Finding, choosing and looking after speakers

A call for papers (CFP) is the open invitation you publish asking people to submit a talk for consideration. Run one properly and it does two things at once: it surfaces people your existing network would never have found, and it gives you a record for arguing, later, that the programme was chosen rather than assembled from favours.

What to ask for in the submission

Ask for a title, an abstract of around 200 to 300 words, the format the talk needs (a straight talk, a panel, a workshop) and the session length it is written for. Ask what the audience will be able to do differently after the talk. A submission that answers that question is easier to judge fairly. Ask for a short speaker bio separately from the abstract, and consider collecting it after the first round of review, because judging the idea before the name is what makes anonymised review possible later on.

Describing the topic without writing it for the usual answer

A call that names a specific technology, job title or well-known problem draws submissions from people already fluent in that vocabulary, which tends to be the same circuit of speakers who submit everywhere. Describe the outcome you want the audience to leave with, and leave room for an answer that comes from a discipline you had not considered. A call asking for talks on API design gets API specialists. A call asking for talks on the decisions that make a system hard to change six months later gets API specialists too, but also product managers, support engineers and people who have only ever worked on one large legacy system.

Where to put the call

A CFP posted only to your existing mailing list and social accounts reaches people who already follow your event, a smaller and more uniform group than the field of possible speakers. Post it to community forums, professional association mailing lists and alumni networks outside your immediate sector, and ask two or three people from underrepresented parts of your field to forward it directly. A personal forward converts into a submission far more often than a public post does. Give six to eight weeks between opening the call and the deadline; a shorter window favours people who already have a deck ready to reuse.

Setting out how proposals will be judged, and how speakers will be treated

State the review criteria and the code of conduct in the call itself. A first-time speaker deciding whether to submit is weighing the risk of the room against the value of speaking, and knowing in advance what the panel is actually scoring, and how the event handles harassment, changes that calculation. Cover fees and expenses in the call too, even in broad terms; it removes any reason for a speaker to assume the platform is the whole payment.

A call for papers (CFP) produces a stack of proposals, and the way you read them decides who ends up on stage as much as the proposals themselves do. A programme committee that reads names before abstracts tends to select for who they already know, whatever criteria they think they are applying. The fix most conferences reach for is anonymised review: strip the speaker's name, employer and any identifying detail from the proposal before it goes to reviewers, and score the content on its own terms.

What anonymised review actually removes

Anonymising a proposal means taking out the speaker's name, job title, company and any biographical line, along with references that give the game away, such as "following on from my talk last year" or a company product name used as the example throughout. What stays is the title, the abstract, the intended audience and the takeaway. Reviewers score that. Names go back on only once scoring is finished, at the point where the committee is deciding the running order and checking for balance across the programme.

Anonymised review will not catch everything. A well-known figure's writing style or a distinctive project is sometimes recognisable regardless, and a small field means fewer plausible authors to guess from. It still removes the most common bias, which is a reviewer's unconscious weighting toward familiar names, and it is worth doing even where it is imperfect.

Criteria that actually discriminate between proposals

A scoring sheet with one column for "quality" produces reviewer disagreement with nothing to resolve it. Useful criteria are specific enough that two reviewers scoring the same proposal land close together. A workable set covers:

• Relevance: does this match what the stated audience for this track actually needs
• Specificity: does the abstract describe an argument, a method or a worked case, or does it promise a topic without saying what will actually be said
• Takeaway: could an attendee describe, the day after, one thing they would do differently having heard it
• Delivery risk: is there anything about the format or content that needs a decision now, such as a demo that depends on the venue's network, or material that needs a sensitivity note

Score each separately rather than folding them into one number early. A proposal that is highly relevant but vaguely written is a different problem, and a different conversation with the speaker, from one that is precise but off-topic.

Conflicts of interest, and why they need a process

A conflict of interest exists whenever a reviewer has a personal or professional stake in a proposal's outcome: their own employer, a close colleague, someone who has previously employed them, or a competitor to a company they work for. The reviewer does not need to act on the conflict for it to matter. The process needs to record it and remove that reviewer's score from that proposal, because a committee that relies on reviewers to self-police in the moment will find that the awkward cases are exactly the ones nobody flags. Ask reviewers to declare conflicts against the full list of submitters before scoring begins.

Safe programme or one that takes a risk

A programme built entirely from proven speakers on familiar topics is lower risk on the day: attendance is predictable, feedback is predictable, nothing goes wrong. It also tends to look, from the audience's seat, like a programme they have half-seen before, because the same reliable people are covering ground they have covered elsewhere. A programme that includes a handful of first-time speakers or genuinely contested topics carries a higher chance that one slot underperforms, against a higher chance that the programme is talked about afterwards for reasons beyond logistics.

The position worth taking here is that a programme with zero risk in it is a design failure, because it has traded the thing a programme is for (the difference between what attendees came in knowing and what they leave knowing) for the appearance of reliability. The practical version is not to gamble the whole programme: put the untested slots in a corridor track (a smaller, parallel session running alongside the main stage) or in a shorter slot where a weak outcome costs the audience ten minutes, and keep the main stage anchored by speakers with a track record. That way the committee can afford to take the risk that makes the programme worth attending, without staking the whole day on it.

How this connects to the rest of the programme, including what a strong CFP asks for and how sessions get shaped once speakers are confirmed, is covered in Designing a Conference Programme.

