip: anything one confident person can carry while the rest watch. That's the exact dynamic you're trying to break, so don't book a version of it.

The remote or hybrid team

You'll know it's you if: it's all business and no glue. Cameras off. Half the team have never met, and nobody really knows what anyone's like as a human.

What's really going on: connection has quietly worn away. This is everywhere now, so don't feel got at. Gallup found around 20% of workers felt lonely "a lot" the previous day, rising to roughly 25% for remote workers. Research cited by Personio put one in five UK employees as often or always lonely at work. That isn't just sad, it's the stuff that makes good people drift off.

What suits it: activities genuinely built to work across a screen, or, better still, an occasional proper reason to get everyone in one room. The rare in-person day matters more for a remote team than for anyone else, so make it earn its place with something interactive rather than a room full of slides and a sad sandwich platter.

What to skip: the default all-hands with a quiz bolted on the end. Better than nothing, sure. But it's not building much, and everyone knows it.

The burnt-out team

You'll know it's you if: heads down, running on empty, a bit flat. Good people, just tired.

What's really going on: they need rest and a real change of pace more than they need a development exercise. Reading this wrong is how you make burnout worse.

What suits it: honestly, this is the one time "just go and have some fun" is exactly the right answer. Something genuinely enjoyable, low-stakes, no hidden agenda, no learning objectives stapled to the back of it. Let them breathe and enjoy each other's company. That has real worth on its own. A team that actually likes each other communicates more openly and cuts each other more slack when things get hard, and sometimes the smartest move is simply to top that up.

What to skip: anything intense, or competitive to the point of stress, or that smells like more work in disguise. Read the room. A knackered team handed a "challenge" will quietly resent you.

The team with quiet friction

You'll know it's you if: the surface is fine but the undercurrent isn't. Little cliques. The odd passive-aggressive email. Meetings where the real conversation happens afterwards, in twos, by the kettle.

What's really going on: trust has broken somewhere, or sub-groups have formed and hardened.

What suits it: activities that deliberately mix people up and hand rival camps a shared goal they can only reach together. Nothing thaws a bit of tension like having to genuinely rely on someone you'd been quietly avoiding all quarter. This is where the shared-goal idea really earns its keep. The big review of team cohesion by Grossman and colleagues found that pulling together on a task predicts performance more than just socialising does. A common goal, even a playful one, beats a forced heart-to-heart every time.

What to skip: anything that could tip into real competition between the factions you already have. You want them shuffled and mixed, not dug further into their trenches.

The thing to avoid whatever your problem is

One pattern runs through every situation above: the passive stuff does less.

Sitting everyone round a dinner table is pleasant, but nobody has to cooperate to eat a lasagne. A wine tasting is lovely, but it asks nothing of the team. Watching a demo, listening to a speaker, being talked at in a nice venue, these are fine as treats, but they're not building much, because your people aren't actually doing anything together.

The activities that move the needle all share a shape. A shared goal. Real interdependence, so you genuinely need each other to get there. Everyone involved, rather than three loud people and a quiet majority. Decisions to make. And enough enjoyment that people forget they're being "developed" at all.

Why active, problem-solving formats tend to win

Which is the honest reason a well-built treasure hunt or scavenger hunt tends to punch above its weight. It isn't the novelty. It's that the format quietly ticks every box the research cares about. A shared goal. Real collaboration, because you can't win it on your own. A clock and some decisions, so it's not purely comfortable. And everyone in it, not spectating from the side.

And here's the bit that ties back to the top of this guide. The best versions aren't pulled off a shelf, they're built around your team's actual situation. The challenges can be shaped around the real thing your people find hard, whether that's getting information across silos, mixing up factions, or just remembering they quite like each other. Match the design to the problem you diagnosed at the start, and a "fun day out" quietly does a great deal more than fun.

The order of operations

So, to put a bow on it. Work out what's genuinely not working. Pick the format that targets that specific thing. Make sure it's active, shared and enjoyable rather than passive and watched. And where you can, get it tailored to your team rather than settling for one-size-fits-all.

Do that, and you'll never book another away day that everyone enjoys and nobody remembers.