Audience Dialogue

Know Your Audience: chapter 14
The co-discovery conference

Every other chapter in this book is about ways of researching the audience. Does the whole idea of "audience research" seem a little one-sided to you? It conjures up the picture of broadcasters (or others) finding out all sorts of things about their audiences, while the audiences themselves sit quietly, watch TV, and be researched. "Audience research" can imply that audiences are passive - which is certainly not the case. (Audience members usually know a lot more about the broadcasters than vice versa.)

Communication is most effective when it is interactive: when it works in both directions, and it is immediate. It takes two to communicate.

So I thought that one way to improve the effectiveness of research would be to enlarge the idea of consensus groups, and to combine research and planning. I remembered the search conference (one of the strands from which consensus groups were derived) and invented a new type of activity, a mixture of research and planning. I call this a co-discovery conference - to show that it's related to a search conference, but has a different emphasis.

Co-discovery conferences work like this: broadcasters (or staff of any organization which has an audience) and audience members meet for a day, find out more about one another, and together map out a future direction for the organization.

The ideal number of participants is about 30, with more audience members than broadcasters: about 18 audience members and 12 staff members. The difference partly reflects the fact that there are many more audience members than staff members, and partly gives more confidence to the audience members, who can feel overwhelmed at the expertise of the staff. The maximum I'd recommend is about 40 people, in the same proportions: 60% audience and 40% staff.

A co-discovery conference lasts for about 8 hours: one working day. Ideally, it will begin in the early afternoon, break overnight, and finish the next morning. The advantage of the overnight break is that it gives participants time to review the first half of the proceedings, and come up with new ideas. People tend to be more creative in the mornings than the afternoons, but the co-discovery conference format needs to go through a fact-finding stage before it can move on to creativity.

When to hold a co-discovery conference

Co-discovery conferences are intensive exercises, which require more staff time than most other types of audience research. They are best done as part of an regular planning exercise. If your organization is a broadcaster or publisher, many of the ideas will be for new programs: something that a media organization needs a constant supply of, and an annual co-discovery conference would be appropriate. For other organizations, the most appropriate situation for a co-discovery conference is when major change is about to occur, and you realize you know very little about your current (or potential) audience.

How to hold a co-discovery conference: Initial planning

Because a co-discovery conference is a participative process, it involves the audience having an influence on the management of the organization. If the senior managers are not comfortable with this idea, there's no point in doing a co-discovery conference at all. So the first thing to establish is to decided whether to go ahead at all. If the answer is positive, you must then organize:

You need to plan at least a week ahead; preferably two weeks. The slowest part is usually finding the audience participants.


The venue needed is a large room, with plenty of wall space to put up the findings. To ensure that there are no hidden agendas, everything written during a co-discovery conference is posted up on the walls, where everybody can read it and comment on it. The room should be at least 10 metres square - or about the size of a large school classroom.

Some of the time, the participants will be divided into small groups, of no more than 8 people each, so the room needs to be large enough that each group can have an uninterrupted discussion in one corner of the room.

The venue should not be the organization's offices. This can discourage participants from speaking their minds, but more important, having the co-discovery conference in the organization's office will encourage staff participants to become distracted from their main task in the conference.

Choosing participants

From the organization: choose staff who are in some way involved with communicating with the audience. For a broadcaster, this will include general managers, broadcasters, and reporters, and other program-makers. It will usually not include people like clerical staff, accountants, or human resources staff. However if there are planning staff they should be included, even if they have no direct contact with the audience, because a co-discovery conference is a planning exercise as much as it is research.

To get 12 staff who will come to the co-discovery conference, a few more will need to be selected, to allow for illness, sudden urgent work, and so on. Another alternative is to nominate an understudy for each designated attendee, to stand in for that person if necessary.

I suggest that all relevant staff should be told about the co-discovery conference, and asked to indicate their interest. There's no point in having delegates who are not interested, unless they are the managers most closely concerned with output to the audience. And if these people aren't interested, there's no point in holding a co-discovery conference at all.

Staff delegates should not be chosen as a favour, or a reward for good behaviour, but so as to include the widest possible variety of people, both in terms of current job and personal characteristics. The 12 or so staff delegates should include all major departments and job types, as well as young people, old people, men, and women, senior and junior staff, long-serving staff and recently appointed staff.

Participants need to be interested in the co-discovery conference, and able to spend a full day there without interruption. When you choose audience delegates, the same principles apply. The participants should include young and old people, men and women, long-term and new listeners (or whatever you call them), city and country dwellers, rich and poor, educated and uneducated - but all should be interested in the co-discovery conference, and able to spend a full day there without interruption.

