Key takeaways
- →Workshop pre-work gets completed when it is under ten minutes, has exactly one deliverable, and produces something the group can see in the room.
- →Asking people to read is a request for attention you cannot verify; asking them to answer one question gives you a completion rate and usable material.
- →Send pre-work 48 to 72 hours ahead, not a week: earlier means forgotten, later means squeezed out by the day job.
- →If nothing in the first thirty minutes depends on the pre-work, you have taught the group that pre-work is optional, and next time it will be.
- →Design a fifteen-minute recovery path for the people who did not do it, and run it without commentary or shaming.
- →Track completion every time. A rate below about 60 per cent is a design problem in the task, not a motivation problem in the group.
Workshop pre-work fails for a predictable reason: it asks for too much, too vaguely, too far in advance, and then does not visibly matter once the session starts. People are rational about this. If the first thirty minutes of the workshop recap the pre-reading anyway, skipping it was the correct decision and everyone learns that for next time.
The fix is not a firmer reminder email. It is a different design: one task, under ten minutes, with a deliverable that shows up on screen in the first activity. When people can see their own submitted answer being used, completion for the next session rises without you asking for anything.
This piece covers what to ask for, how long it should take, when to send it, how to build the session so the pre-work is load-bearing, and the recovery path for the third of the room who will not have done it regardless.
Why pre-work goes undone
The first reason is size. A twelve-page strategy document sent to fourteen people is roughly four hours of collective reading time requested by email with no deadline enforcement. Most people will open it, scroll, and decide to skim it in the ten minutes before the session.
The second is ambiguity about the deliverable. 'Have a think about our priorities for next quarter' cannot be completed, because there is no state in which you have finished it. Thinking expands to fill nothing at all. A task with a visible end state, such as 'name the one thing you would stop doing', can be finished on a train.
The third is that it is invisible. If the facilitator opens with a summary of the material, everyone who read it has been penalised with duplicate effort and everyone who did not has been rewarded. Groups calibrate on this within one session.
The fourth is timing. Sent a fortnight out, it drops out of memory. Sent the night before, it collides with everything else that day. There is a narrow window, roughly two to three days ahead, where it is close enough to feel real and far enough to be schedulable.
✓ Pros
- •One question with one answer, submitted somewhere you can see it
- •A hard cap you state up front: 'this takes eight minutes'
- •Material the group produces rather than material they consume
- •A first activity that literally cannot run without the submissions
- •A visible completion count, so the group knows it is being tracked
✗ Cons
- •A reading pack with no question attached to it
- •'Come with your thoughts on the roadmap'
- •Anything requiring a login people do not already have
- •Pre-work sent more than a week ahead with no reminder
- •A session that opens by summarising the pre-work for everyone
The five properties of pre-work that gets done
Test any pre-work task against five properties before you send it, and rewrite until it passes all five. This takes about five minutes and is the highest-return five minutes in the whole session design.
It has one deliverable. Not read this and consider that and bring some examples. One thing, phrased as an instruction with a verb: submit, rank, name, choose.
It has a stated time cost, and the cost is honest. Say 'eight minutes' and mean it. An accurate small number does more for completion than any amount of explaining why the workshop matters.
It is doable on a phone. A meaningful share of pre-work gets done on a commute or between meetings. If it requires a spreadsheet, a VPN or a document that renders badly on mobile, you have lost that share.
It produces something. Silent reading leaves you with no material and no completion data. A submitted answer gives you both, and gives the group a head start on the session's first activity.
It is used in the first twenty minutes. If the material appears an hour in, the people who did it have already sat through an hour that did not need them.
One instruction, one verb
Name the single thing you would change. Rank these five. Submit one example. If the task needs an 'and', split it or cut it.
An honest time cost
State the number of minutes in the first line. Eight minutes gets done; 'shouldn't take long' gets postponed indefinitely.
Phone-completable
One link, no login, no document. If it cannot be finished on a train with one thumb, expect completion to halve.
Produces material
The output is the session's raw input: answers to sort, ideas to vote on, estimates to compare. Reading produces nothing you can use.
