How to produce great think tank events

Conference season is upon us. Let’s make it worth people’s time.

Conference season is beginning. The UN General Assembly, the World Bank–IMF Annual Meetings and COP will bring thousands of people together—and inspire many think tanks to organise events around them. Participants will travel across countries and continents, navigate crowded schedules and invest considerable time and money to attend.

Even a local event asks something significant of people. It takes them away from their work, families and other commitments. 

And think tanks are organising more events than ever. In-person, online and hybrid, think tank events are at the centre of their communication strategies. 

Every invitation is therefore asking someone to give us one of their scarcest resources: their attention.

We are very conscious of this at On Think Tanks. Anyone who comes to an OTT Conference, joins the School for Thinktankers, or attends any of our local and online events should feel they received more than they put in. They should leave with useful ideas, honest conversations, new relationships and something they can apply. Most importantly, we want them to want to come back.

Todd Moss’s recent article on think tank events makes several very good points. Too many events are organised because think tanks traditionally hold events—not because anyone has decided what the gathering should accomplish or why people should attend.

This is OTT’s take on what makes a think tank event genuinely worth people’s time.

We have tried to apply these principles to our own gatherings. OTT Conferences combine participant-led sessions, practical workshops and candid conversations with long breaks, shared meals, fringe activities—and formats such as Unpopular Opinions and F@ck-up Night. We design the formal and informal parts together because the conversation over lunch or on the walk back from dinner may prove as valuable as anything on the published programme.

And we are genuinely pleased when other organisations copy our formats. Good ideas for convening should travel. We would much rather see an OTT experiment adapted, improved and used elsewhere than guarded as institutional intellectual property.

If you would like to experience—and help shape—this approach, the 2027 OTT Conference will take place in Barcelona on 12–13 May, with fringe events on 11 and 14 May. Its theme is Think tanks and People, but, as always, the programme will be shaped by the questions and challenges participants bring.

So, what have we learnt? Our central advice is simple: do not begin with the event format; begin with the change you want the event to produce.

A great think tank event is not simply a well-run panel, a successful launch or a large audience. It is a carefully designed intervention that helps the right people encounter useful evidence, test ideas, build relationships and decide what to do next.

The 12 principles below are an update on advice we developed in 2014: produce an event, don’t just organise one. They guide our own event design, as well as the advice we offer think tanks and funders on organising theirs. Please contact us if you would like to know how we can help you.

Twelve principles for great think tank events

1. Give the event a job

Be precise about its purpose. Is it intended to:

• inform a particular policy debate;
• test or improve an argument;
• build relationships with decision-makers;
• connect people who would not normally meet;
• develop a shared agenda;
• increase the visibility of an issue or organisation;
• generate commitments or initiate collaboration?

Trying to achieve all these things in one event usually produces a generic discussion – and confusion among participants! Decide what the event’s primary job is and design everything around it.

2. Design for a specific audience

“The public” and “stakeholders” are not useful audience definitions. Identify who needs to participate and why:

• policymakers with authority;
• officials with technical or operational knowledge;
• political advisers and intermediaries;
• researchers with relevant evidence;
• journalists and communicators;
• funders, practitioners or affected communities.

Invitations should explain why that person’s participation matters—not merely announce the topic and speakers.

But be careful about who is in the room. Diversity has lots of benefits but also makes it much harder to manage. Having policymakers, journalists, funders and thinktankers in the same room may look good on paper, but unless you have the resources and skills to create the right connections between them, you’ll find that the conversation between these groups will not flow – and may lead to mistrust and conflict. 

3. Treat participants as contributors, not spectators

The value of a think tank event often lies in the knowledge distributed around the room. Participants need opportunities to question, compare, challenge and contribute.

This means fewer presentations and more:

• facilitated conversations;
• small-group exchanges;
• interviews and fireside conversations;
• case clinics;
• structured debates;
• problem-solving exercises;
• opportunities for informal interaction.

OTT’s experience with the School for Thinktankers reinforces this: participants frequently report that learning from one another is as valuable as the formal content.

4. Start with a good question—not a broad topic

“Discussing the future of democracy” is not a useful event proposition. A stronger framing might be:

“What should independent policy organisations do when evidence is valued but politically inconvenient?”

A productive question introduces a choice, tension or unresolved problem. It should be narrow enough to generate a meaningful conversation but open enough to allow disagreement.

A specific question also helps speakers focus and makes their job much easier.

