A leadership offsite changes something when it's designed around one real conversation the team has been avoiding, and when the week after gets as much design as the event itself. Most offsites are designed around an agenda instead, which is why most offsites produce a nice dinner, a full flipchart, and a team that's identical on Tuesday.
I say this as someone who runs offsites. The event is the easy part. The change is a reentry problem, and almost nobody plans for reentry.
Why don't most leadership offsites change anything?
Because they're built to feel productive instead of to be consequential.
The standard failure is the packed agenda: strategy review, breakout groups, a speaker, vision work, maybe a personality assessment with colors. Every hour is full, which guarantees no hour goes deep, and the topics the team most needs to address never make the schedule because nobody wanted to put "why Sales and Ops haven't really spoken since March" on a slide.
The second failure is treating openness at the offsite as the achievement. Teams genuinely do open up at these things. Off the clock, away from the building, after a glass of wine, people say true things to each other. Then Monday arrives, the conditions that made truth expensive are exactly where the team left them, and the openness evaporates so completely that mentioning it feels embarrassing. I've heard executives refer to a past offsite the way people refer to a vacation fling.
What should a leadership offsite actually be for?
One thing: having the conversation that can't seem to happen at the office, with enough time and safety that it actually finishes.
You already know what the conversation is. Every CEO does. It's the one you thought of when you read the last sentence, the strategy doubt, the broken peer relationship, the commitment everyone made and nobody believes. An offsite earns its cost by giving that conversation four protected hours instead of the eleven rushed minutes it would get at the end of a Tuesday agenda.
Everything else, the strategy review, the planning, the dinner, is fine as supporting material. But if you can't name the one conversation, postpone the offsite, because you're about to spend $30K cater-feeding the avoidance.
How do you design an offsite that sticks?
Strip the agenda until it's almost uncomfortable. One real conversation, with room to breathe, beats six topics every time. Teams panic at white space on an agenda, and white space is precisely where the real material comes out.
Do the surfacing work beforehand. Short private conversations with each executive ahead of the event, what does this team need to talk about that it doesn't, change everything, because the hard topics arrive at the offsite already named instead of needing to be excavated on-site, which burns half a day and most of the courage in the room.
And the CEO opens with the most honest thing said all year. I watched a CEO start an offsite with: "Our biggest customer is probably leaving, I've known for three weeks, and I haven't told you because I didn't trust how this team would handle it." Brutal opening. It also set the honesty bar for the entire two days, and the team spent them having conversations they'd deferred for two years, including what it meant that their CEO didn't trust them with bad news. That admission was trust work, the conversations that followed it were candor, and the commitments that came out the other side actually held, because people made them with the truth on the table. The order those things happened in wasn't luck. It was the Six Shifts running at compressed speed, 48 hours instead of a year, and the compression only held because the weeks afterward defended it.
What happens the Tuesday after?
The offsite's fate gets decided in the first two weeks back, so design those weeks before you design the event.
Whatever got honest at the offsite needs a scheduled continuation in the regular rhythm: the conversation reopens at the next leadership meeting, in the building, at normal altitude. The first time someone says "like we discussed at the offsite" and the room actually picks the thread up instead of smiling politely, the event starts paying. And the first hard moment after reentry is the real test, the first miss, the first conflict. If those get handled the old way, the team concludes, accurately, that the offsite was theater, and the next one starts in a deeper hole.
One commitment beats ten. Teams leave offsites with poster paper full of intentions and keep none of them. Leave with one behavioral agreement, ours was "disagreements about the forecast get voiced in the meeting, that week," and enforce it like payroll.