Speaker plans to deliver the submitted conference paper as written
A paper written for silent reading and a paper written for a live room are two different documents, and treating them as one is the single most common cause of an overrun. Academic prose is built from subordinate clauses, citations, and qualifying phrases that a reader can pause over; a listener cannot pause a speaker, so the same density that works on the page becomes incomprehensible read aloud at speed.
The practical consequence is timing. A paper of around 3,000 words, read at a natural pace, takes roughly 20 minutes — but conference slots are usually set at 10, 12, 15, or 20 minutes including a chair's introduction, which leaves far less room than the full text requires. When a speaker tries to compress the whole paper into that window, the talk runs long, the chair gives the time warning, and the conclusions — the part the audience most needs — get cut or rushed.
The fix is to treat the spoken talk as a separate text built from the paper rather than a trimmed copy of it:
- Identify the two or three findings the audience must leave with, and drop everything that does not support them.
- Cut the literature review to a sentence or two of context — the audience came for the new result, not the background chapter.
- Replace methodological detail with the minimum needed to judge the result's credibility; put the rest in slides or a handout for those who ask.
- Write the spoken version in short, declarative sentences, the way the point would be said in conversation, not the way it was written for a journal.
- Read the script aloud against a timer before the session and cut again if it runs over — not by speaking faster, but by removing content.
This mirrors the structure most guidance on giving a conference presentation recommends: treat the talk as a distinct communicative act from the paper, built around a small number of take-home points rather than full coverage.
Slides are built by pasting paragraphs and tables from the paper
Slides copied straight from the paper fail because the audience cannot read dense text and listen to the speaker at the same time — they will do one or the other, and usually choose reading, which means the spoken explanation is wasted. A table built for a print page, with eight columns and footnoted sample sizes, is unreadable from the back of a conference room regardless of font size.
The corrective is a stated density limit rather than a vague instruction to "simplify." The 5/5/5 rule is one such guideline: no more than five words per line, five lines per slide, and five slides in a row that are mostly text, forcing the presenter to convert prose into a small number of short visual cues rather than a transcript. Guides on building conference presentation slides describe the same underlying discipline — one point per slide, supported by an image, chart, or single statistic, with the full explanation delivered verbally rather than printed.
Applied to a results table, this means picking the one or two comparisons the argument depends on and showing those as a simple chart or highlighted figure, while the full table stays in the paper or an appendix slide held back for Q&A. Applied to a methods section, it means a single diagram rather than a paragraph of procedural text. The slide's job is to anchor what the speaker is saying, not to replace it — a distinction Enago's overview of presentation types also makes in separating oral papers from poster formats, where the poster is expected to carry more standalone text precisely because no one is speaking alongside it.
A timing check for slide count
A rough planning figure is one slide per one to two minutes of talk, excluding the title and acknowledgements slides — so a 12-minute talk supports somewhere around 6 to 10 content slides, not the 20 or 30 that a full paper's section headings might suggest. Counting slides this way before the talk is drafted catches an overrun risk before rehearsal does.
Call for papers and abstract acceptance place the paper in a session
The session format is set before the talk is written, and it determines the time limit, the expected slide density, and the length of Q&A — so the format should be confirmed, not assumed, as soon as the abstract is accepted. A call for papers typically specifies whether a submission has been accepted for an oral paper session, a shorter "lightning" or pecha-kucha-style slot, a poster session, or a panel contribution, and each of those carries different constraints on what the speaker can attempt.
| Format | Typical time allowance | Audience expectation |
|---|---|---|
| Oral paper in a parallel session | Short, fixed slot plus brief Q&A | Linear argument, one or two findings, minimal methods detail |
| Plenary or keynote paper | Longer, sometimes with no hard Q&A cutoff | More context, more development of implications |
| Poster session | No scheduled speaking time; conversation-driven | Visual self-explanation, detail available on request |
| Panel contribution | Very short framing statement, most time is discussion | A position or provocation, not a full paper |
This is drawn from the format distinctions commonly made in guidance on types of conference presentation, which separates oral, poster, and panel contributions specifically because each demands a different balance of visual material, spoken explanation, and audience interaction.
Because the acceptance email or the conference programme usually states the exact minute allowance for the session, that number — not a generic assumption about "a conference talk" — should be the first constraint written down before drafting begins. A paper accepted into a 15-minute parallel session with five other speakers and a shared Q&A block needs a far tighter script than the same paper accepted into a 45-minute invited session. Checking the format also surfaces whether questions are taken individually after each talk or batched at the end of the session, which changes how much the speaker needs to anticipate in the talk itself versus hold back for the discussion.
Audience
The audience is whoever is in the room, and their prior knowledge — not the speaker's — sets what the talk can responsibly cover. A session audience at a disciplinary conference may share the speaker's technical vocabulary; an audience at a multidisciplinary conference, a public engagement session, or a mixed academic-industry event usually does not, and a talk that assumes shared jargon loses that audience in the first two minutes.
Pitching to the audience means making a deliberate choice, before the talk is written, about which register it is written in:
- Specialist audience in the same subfield — technical terms and field-standard abbreviations can be used without definition; the talk can spend its time on what is new rather than what is established.
- Mixed academic audience across subfields — key terms need a one-sentence gloss the first time they appear; the motivation for the work needs more space than it would for a specialist crowd.
- General or public audience — jargon is minimised almost entirely, analogies replace technical description, and the "why does this matter" framing moves to the opening rather than the conclusion.
Advice on presenting at discipline-specific conferences, such as the British Society for the History of Science's guidance for early-career presenters, stresses that even within one discipline a session audience often includes people outside the speaker's immediate specialism, which argues for erring toward slightly more explanation rather than less. Checking the session's programme description in advance — many conferences group sessions by theme rather than strictly by subfield — gives a reasonable estimate of who will actually be in the room.