Visibly used early
The submissions appear on screen inside the first twenty minutes, in the participants' own words.
Choosing the right kind of pre-work
Match the task to what the session actually needs from people, rather than defaulting to a reading pack because it is easy to attach.
If the session needs shared context, send a summary of no more than 400 words with one comprehension question attached. The question is not a test; it is the thing that converts reading into a completable task, and it tells you who is arriving cold.
If the session needs divergent input, collect it in advance. Ask each person to submit two or three ideas independently before the session, which is the pre-work version of the nominal group technique and avoids the production blocking you get when fourteen people try to generate ideas in the same room. You arrive with sixty ideas already on the board and spend the session choosing rather than generating.
If the session needs positions, ask for one. A single-line answer to 'what should we do about X' from every attendee, collected privately, gives you a genuine map of disagreement before anyone has anchored on the loudest voice in the room.
If the session needs data, ask each person for their own number. Estimates, current counts, this quarter's figures. Collecting these live wastes twenty minutes and produces guesses.
If the session needs a skill refresher, use a short quiz rather than a document. Retrieval practice is a more efficient use of ten minutes than re-reading, and the score distribution tells you exactly which part of the training to spend the session on.
How long pre-work should take
Ten minutes is the working ceiling for a routine internal session. Beyond that you are asking for a meaningful piece of someone's working day and you need a matching justification, an explicit ask from the person who called the meeting, and time visibly protected for it.
The numbers below are the defaults worth starting from. They are guidance rather than research findings, but they are the shape that survives contact with real calendars, and each of them exists to keep the request smaller than the resistance to doing it.
The send: timing, channel and wording
Send 48 to 72 hours ahead
Close enough that the session feels real, far enough that people can find ten minutes. Avoid Friday afternoons and the evening before, which is when pre-work goes to die.
Put the time cost in the first line
'This takes eight minutes and closes on Wednesday at 5pm.' Not in paragraph three, not after the context. People decide whether to open the link on the first line.
Say what it is for, in one sentence
'Your answers are the starting list we will rank in Thursday's session.' The point is not motivation, it is showing that skipping it removes your input from a decision.
Send it in the channel people actually read
For most teams that is chat, not email. One link, no attachment, no login. Every additional click costs you completions.
Set a close time before the session
The evening before works well. It gives you the night to sort the responses and gives late finishers a real deadline rather than a vague one.
Reminder once, twenty-four hours out
One reminder with the completion count so far. Two or more reminders stop being information and start being nagging, and the group learns to wait for the last one.
Making the pre-work load-bearing
This is the part most facilitators skip, and it is the only part that changes behaviour for the next session. The pre-work has to be structurally necessary, not merely referenced.
The strongest version is opening the session with the submissions themselves on screen: the raw list of ideas, the spread of positions, the distribution of estimates. The first activity is then sorting, clustering or voting on material the group made, which cannot happen if the material does not exist. With Session Flo you can collect the responses ahead of time and open the room straight onto the results, which removes the twenty-minute warm-up most workshops spend generating what could have arrived in advance.
A weaker but still workable version is to reference specific contributions early: 'four of you raised the handover problem, so we are starting there.' It has to be specific enough that people recognise their own answer.
The version to avoid is the recap. Summarising the pre-reading for the benefit of those who missed it is generous and it is exactly what teaches the group to skip it. Do not do it, even when you can see people arriving cold.
T-72 hours: send
One link, time cost stated, close time given, sent in the channel the team actually reads.
T-24 hours: one reminder
Include the current completion count. Share it as information rather than pressure.
T-12 hours: close and sort
Cluster the submissions, remove duplicates, keep the wording people used. Note the completion rate.
Minute 0: open on the material
The first thing on screen is what the group submitted, not your agenda slide.
Minute 5: absorb the non-completers
Pair them with someone who submitted. Ten minutes, no commentary, everyone back in the room.
Minute 15: first decision
Rank, vote or cluster the pre-work output so the session's first outcome comes from it directly.
When a third of the room has not done it
Plan for this every time, because it happens every time. The goal is to get those people usefully into the session inside fifteen minutes without spending the group's attention on the fact that they arrived unprepared.
