How to run a workshop when you're not a trained facilitator

The job is smaller than the title suggests

"Facilitator" sounds like a discipline you need a qualification for. In practice, running a workshop well comes down to three responsibilities, and none of them requires training.

You decide what the group is trying to produce. You decide how they will produce it. And you protect the conditions that let them — mainly time and airtime.

That is the whole job. Everything else people associate with facilitation, the flip charts and the sticky notes and the energiser games, is technique in service of those three things. Technique is useful, but a session with a clear purpose and a bad flip chart beats a beautiful session with no purpose every time.

Decide the output before you plan a single activity

The most common reason workshops fail is that they were designed as an agenda rather than towards an output.

An agenda answers "what will we do?". An output answers "what will exist at the end that does not exist now?". The second question is much harder and much more useful. A prioritised list of five things. A decision about which option to take. A written definition everyone agrees with. A list of risks with owners.

If you cannot state the output in one sentence, the session is not ready to plan. And if the honest answer is "shared understanding", push harder — shared understanding of what, and how would you know afterwards whether you got it?

Design backwards from the output

Once the output is fixed, the plan mostly writes itself. Ask what the group needs to have done in order for that thing to exist, then what they need before that.

A prioritised list requires a long list to prioritise from, which requires people to generate options, which requires them to share the same understanding of the problem. That is your session: frame the problem, generate options, narrow, prioritise. Four blocks, in that order, and now you know roughly how long each needs.

This is also how you find out the session is impossible in the time available — which is much better discovered while planning than at minute fifty.

The five ways it goes wrong

1. One person dominates. Usually the most senior or the most confident, usually not deliberately. The fix is structural: have people generate independently before any discussion, so the room has its own ideas before it hears theirs.
2. The group solves the wrong problem. Everyone arrives with a slightly different idea of what is being discussed, and nobody checks. Spend the first block establishing the problem, and get people to state it back rather than nodding.
3. It runs out of time at the important part. Almost always because the opening discussion was allowed to expand. Set a visible time for each block and end blocks on time even when they are not finished. Unfinished is information.
4. Everyone agrees, and nothing was true. Politeness looks like consensus. If a room agrees quickly on something difficult, you have not found agreement, you have found the cost of disagreeing. Ask the difficult question privately and show what comes back.
5. Nothing happens afterwards. The most common failure of all. A workshop that produces a good output and no owners produces nothing. Do not let people leave until each item has a name and a date attached.

What to do in the first five minutes

The opening sets what kind of session this is going to be, and most inexperienced facilitators spend it on housekeeping.

Say what the output is, in the sentence you wrote earlier. Say how long each part will take. Say explicitly what is in scope and what is not, because half the derailments come from someone raising a legitimate issue that belongs in a different meeting.

Then get everyone to say something, immediately, before the session properly starts. Not an icebreaker for its own sake — one short question relevant to the topic that everyone answers at once. The purpose is that everyone has now spoken, which makes speaking again much cheaper. A person who has been silent for forty minutes will usually stay silent.

Techniques that are worth the effort

• Independent generation before discussion. Everyone writes their own ideas privately, then you collect them together. This single technique prevents the dominance problem and the anchoring problem at once, and it is the closest thing to a universal improvement.
• Ask for the objection, explicitly. "What is the strongest argument against this?" gets you something. "Any concerns?" gets you silence. Naming it as a requested contribution removes the social cost.
• Timebox and say so out loud. "We have eight minutes on this" changes how people use eight minutes. A visible countdown does more work than any amount of encouragement.
• Write it where everyone can see. Half the disagreements in workshops are about what was said ten minutes ago. Capturing decisions visibly ends those before they start.
• Ask what is still unclear at the end. Privately, in one line each. This is the highest-yield ninety seconds of most sessions and almost nobody does it.

Techniques you can safely skip

You do not need energisers, physical games, or anything that requires people to stand up and move around. They work for some rooms and actively alienate others, and they are not load-bearing.

You do not need to be entertaining. Groups forgive a dull facilitator far more readily than a directionless one.

You do not need to fill silence. After you ask a question, the pause before someone answers will feel much longer to you than to anyone else. Most people fill it and answer their own question, which teaches the room not to bother.

And you do not need a script. You need the output, the blocks, and the times. Everything between those is judgement, and judgement is what you already have.

The gap between an untrained facilitator and a trained one is smaller than the gap between a session designed towards an output and one that was not. Almost everyone can close the first gap by planning carefully. Very few sessions survive the second.

Decide what should exist at the end. Work backwards. Protect the time and the airtime. That is most of it.