Your annual strategy retreat is wasted. Here is a better one

Most annual strategy retreats fail for a simple reason: they are designed as events, not as decisions. Leaders spend two days hearing presentations, generating priorities and enjoying a good venue, then return to the same calendar and the same budget. A better retreat is smaller in ambition and harder in content: it reviews last year's bets honestly, makes three to five explicit choices, decides what stops, and books the follow-up before anyone leaves.

The belief: a retreat is where strategy gets made

The belief is reasonable. Senior people are busy; taking them out of the office for two days seems the only way to think beyond the next quarter. A new setting, a facilitator and some distance from email should produce clarity.

And retreats do produce things: energy, goodwill, a shared vocabulary, sometimes a better relationship between two executives who rarely speak. Those matter. But they are not strategy, and they fade within weeks if nothing in the organisation changes on Monday.

Why it fails in practice

It starts from a blank page. Many retreats open with "where do we want to be in five years?" without first looking at what happened to last year's priorities. Without that review, the room repeats the same aspirations in new words, and nobody asks why last year's list did not move.

It adds and never subtracts. Retreats are good at generating priorities and bad at removing them. Ten new initiatives join last year's fifteen. Nobody has more people or more hours, so everything gets a little attention and nothing gets enough.

The decisions are not really decisions. "We will focus on customer experience" commits nobody to anything. A real decision names an owner, a budget, a date and something that will no longer be done.

The agenda is presentations. When each executive presents their area for forty minutes, the retreat becomes a long management meeting in a nicer room. Presentations defend; they do not choose.

Seniority shapes the outcome. In most rooms, the most senior voice speaks first and the rest calibrate. This is a feature of human groups, not a flaw of any one leader, but a retreat that does nothing to counter it ends up as a well-catered confirmation of what the chief executive already thought.

There is no Monday. The retreat ends with a group photo and a promise to circulate the slides. The slides arrive three weeks later. By then, the next quarter's operational pressure has filled every calendar.

What works instead

A useful retreat is designed backwards from the decisions it must produce. Here is a structure for one and a half days that works for a leadership team of eight to fifteen.

Before: four weeks out

• Write the questions, not the agenda. Agree with the chief executive the three to five decisions the retreat must make. For example: which two search fields get innovation funding next year; which of last year's initiatives stop; what the target mix between core and new looks like.
• Prepare an honest review of last year. One page per priority from the last retreat: what was promised, what happened, what was learned. Written by someone neutral, not the owner.
• Map the current portfolio. Every major initiative on a portfolio map, so the room sees where effort actually goes.
• Collect views anonymously. A short survey: what should we stop, what are we avoiding, what will hurt us in three years. Anonymous answers surface what nobody says in the room.

Day one, morning: look back before looking forward

Run a structured retrospective on last year's priorities. What worked, what did not, what we would do differently. Keep it about decisions and conditions, not about individuals. Then read the anonymous survey results aloud.

Day one, afternoon: make the choices

Use the strategic choice cascade or three horizons to frame the decisions. For each decision question:

1. Everyone writes their answer silently before anyone speaks.
2. Answers go on the wall at the same time.
3. The chief executive speaks last.
4. The room settles on a choice or, if it cannot, names who will decide and by when.

Day two, morning: subtract and stress-test

• Build a stop-doing list. Rule of thumb: nothing new enters next year's priorities unless something leaves.
• Run a premortem on the chosen priorities. Gary Klein's technique asks the team to imagine it is a year later and the plan has failed, then to write down why. It legitimises doubt that seniority would otherwise silence.

Day two, before lunch: owners and Monday

• Each choice gets one owner, first-quarter objectives and key results, and a first action within two weeks.
• Book four quarterly review dates in the room, on the calendar, with the same people.
• Agree a one-page summary that goes out within 48 hours, not three weeks.

Illustrative example

Illustrative. The leadership team of a mid-sized engineering university in Australia had held the same retreat format for six years: a coastal hotel, presentations from each faculty dean, and a closing session generating "strategic themes". Each year the themes grew; none were ever retired.

The new format started with a neutral review of the previous year's seven themes. Two had clear progress; five had no owner or budget. The anonymous survey's most frequent answer to "what should we stop?" was a set of overlapping committees.

In the afternoon, the team made three choices: one research area to build, one change to the first-year student experience, and a simplified committee structure. On day two, the stop-doing list retired four of the old themes and merged three committees. The premortem raised a risk nobody had voiced: the student-experience change depended on a timetable system the IT department had not been asked about. The IT director was added as co-owner before lunch.

The summary went out two days later. The first quarterly review was held on the date booked in the room.

Do's and don'ts

Do

• Agree the decisions the retreat must make before booking the venue.
• Review last year's priorities honestly before setting new ones.
• Collect anonymous views in advance and read them aloud.
• Make people write before they speak, and let the most senior person speak last.
• Leave with owners, dates and booked review meetings.

Don't

• Don't fill the agenda with departmental presentations. Send them as pre-reads.
• Don't add priorities without removing others.
• Don't invite thirty people to a decision meeting. Brief them afterwards.
• Don't confuse team building with strategy. Do both if you like, but separate them on the agenda.
• Don't let the summary take three weeks.

Further reading in this hub

The retreat is one moment in a year-long cycle. The full process for building the strategy behind it is in How to build an innovation strategy in a few workshops, and the quarterly follow-up is described in Innovation portfolio management: fund, scale, pause or stop. For the stop decisions the retreat will produce, see How to stop a project well.

Questions people ask

How long should a strategy retreat be? A day and a half is enough for most leadership teams if the preparation is done. Longer retreats usually fill the extra time with presentations rather than decisions.

Should we use an external facilitator for our leadership offsite? Often, yes, because an outside facilitator can enforce silent writing, time limits and speaking order without the political cost an insider would pay. The facilitator should design the session around your decision questions, not bring a standard agenda.

What should a strategy offsite agenda include? A review of last year's priorities, a short set of decision questions, a stop-doing session, a premortem on the chosen priorities, and a closing session that assigns owners and books follow-up reviews.

Where to go next

The guides used above are free: retrospective, stop-doing list, premortem and strategic choice cascade.