Audience delegates can be selected in various ways:

All of these methods are explained in the chapter on sampling. Don't choose only people whom the staff already know. This is easy to do, but it won't produce a representative sample of the entire audience.

Conference managers

You need at least two people to manage the conference. As with a consensus group, these take the roles of moderator and secretary. But, unlike a consensus group, the moderator and secretary don't play an active part in shaping the outcomes. They are there only to manage the process. The participants themselves must do the rest.

One of the moderators can double as secretary, or a less-skilled secretary can be chosen: a junior clerical worker who will write up the comments without necessarily understanding the whole process. When there are two moderators, one is normally experienced and the other is a learner. Moderating a co-discovery conference is easy, most of the time. It's only when something goes wrong that management skills are needed - e.g. if there's a violent disagreement among participants, or the group runs far over time.

The role of the moderator is to explain the process to participants, make sure that everything runs more or less to time, and solve any minor problems that get in the way of the smooth running. The moderator should not answer questions about the content of the conference. If he or she makes that mistake, the participants will come to depend on the moderator (with his or her apparently superior knowledge) and won't be so committed to produce their own plans - or to follow them, after the moderator leaves. The moderator should not be an important manager, who would not be able to be objective in the face of criticism from the audience (or staff). However a manager with no direct authority over other staff present can be OK - e.g. the planning manager - as long as he or she is willing to disagree with the general manager if necessary.


It's not feasible to have more than about 15 staff attend a co-discovery conference, but many more staff will be needed to implement the plans that the conference produces. Therefore the other staff should have the opportunity to see what happened at the conference. They probably won't have time to watch or listen to a full 8-hour record. A one-hour summary is usually enough to show how the conference worked. And remember that the conference findings have been recorded on paper, for everybody to read.

To record a conference, you can use either a still camera, a video camera, or audio tape. Each photo, video scene, or audio tape should be clearly labelled, so that the viewers of the record know where each part fits in to the whole. It's best to have a recording team of two people: one to operate the camera or microphone, and the other to record what is being said at each moment.

A video report can be very persuasive - far more so than a written report. Obviously, it requires a lot of skill to make a good video report, but if your organization is a broadcaster, you should have staff with these skills.

If audio tape is used, the recordist needs to understand what is happening, and to sometimes talk to the tape recorder, telling it things which are obvious at the time, but will not be obvious later, to people who were not at the conference.

It's not usually feasible to videotape the small-group sessions. Multiple video cameras and operators would be needed, and in these sessions people usually sit in a circle, so there's no way to see all of them at once. However, still photography is effective - a photographer can roam around capturing important moments, or recording something on request. Flashes should never be used - they're too distracting. Instead, the camera should have a large lens and use high-speed film.

A good record of a conference - whether on videotape, photos, or audio tape, can help to convince staff who were not present of the value of the conference and its findings. They will then be more willing to implement the plans made at the conference.

Co-discovery conference procedures

The conference process falls into four main parts:

1. Introduction (about 1 hour)
2. Research (3 hours)
3. Planning (3 hours)
4. Closing, and looking ahead (1 hour).

Within each part, there's a mixture of individual work, small-group sessions, and plenary sessions.

1. Introduction

The first stage has four main parts, beginning with a general introduction. After that, participants discuss their hopes and expectations for the conference, the two groups (staff and audience) give their impressions of each other as a group, and some personal accounts are selected for presentation.

1a. Explanation

The moderator explains the concept of a co-discovery conference to the participants. The principles stressed are:

Participants are then asked if they have any questions. If there aren't a lot of questions, this stage usually takes about ten minutes,

1b. Recording expectations

Everybody is asked to write down the main outcomes they hope to get from the conference. These are collected, and read out by the moderator, who comments on the feasibility of achieving the various expectations. The secretary writes all the different expectations on a wall chart, crossing out those beyond the scope of the conference. All this takes about 15 minutes.

1c. Impressions of the Other

For the staff members, the Other is the audience. For the audience members, the Other is the staff. This exercise is designed to break down the barriers between the two, and reveal any misconceptions that each group has about the other.

Participants are divided into about 5 groups of 6 people each: 2 groups for the staff (senior and junior, or managers and program-makers), and 3 for the audience. Depending on the purpose of the co-discovery conference the audience groups can be

The split is based on whatever is likely to produce the most similarity of people within each group - and thus the largest difference between the three groups.

The staff groups are asked to describe a typical audience member. The audience groups are asked to describe a typical staff member. Each group is asked to nominate a secretary (somebody who can write quickly and neatly), and a presenter (who is willing to present that group's findings to all participants). It's fun to show the Other as a cartoon diagram - including a cross-section plan of the head, revealing what they supposedly think about.