Pair them. Put each non-completer with someone who submitted and give the pair one task: the person who did the pre-work explains their answer, the other asks two questions and adds one thing. Ten minutes. The prepared person gets to articulate their thinking, which is a genuine benefit rather than a consolation, and the unprepared person arrives with a position rather than a blank.
Do not run a catch-up briefing for the whole room, do not apologise for the pre-work, and do not ask who did it by a show of hands. Public accounting produces defensiveness and turns a design problem into a discipline problem.
If completion comes in under about 60 per cent, treat that as feedback on the task rather than on the group. Nine times out of ten the task was too long, had more than one deliverable, or was sent too far ahead.
Tracking completion and improving the next one
Top Tips
- Record the completion rate for every session in the same place, with the time cost you claimed and the actual lead time. Three data points will tell you your team's real ceiling.
- Ask one question at the end of the session: was the pre-work worth the ten minutes? A single yes-no from the group is enough to catch a task that quietly wasted everyone's time.
- If the same three people never complete, that is a conversation to have privately, not a reason to redesign the task for everyone else.
- Watch for pre-work inflation. Once it works, the temptation is to add a second question, then a document, then a spreadsheet. Guard the ten minutes.
- Reuse the format, not the content. Once a team knows what your pre-work looks like, the friction to complete it drops because they already know it will be short.
- When completion is high but the material is thin, the question was too easy rather than too long. Ask for a position rather than an observation.
Nobody skips pre-work because they do not care. They skip it because the last three times, the session ran exactly the same for the people who did not.
Frequently asked questions
How long should workshop pre-work take?
Eight to ten minutes for a routine internal session, and say the number in the first line of the message. Anything over about twenty minutes needs an explicit ask from whoever called the meeting and time genuinely protected in people's calendars, because otherwise it competes with delivery work and loses. Long pre-work is not automatically wrong, but it needs to be treated as work rather than as a favour.
What if people just read the pre-work in the meeting itself?
That is a legitimate design if you plan it. Blocking the first ten minutes for silent reading in the room guarantees a shared starting point and removes the completion problem entirely. The trade-off is ten minutes of expensive collective time, so it suits short briefs where alignment matters more than volume. What does not work is doing both: sending pre-work and then also reading it in the room.
Should pre-work be anonymous?
Anonymous when you are asking for positions, risks or anything with a political cost, because named submissions in advance get self-censored more heavily than live answers. Named when you are collecting data, questions or examples you will need to follow up on. If you go anonymous, say so in the request itself; people decide how honest to be before they start typing, not afterwards.
How do I get senior people to do pre-work?
Make it shorter than everyone else's and make the consequence of skipping it visible. One question, three minutes, and their answer appears on screen alongside everyone else's in the first activity. Senior attendees skip pre-work when it looks like homework and complete it when it looks like the mechanism by which their view gets into the room. Never send them a reading pack and expect it read.
Is pre-work worth it for a one-hour meeting?
Often yes, and it is where the return is highest. Collecting the ideas, questions or positions in advance can turn a sixty-minute meeting that spends thirty minutes generating material into one that spends fifty minutes deciding. Keep it to three minutes and one question, and open the meeting on the results rather than on an agenda slide.
What do I do if the pre-work responses are poor quality?
Use them anyway, in the participants' own words, and fix the question next time. Thin answers usually mean the prompt asked for an opinion rather than a choice: 'thoughts on the roadmap' produces mush, while 'name the one item you would cut' produces something you can rank. Rewriting or upgrading the submissions before you display them is worse than showing thin material, because the group stops trusting that what appears on screen is what they wrote.
Keep reading
- building a workshop agenda from outcomes backwards — Deciding what the session must produce before you decide what to ask for in advance.
- synchronous vs asynchronous workshops — How much of the work can move out of the room entirely.
- silent brainstorming and independent idea generation — Why collecting ideas before the session produces more of them.
- opening a workshop in the first five minutes — Putting the pre-work material on screen instead of an agenda slide.
- collect pre-work responses with Session Flo — Gathering submissions in advance and opening the session straight onto the results.