Most conferences fall into one of two payment models: a flat speaker fee or expenses-only, with the choice depending on the type of speaker and the nature of the event. Trade and professional conferences with a commercial audience typically pay keynote speakers a fee, while community, academic and volunteer-run events more often cover costs only and are explicit about that from the first email. Whichever model applies, it should be stated before the speaker prepares a proposal.

The UK fee range and what moves it

Speaker fees for UK conferences vary widely, and the honest answer is that there is no single going rate: a practitioner speaking on their own expertise at a mid-sized industry event, a well-known external keynote, and a first-time speaker presenting original research sit in very different bands. What moves the fee within any band is the speaker's public profile, whether the talk is bespoke or reused, the size and commercial nature of the audience, and how much preparation the brief requires. A speaker asked to build a new talk for your event is doing more work, and that should be reflected in what you offer.

Why 'exposure' is not payment

Offering a platform in place of a fee only works when the platform genuinely delivers something the speaker cannot get elsewhere, such as access to a specific buyer audience they are trying to reach commercially. For most speakers, 'exposure' costs them a day's preparation and a day's attendance and gives back nothing they can invoice against. If a speaker is taking time away from paid work to appear, name the trade-off honestly in the invitation, so they can decide with the actual terms in front of them.

What to cover regardless of the fee

Even where no fee is paid, reasonable costs directly caused by speaking should normally be covered. That typically means:

• Travel, at a standard reasonably close to what the speaker would use for their own work travel.
• Accommodation, where the event requires an overnight stay, including the night before an early morning slot
• Meals and incidentals for the day of travel and the day of the event
• Any direct costs of preparing the talk that you have specifically asked for, such as commissioned diagrams or accessibility formatting

Set out what is covered and what is not in writing before the speaker accepts, including whether they book and claim or whether you book directly. A speaker who has to chase an expense claim after the event is less likely to speak for you again, and word travels between speakers faster than most organisers expect.

Where a code of conduct and access needs sit in this decision

The fee and expenses conversation is also where a code of conduct and accessibility needs should be raised. Ask directly whether the speaker needs any adjustment to travel, room setup or session format, and confirm what conduct standards apply to speakers and attendees alike. The Equality Act 2010 places a duty on organisers to make reasonable adjustments, and what counts as reasonable depends on the size of the event and the practicability of the change; the guide to accessibility and codes of conduct covers this in more detail.

Pay-to-speak, and why it sits outside this

A pay-to-speak arrangement is where a speaking slot is sold as a sponsorship benefit. Where sponsorship funds a slot, say so on the programme. The section on reviewing proposals fairly, elsewhere on this page, covers how to keep sponsored and merit-selected slots visibly separate.

For the wider question of what a speaker needs to prepare well and what to brief them on before the day, see Speaking at a Conference: Getting Selected and Being Good. For the budget context that a speaker fee sits inside, see Venues, Catering and the Costs Nobody Budgets For.

A speaker briefing is the information you send in advance so a speaker walks in knowing the format, the room, the audience and the boundaries the session is working to. Speakers who ask questions on the day that the briefing should have answered are usually reacting to a briefing that arrived too late or said too little.

What to send, and when

Send the first briefing when you confirm the speaker. It should cover the session format (panel, keynote, workshop), the length of the slot including questions, the audience's likely level of prior knowledge, and the theme or question the programme is built around. A second, shorter briefing two to three weeks out should confirm timing, slide deadlines and logistics such as travel and check-in.

• Format and length of the slot, with the split between talk and questions stated separately
• Audience: numbers, seniority, what they already know
• Slide template and deadline for submitting slides, if slides are reviewed centrally
• Arrival time on the day and where to go first
• Who to contact with questions before the event

The technical brief: what actually goes wrong

Most on-the-day problems are technical and avoidable: a laptop that will not connect to the house AV, a slide aspect ratio that does not match the screen, a clicker the speaker has never used before. State the screen resolution and aspect ratio, whether speakers present from their own laptop or a shared one, and whether there is a lectern microphone, a lapel mic or a handheld. Offer a tech check slot the day before or on arrival, and say so in the briefing.

What the green room is for

The green room is the space set aside for speakers before and after their session, somewhere to wait, run through slides, or simply sit down away from delegates. It works when it is quiet, has water and a power socket, and is close enough to the stage that a speaker is not walking through the main crowd five minutes before going on. A green room with no clear signage, or no member of staff who knows who is on next, causes more anxiety than the talk itself usually would.

The small things that decide whether a talk lands

A speaker who knows how much time is left, without checking their own phone, generally paces a talk better than one who is guessing. Agree a timing signal in advance: a card at five minutes and one at time, held where the speaker can see it without breaking eye contact with the room. Introduce the speaker with the introduction they gave you, and confirm before the session whether they want to take questions standing at the lectern or moving into the room.

Accessibility and conduct, briefed not bolted on

If the event provides captioning (CART) or a British Sign Language interpreter, tell the speaker before the day. It affects pacing, because a speaker who talks over captions or reads slides aloud too fast makes both harder to follow. If a code of conduct covers how attendees may question or approach speakers, send it to speakers as well as delegates, so they know what support is in place if a session or the questions afterwards become difficult. Reviewing accessibility needs against the venue is covered in Accessibility and Codes of Conduct, and the same audit should feed directly into the speaker briefing.

After the session, a short thank-you and a note of anything that went wrong technically is useful if the same speaker is invited back. The programme structure that determines how much room a speaker's slot actually has is covered in Designing a Conference Programme.