I usually give the example of a cat, by drawing a circle with ears and a nose: a diagram of a cat's head, as seen from above. Inside the circle, a large area is set aside for food, another large area for sleep, and a small area for chasing mice. Audience groups are asked to make a similar diagram of a typical staff member's head - and vice versa. About 15 minutes is allowed for each group to produce a diagram of what they think occupies the mind of the Other (staff or audience).

Inside a cat's head

The presenter for each group then displays the picture, and describes the group members' perception of the Other type of participant. The Other group is then asked to vote on the reality of this perception. The moderator can say "Please put up your hand if you know anybody like this." Usually, nobody does. These stereotypes will create a lot of laughter - which serves to weld all the participants together, and helps them to realize that the Other - who are of course right there in the room - are really not so different after all. Many staff members hold assumptions about their audience, but don't realize this, and can't state them when asked. Presenting labelled cartoons in this format often helps them to let go of these assumptions, and to realize that the audience is more varied and better informed than they expected.

1d. Personal stories

A valuable part of the co-discovery conference, specially for staff, is how audience members use their service. Audience members can also understand the organization's constraints better by knowing the work patterns of the staff. So each group is now split into smaller groups, of 3 (or maybe 4) people, with at least 2 audience members and 1 staff member. Within each smaller group, each audience member is asked to give the others an account of a recent and not atypical experience with the organization. This must be a specific story, not a general one: not "on weekdays I usually wake up early and..." but more like "last Tuesday I woke up at 5 a.m and..."

Within each small group, the members are now (not earlier) asked to choose one story - the most interesting and/or the best told. The teller of that story is asked to repeat it to all participants. The story is recorded on tape, and during the next session the secretary transcribes it onto a large sheet of paper and puts it on the wall. With 30 participants, there will be 10 stories. When read together, they will give everybody a varied picture of the way the station's output is used by its audience. The assumptions about the Other, ridiculed in the cartoon session, are now replaced by a wide range of genuine stories.

2. Research

Now that the mental stereotypes have been swept away, the research part of the co-discovery conference can begin. The small groups are re-formed. From this point on, the audience and the staff are always mixed together, further emphasizing the unity of the conference.

The next step is to hold consensus groups: three at once. With a total of 30 participants, there will be 10 in each consensus group: about 4 staff members and 6 audience members. The focus now is on the audience experience, and staff members will respond not in their staff role but in their private audience role.

To run three consensus groups at once, you need three people who know how to moderate and three who can be secretaries. If you don't have people trained in these skills, you will need to organize consensus-group training sessions before the co-discovery conference.

2a. First set of consensus groups: audience point of view

Depending on the purpose of the co-discovery conference, the three consensus groups can either discuss the same topic (something to do with the organization as a whole), or three different topics, or three different approaches to an overall topic. We've found the latter to be best, bearing in mind that the main purpose of the research stage is to help staff and audience find out more about each others' interests and capabilities. For example, the three groups can be asked to discuss:

1. What are the best aspects of the organization?

2. What are the worst aspects of the organization?

3. What aspects of the organization most need improving?

As usual with a consensus group, these groups begin with a self-introduction by each participant.

After about an hour and a half, the consensus groups are asked to compile their findings. From each group, a presenter (who is a member of the group, not necessarily a moderator) presents the findings, which will be displayed on large sheets of paper, and sticks these on the wall. As the staff and audience have been about equally mixed in the three groups, the three sets of findings should be fairly consistent.

For each group, any findings on which there was a clear disagreement between staff and audience should be noted, and discussed, to make sure that the disagreement is real, and not just a matter of the vocabulary used.

2b. Second set of consensus groups: producers' point of view

The three groups are now re-formed, with different people in each, but still about 40% staff and 60% audience in each group. As the previous research exercise saw the organization's output from the audience's point of view, this session will review the output from a producer's point of view. Just as, last time, the staff were asked to take on a role as audience members, this time the audience are asked to put themselves in the position of staff members, choosing and evaluating programs. Don't worry: they can do it! No technical knowledge is needed here - and for many organizations, the audience have longer experience than the staff.

Again, three consensus groups are done. They will be faster this time, because everybody now knows what is expected. Introductions are repeated (because the mix of people is different this time). The broad questions to be considered by the consensus groups are similar to those considered in the previous session, e.g.