Conference paper presentation
A conference paper presentation is the oral delivery of research to an audience within a scheduled session, distinct from both a lecture and a poster in what it assumes about the audience and what it demands of the speaker. Unlike a lecture, which typically has one speaker addressing students over an extended period with room to build an argument slowly, a conference presentation is short, judged by peers rather than students, and usually followed immediately by other presentations covering unrelated work — so the audience's attention has to be captured and the argument delivered within a narrow, fixed window.
Unlike a poster, which stays on display for an extended period and is read and discussed one-to-one or in small groups at the presenter's own pace, an oral presentation happens once, to everyone in the room simultaneously, with no opportunity to go back and re-explain a missed point. This is why pacing, slide density, and a rehearsed script matter so much more for the oral format than for a poster, where the visual document itself has to do more of the explanatory work because the presenter cannot assume continuous attention from every person who stops to look. The distinction is set out in overviews of presentation formats used at academic conferences, which separate oral papers, posters, and panels by exactly this difference in audience interaction.
A practical primer on the genre, intended for researchers presenting for the first time, frames the oral conference paper as a performance built for listening rather than a recitation of the written paper — a framing echoed in guidance aimed specifically at early-career researchers preparing their first talk, which recommends treating the spoken delivery, the slides, and the handling of questions as three separate skills to prepare rather than one.
Conference paper (written text)
The written paper is the source material for the talk, not the talk itself, and the two documents are expected to diverge once the oral version is drafted. The paper exists to be read closely, cited precisely, and checked by reviewers against its methodological claims; it is written with full detail because a reader can reread a difficult paragraph, follow a footnote, or check a reference at their own pace.
The talk draws on that paper's argument and evidence but is built as a separate spoken text, because none of the advantages a reader has — rereading, footnotes, pausing — are available to a listener. This is also why the submitted abstract matters beyond its role in getting the paper accepted: the abstract is usually the clearest existing summary of the paper's core contribution, and rewriting the spoken introduction around the abstract's framing, rather than the paper's own introduction, often produces a tighter opening than starting from the paper's literature review.
Keeping the two documents distinct also protects the written paper from compromise. A paper that has been cut and simplified to fit twelve minutes of speaking time is a weaker paper if someone later asks to read it — so the cutting and simplifying should happen in the speaking script, with the full paper left intact as the archival record of the work, available to anyone who wants the complete methods and results after the talk.
Conference session
A conference session is the scheduled block — usually 60 to 90 minutes — in which several presenters speak in sequence under a chair, and the session's total length divided by the number of speakers is what actually produces the individual time limit, not a conference-wide standard. A session with five accepted papers in a 90-minute block, minus time for the chair's introductions and transitions between speakers, might leave genuinely little more than 10 to 12 minutes per paper including questions — considerably tighter than the headline "20-minute slot" a programme sometimes advertises in the abstract.
The chair's role in that session is to enforce the shared time budget fairly across all speakers, which is why chairs give time warnings and, if necessary, cut a speaker off before the conclusion: a single speaker running long does not just cost that speaker their own closing remarks, it eats into every subsequent speaker's slot and the session's Q&A. Knowing in advance how many papers share the session, and asking the chair or the programme for the exact per-speaker allowance rather than assuming the published session length divides evenly, removes most of the uncertainty that leads to a surprise cutoff.
Sessions run differently depending on the conference's delivery mode, and preparation should account for the mode as well as the time limit:
- In-person sessions rely on the room's own screen and clicker, slide transitions are near-instant, and the chair's time signals are usually visual (a hand, a card) rather than verbal.
- Virtual sessions typically run through screen-share, which can introduce a visible lag between slide changes and the speaker's voice, so pacing should allow a beat longer per slide than an in-person talk would need.
- Hybrid sessions combine both, and the speaker should confirm beforehand whether questions from the remote audience are read aloud by a moderator or taken directly, since that changes how Q&A time is allocated.
Time limit
The time limit is the fixed speaking window for the talk, and it is the constraint every other preparation decision should be checked against before the talk is finished, not after. Working from a stated limit, a usable rule of thumb is roughly 100 to 130 spoken words per minute for an academic talk delivered at a measured, audience-friendly pace — faster than that and comprehension drops, particularly for an audience listening in a second language or encountering the topic for the first time.
That gives a rough word-count target for the spoken script:
| Time limit (including Q&A buffer) | Approximate spoken-script word count | Approximate content slides |
|---|---|---|
| 10 minutes | 1,000–1,200 words | 5–7 |
| 15 minutes | 1,500–1,800 words | 7–10 |
| 20 minutes | 2,000–2,400 words | 10–14 |
These figures are a planning starting point, not a fixed formula — pacing varies by speaker, by how data-heavy the slides are, and by how much pause time the talk needs for audience comprehension, a point made in guidance on preparing an academic talk for a live audience, which recommends rehearsing with a timer rather than relying on an estimated reading speed alone.
Q&A preparation belongs inside the time limit, not after it. Before the session, it is worth listing the two or three questions most likely to be asked — about sample size, a methodological choice, or how the finding compares to a specific prior study — and drafting a short, direct answer to each rather than improvising. For a question with no good answer (a genuine limitation, a result that is not yet replicated, a gap in the data), the stronger response is a brief, honest acknowledgement of the limitation followed by what would be needed to resolve it, rather than an attempt to argue the gap away; reviewers and audience members generally read a direct admission as more credible than deflection.
Confirm the exact time limit, the Q&A format, and whether the session is in-person, hybrid, or fully virtual as soon as the acceptance notice arrives, and build the spoken script and slide deck against those numbers from the first draft rather than retrofitting them the night before.