5. Choose speakers for the conversation they can create

Reputation alone does not make someone a good speaker. Look for people who bring:

• different forms of knowledge;
• contrasting positions or experiences;
• candour and an ability to engage;
• geographic, institutional and generational diversity;
• genuine relevance to the central question.

Avoid panels made up entirely of celebrity intellectuals or organisational leaders delivering polished institutional statements. The person closest to implementation may contribute more than the most senior person.

Think tanks often spend all their resources securing celebrity speakers and pay little attention to the real contribution they will make to the discussion. 

6. Do not be afraid to tell speakers what you want them to say

Inviting someone to speak should not mean giving them an empty slot and hoping for the best. Tell speakers clearly what contribution you need: the argument, experience, example or provocation they can uniquely bring—and what the other speakers will cover.

Develop a detailed script or run-of-show, including the questions, transitions, timings and planned moments of interaction. This is not about putting words into speakers’ mouths or suppressing disagreement. It is about producing an event in which individual contributions combine into a coherent and engaging experience for the audience.

Good events may feel spontaneous, but they are usually carefully produced.

7. Give the moderator real authority

Moderation is an editorial role, not a ceremonial one. The moderator should:

• understand the issue;
• agree on the purpose with speakers beforehand;
• prevent long opening statements;
• identify areas of disagreement;
• ask follow-up questions;
• bring participants into the conversation;
• synthesise what is emerging;
• finish with implications rather than generic thanks.

A useful rule is: the moderator represents the audience, not the speakers. Chatham House reportedly trains its staff on how to moderate events!

8. Replace presentations with preparation

Ask participants to read a concise briefing, examine a chart, respond to a proposition or consider two questions before arriving. Do not use the event to reproduce a report chapter by chapter.

If you are planning to present a book or a report, share it with your audience in advance. Allow them to digest the content, reflect on it and prepare questions or comments. These will be much better than the questions or comments they will utter on the spot. 

The event should add something the publication cannot provide: interpretation, challenge, connections, decisions or collective intelligence.

If a presentation is necessary, make it short and purposeful. Its job is to give participants what they need for the conversation—not to demonstrate everything the researchers know.

9. Design the room—and the rules—for trust

Some conversations require publicity; others require confidentiality. Decide explicitly whether the meeting is:

• public and on the record;
• attributable;
• held under the Chatham House Rule;
• private and non-attributable.

Smaller, carefully curated spaces can generate more influence than large public events. This is particularly important in polarised or restricted political environments, where participants need confidence that disagreement will not expose them or their organisations.

The physical or virtual setting matters too. Seating, timing, language, technology and accessibility all affect who speaks, who listens and whose knowledge is taken seriously.

10. Plan the event as part of a longer journey

A good event has a “before” and an “after”.

Events should sit within a broader communications, engagement or policy-influence strategy—not operate as isolated activities.

11. Use the event to generate content—and evidence

An event should not only communicate work that has already been completed. It can generate new knowledge, evidence and material for future work.

Plan from the beginning to capture what the event produces through:

• keynote and session videos;
• short interviews with speakers and participants;
• session summaries and synthesis articles;
• blogs and opinion pieces;
• podcast episodes;
• photographs, quotations and short social-media clips;
• new case studies, examples and research questions;
• insights that feed into ongoing studies.

Following the 2026 OTT Conference, we produced five keynote videos, 18 parallel-session summaries, more than a dozen blogs, interviews, podcast episodes and other outputs. Several sessions also contributed directly to ongoing OTT studies.

This does not mean recording everything mechanically. Decide what is worth capturing, secure participants’ consent and assign people to document the discussion properly. A transcript is not yet a useful output: it needs editorial judgement, synthesis and follow-up.

Events are valuable opportunities to generate evidence—not simply share it. Remember: the knowledge is in the room!

12. Judge success by what happens next

Attendance, social-media impressions and speaker seniority are useful operational measures, but they are weak measures of value. Better questions include:

• Did the right people participate?
• Did they engage seriously with the evidence?
• Did anyone change or refine their position?
• Were new relationships established?
• Did participants identify decisions, experiments or collaborations?
• Was the organisation invited into subsequent conversations?
• Did the event reveal something that should change the research?
• Did anything happen afterwards that would not otherwise have happened?

The practical OTT test

And one final, slightly provocative principle: not every publication needs a launch, not every discussion needs a panel, and not every important conversation should be public.

The best format is the one that serves the purpose—not the one think tanks are most accustomed to organising.

So before approving an event, ask these two questions:

1. Why this event, for these people, at this moment—and what should be different afterwards?
2. Can we deliver this result in a more cost-effective way?