1. Which of the existing programs should be kept?

2. Which of the existing programs should be dropped?

3. What new programs should be created?

As before, the agreed findings are written on large sheets of paper, and an elected presenter describes these findings to all participants after the consensus groups have finished

Usually there will be a lot of similarity between the audience-oriented and the staff-oriented findings. Any differences between the two can now be discussed in the full groups, with the moderator seeking reasons for the differences. These reasons are also written on large sheets of paper, and put on the wall. As before, clear disagreements between staff and audience should be discussed further, to clarify the exact area of disagreement.

By the end of this second stage, when staff members have thought in an audience role (and vice versa) both sets of participants should have a much clearer understanding of each other. It's time to move on to the second half of the co-discovery conference, which is the planning stage. If feasible, there should be a gap here, to give everybody time to reflect on the findings. The ideal is to leave it overnight, if this was an afternoon session. Next best is to continue the next day, after a morning session. Third best is to break for lunch, and continue soon afterwards.

3. Planning

The planning stage consists of 3 sessions, each lasting about an hour. In this, ideas are generated, discussed, and put forward as tentative plans.

3a. Generating ideas

Now that the audience and the staff have a better understanding of one another, they have a clearer view of what is acceptable and what is possible. So they are ready to begin generating useful ideas. The technique of brainstorming can be used at this stage.

The purpose of brainstorming is for people to produce as many ideas as possible in a short time. All ideas suggested are written up by the secretary, and no criticism is allowed at this stage (it discourages people from producing more ideas). People call out ideas as they think of them. Sometimes the ideas come so quickly that two secretaries are needed - each one recording every second idea. The flow of ideas usually begins slowly, and builds up to a maximum in about 15 minutes. After about 30 minutes, people begin to run out of original ideas, and it is now time for the moderator to stimulate more. For example, participants can be invited to combine two existing ideas, or to build on another idea that they don't quite agree with. No criticism is allowed - but extension of ideas is welcomed, even if the extension completely changes the original idea.

Research has found the ideal group size for brainstorming to be about 12 people, so if there are 30 participants, they can be divided into 2 or 3 groups. Usually, 2 groups will be most convenient - one on each side of the large room. In each group, audience and staff should be well mixed.

The secretary should write each idea on a separate piece of paper. A4 size is good, and pages can be stuck on the wall in advance. Alternatively, the secretary can sit down to write the ideas, and another person can stick them on the wall.

After about 45 minutes, the brainstorming sessions can finish. People are usually exhausted by this, and will welcome a short break. During the break, the secretary and moderator take down all the ideas from the walls, and rearrange them, so that similar ideas are together. If two brainstorming groups came up with the same idea, one page can be stuck over the top of the other.

3b. Adding detail to the ideas

When the participants come back from their break, they'll find the ideas laid out logically on the wall. Some will have had new ideas. Each of these can be written on a piece of paper, and put on the wall in an appropriate place.

New groups of participants are now formed. The ideas are divided into a number of topic areas which is about the square root of the number of participants. For example, if there are 30 participants, the square root of 30 is about 5.5, so there should be 5 or 6 people in each group, specializing in a particular topic. If there are less than about 4 people in one group, or more than about 7, the topics may need to be divided, combined, or redistributed. With all the ideas laid out logically on the wall, it is not difficult to divide the topics fairly evenly.

Each group now takes its ideas down from the wall, and prioritizes them. A good way to do this is to sort the ideas into 4 heaps:

1. Ideas agreed by most people to be good ones, and worth following.
2. Ideas agreed to be bad, not worth following.
3. Ideas that some participants say are bad and others say are good.
4. Ideas whose meaning is not clear to everybody.

When all the ideas have been sorted, heap 4 is considered first. Each idea in this heap is briefly discussed, and its meaning clarified, perhaps by adding to it. New ideas can still be added at this stage. Heap 4 is then redistributed into heaps 1, 2, and 3, and each of these heaps is considered in turn. Near-duplicated ideas can be combined at this stage.

The good ideas in heap 1 should be elaborated a little more, spending a few minutes on each to describe how it could be used. The reasons for choosing each idea should also be written up. This can be helpful if an idea is later found to be not feasible; if the reasons for it are known, some other idea could take its place.

As for heap 2, the bad ideas, participants should decide why the idea is bad. Too expensive to implement? Already being done? Whatever the reason, this should be written on that page.

With heap 3, the nature of the disagreement should be explored. Why are people disagreeing? In particular, are the staff disagreeing with the audience? Is it possible to split the disagreed item into two: one part agreed to be good and the part bad? The aim is to reduce heap 3 to items on which there is clear disagreement.

If disagreement can't be resolved, it should be maximized. Can the statement be changed so that half the group are strongly in favour of it, and the other half strongly against? This will help focus the disagreement.

