What is a lightning talk?

Five minutes, one idea

The defining rule is time, not slides. PyCon US describes its lightning talks as a maximum of five minutes on any topic of interest, and explicitly allows one minute talks too. Within that limit the speaker chooses how many slides to use or whether to use slides at all. Most good lightning talks do one of four things: show a tool, share a lesson learned, make a small argument, or ask the room for help. Anything that needs background, caveats and a literature review is a full talk in disguise. A useful test is whether the idea can be said in one sentence to someone in a corridor. If it can, five minutes is enough to give the sentence, the evidence and the next step. If it cannot, the talk is not ready yet, regardless of how many slides exist.

Where the name comes from

The term is usually traced to the Yet Another Perl Conference in Pittsburgh in 2000, where Mark Jason Dominus organised a session of five minute talks. A report from that conference on perl.com describes the session as a big success, notes that all the speakers kept to their allotted time, and records that Dominus marked the limit with a gong. The format spread through open source and developer events, then to academic conferences, company all hands meetings and meetups, and inspired stricter cousins such as Ignite with its automatically advancing slides.

How to structure one

A reliable shape has four beats. Open with the problem in one sentence the audience recognises. Show the single thing you want them to know, whether a number, a demo or a before and after. Explain just enough of how it works for them to try it. End on what to do next and where to find it. That fits comfortably in five to eight slides, each carrying one idea in large type. Rehearse aloud with a timer at least three times, and cut the part you are most attached to if you run over, since it is usually the background.

Common mistakes

Trying to compress a thirty minute talk into five is the usual failure: the speaker talks faster, skips slides and finishes nothing. The second is spending the first minute introducing yourself and your company, a fifth of the talk gone before the idea arrives. The third is a live demo with no fallback, since there is no time to recover from a failed login. The fourth is ending on a slide of links nobody can copy before the next speaker plugs in. One short address shown for the whole final minute works better than five.

Why events like the format

A block of lightning talks gives first time speakers a low stakes stage, lets an audience sample many topics, and surfaces people worth inviting back for a full slot. Organisers usually load everyone's slides onto one machine to avoid changeover delays, which is why speakers are often asked for a PDF in advance rather than bringing their own laptop. Planning for that shared machine avoids a surprise with fonts, videos or animations that do not survive the handover. It also means your first slide appears while you are still walking to the microphone, so make it one the room can read without you: the title of the idea, your name and nothing else.

Where it shows up in the product

A lightning talk sits at the short end of the generator's length slider, which runs from 3 to 20 slides and starts at 3, so move it to six or eight before generating. Describe the one idea, the audience and the ending you want in the prompt. For the shared machine, export a PDF, which is free and renders every slide as an image, so fonts and layout arrive exactly as designed, though its text cannot be selected. Present mode advances with the arrow keys, Space or Page Down, and N opens your notes if you are presenting from your own laptop.

Questions people ask