The danger is spending too much time on the disagreed items, so somebody should time this. If an issue can't be resolved (by only about 6 people) in 3 minutes, it probably never will be, so one member of the group should act as timekeeper. When the 3 minutes is up, and no agreement has been reached, the statement should go back on heap 3.

Most time should now be devoted to heap 1: the ideas that most people think are good.

The time taken by this stage will vary greatly, but should be no more than about 90 minutes. If it's important to finish on time, the moderator will need to work out which groups are slowest, and hurry them along.

3c. Ideas Fair

In the final planning stage, the ideas set out previously are now explained to all other participants. This is done by re-forming the groups, with the same number of groups as before, but each new group has one member from each of the previous groups. Each previous group has left its papers in the area it worked in. Each new group considers each set of papers in turn. The member who was in the previous group explains the findings, and answers the others' questions about them. If no member of the current group was in the group that produced these papers, this group will need to borrow somebody from another group. The controversial ideas in heap 3 are voted on by anybody who is interested - e.g. by putting a tick if they agree and a cross if they disagree.

The rotation in the "ideas fair" ensures that every participant will have a chance to have every idea explained. This stage helps to make the reason for each idea very clear.

3d. Setting priorities

This is a plenary session, in which everybody meets to discuss the ideas they have now all seen. The only ideas put forward are heap 1 (with general approval) and those from heap 3 (controversial) on which ticks outnumber crosses by at least 2 to 1.

The main purpose of this session is to prioritize the ideas, on two key dimensions:

1. How easy is it to implement: how quick, how cheap, and how well does it in with the organization's priorities? On this dimension, the staff will have more to say.

2. How valuable or important is it? Hardly at all, moderately, or vitally important? This dimension will be more important to the audience.

Also at this stage, if any ideas (considered by separate groups) are practically identical, they can be combined.

Now it's time to vote on the ideas. An easy way to do this is to give each participant two sets of sticky dots, with one colour for "how easy" and another for "how valuable." Staff and audience members can have different colours. With each idea on the wall, participants can stick on it as many dots as they like. The ideas with the most dots are the winners, and put forward for further consideration by the organization's staff.

4. Final stage

Now we are near the end. The final stage usually takes less than an hour.

4a. Discussion of expectations

The very first session of the co-discovery conference was when everybody wrote down their expectations. Now is the time (before it is too late) to discuss how far those expectations have been fulfilled. The moderator will have said, after the first session, whether some expectations were not achievable, but the conference should try to meet all other expectations. As a list of expectations already exists, and will still be up on the wall, the moderator goes through it - ignoring any expectations that were crossed out because they were clearly not achievable.

The people who gave each remaining expectation are now asked whether it has been met. If not, and if most participants regard it as important, something will need to be done about it - for example, organizing a brief small-group session to discuss that topic. This may take about half an hour - but if all feasible expectations have been met, it will take only a few minutes.

4b. Ending

The staff participants will meet again, back at work, so there's no need to discuss follow-up procedures in detail. However, the audience members, who do not have continuous contact with the organization, will want to know what happens next. At this point, a senior manager from the organization may arrive, thank them, and give them some assurance that they have not wasted their time. A follow-up team should be announced, to make sure that everything planned at the conference will really happen.

That's the end of the conference. By this time, the participants are usually enthusiastic, even excited. It can come as a shock for the moderator to declare "you can all go home now," because many participants don't want to finish. So it usually seems best to end with an informal break, or a meal.


Usually a report will be written, summarizing the co-discovery conference. It will be easy to put together, because it will consist mostly of the papers that have been stuck on the wall. It's a good idea to include some photos in the report, too. When participants read the report later, the photos will remind them of the process and what was agreed; photos also help staff members who couldn't attend the co-discovery conference understand the proceedings. If the organization is committed to audience participation, it should send all the audience participants a copy of the report. If it's a media organization, it can publish the report - even if only on its web site.

If there's a videotape or audio report, this is done separately from the written report, but if photos are used instead of video or audio, these should not be in a separate report - to make sense, they need to be labelled, and in context.


The big danger with a co-discovery conference is that, at the end of the day, everybody is clear about what should happen, but nobody ever gets around to doing it. All that effort is wasted. That's why it's important to assign an effective team to follow up the conference: staff who are important enough that everybody respects them, but not so busy that they have no time to make sure that the conference plans are carried out.


In a single day's work, the staff involved will have gained both a good understanding of their audience, and a clear set of prioritized plans to work on. I know of no other way of achieving all this, in so short a time.