ESL Speaking Activities: The Complete Guide
The widely-cited rule of thumb in language teaching is the 70/30 principle: students should be talking for 70% of a lesson, teachers for the remaining 30% — with some methodologists pushing for 80/20 where possible. Most classrooms don't get there. That gap is the entire reason ESL speaking activities exist as their own category of lesson design, separate from grammar or vocabulary work: the goal isn't covering material, it's engineering enough genuine student talk time to close that ratio.
This page is the hub for everything English, no kidding has published on the subject — three deep-dive posts, each covering one piece of what actually makes a speaking activity work, plus the five apps built specifically around spoken production. If you're looking for one specific angle, jump to the section below. If you want the full picture, read top to bottom.
What actually makes an activity a "speaking activity"
Not everything that involves talking counts. A worksheet read aloud, a vocabulary list recited in unison, a single student answering a teacher's question while 23 others wait their turn — all involve speech, none of them build spoken production the way CEFR describes it. A genuine ESL speaking activity has three properties, and each one maps to a specific problem this hub's spoke posts cover in depth:
- It survives repetition. A game that's identical every time it's run stops producing real language the second time a class plays it, because there's nothing left to figure out. Covered in No-Prep ESL Games.
- It creates a genuine information gap. Two students each holding information the other lacks negotiate meaning without being told to. Two students told to "discuss" the same shared prompt run out of things to say within thirty seconds. Covered in Peer-to-Peer Speaking Activities.
- It demands the right task for the level. A B2 student doing a B1 task with harder vocabulary isn't doing B2 work — CEFR asks for a different cognitive job at each level, not a bigger word list. Covered in B1 vs B2 Speaking Activities.
Miss any one of these and the activity quietly stops working, even if it looks identical on the surface. A game can be built around a real information gap and still fail if it's the fourth time the class has played it. A pair task can be brand new every round and still fail if it doesn't ask enough of a B2 student to actually stretch them.
The three building blocks, one level deeper
Reuse without repetition. The classic no-prep games — word chain, board race, two truths and a lie — survive because their mechanic is the whole activity, not because the content stays fresh. That works for a first run and starts to feel thin by the third. AI-generated activities solve the same zero-materials constraint differently: same instant setup, but the prompt, topic, or vocabulary set is different every time a teacher generates a new round. The full breakdown, including which classics pair well with which generated activities, is in No-Prep ESL Games: What to Run When You Have Zero Minutes to Plan.
Information gap. The British Council's own definition of an information gap task credits it with producing genuine communication and forcing "sub-skills such as clarifying meaning and rephrasing" — exactly the negotiation a scripted Q&A never requires, since there's nothing to clarify when the answer is already known. The catch is that a gap only works once; run the same pair prompt twice and the second round is recitation, not speaking practice. That's why the gap has to regenerate every time, which is a harder problem than it sounds for a teacher writing prompts by hand. Full detail in Peer-to-Peer Speaking Activities for ESL.
CEFR-calibrated demand. The Common European Framework's own spoken-production descriptors don't ask a B2 student to say more — they ask for a different task: not narrating an experience, but arguing a position with advantages and disadvantages, sustained under mild pushback. Swapping in harder vocabulary while keeping the same "describe X" task leaves a class doing fluent B1 work in bigger words, not genuine B2 work. The concrete swaps that turn a B1 prompt into a real B2 one are in B1 vs B2 Speaking Activities: What Actually Changes at Each CEFR Level.
The five ENK apps built around spoken production
These are the apps in the English, no kidding lineup that this hub's spoke posts draw their concrete examples from — all built around getting students producing language directly with each other or the class, not answering into a worksheet.
| App | Formation | CEFR range | Core mechanic |
|---|---|---|---|
| Spin & Speak | Free Circle | A1–C1 | A spin generates a topic and keyword scaffolds; the student speaks for up to 60 seconds without stopping while the class tracks fluency live |
| Dialogue Factory | Pairs | A2–C1 | Pairs rehearse an AI-generated dialogue, then progressively lose the lines until they perform the scene from memory alone |
| Ask, Tell, Reveal | Free Circle | A1–C1 | Students take turns asking, answering, or revealing something personal, each prompt aimed at one specific classmate rather than the room |
| Iconic Stories | Small Groups | B1–C1 | A group builds one collaborative story live, each student extending the narrative in turn with no notes allowed |
| Sentence Lab | Whole Class | A2–C2 | Three random words drop on screen; students build and defend one sentence that uses all three |
Every one of these regenerates its content on demand — the same app run again next lesson produces a different topic, dialogue, or word set, which is the specific property that keeps a speaking activity from going stale the second or third time a class runs it.
Sequencing a speaking-first lesson
A realistic build across a single lesson, using the properties above in order:
- Open with a no-prep warm-up — word chain, two truths and a lie, or a quick Sentence Lab round — to get language moving with zero setup cost.
- Move to a genuine pair task — Dialogue Factory or Ask, Tell, Reveal — so every student is producing language simultaneously, not waiting in a queue for a turn.
- Calibrate the demand to the class's actual level — a B1 group needs narration and basic reasoning; a B2 group needs comparison, hedging, and a position that survives a follow-up question. Swap in the appropriate tier of Spin & Speak or Iconic Stories rather than just harder vocabulary on the same task.
- Close with a whole-class moment — Sentence Lab's class vote or a Spin & Speak round in front of the group — so the pair work that just happened gets a visible, shared payoff instead of disappearing the moment the pairs stop talking.
This sequence works across a single 45-minute lesson or across a full unit — the same three moves (reuse without repetition, genuine gap, calibrated demand) apply at either scale.
Where speaking activities usually break down
Most speaking activities that fail don't fail on day one — they fail on the second or third run, once the novelty that carried the first attempt wears off. Three failure patterns cover almost every case: a game gets reused until students have memorized the mechanic and stop producing new language; a pair task shares its "gap" so obviously that both students already know the answer before they start talking; or a task stays identical across levels, so a B2 class ends up doing the same describe-and-explain job as a B1 class, just with fancier words. Each of these maps directly onto one of the three properties above, which is exactly why the fix is rarely "try a completely different activity" — it's usually "regenerate the content," "tighten the gap," or "raise the task demand," using whichever of the three spoke posts covers that specific failure.
All five apps above sit inside the same account as ENK's other 21 apps, with 100 AI credits a month at no cost, or from €5/month for 300 credits on the pricing page — no separate tier for speaking-focused formats.
FAQ
What makes an ESL speaking activity different from any activity that involves talking? A genuine speaking activity survives repetition without going stale, is built around a real information gap rather than a shared prompt everyone can already see the answer to, and demands the right task for the student's CEFR level rather than just harder vocabulary on an easier task.
What's the single biggest mistake teachers make with speaking activities? Treating a harder version of a task as the same thing as a higher-level task. A B2 student describing a holiday in more advanced vocabulary is still doing B1-level work — B2 speaking activities need to ask for a comparison, a hedge, or a position that survives pushback, not just bigger words.
Do ESL speaking activities work for large classes? Pair and small-group formats (Dialogue Factory, Iconic Stories) scale to large classes far better than whole-class Q&A, since every student is talking at once instead of waiting for one turn. A class of 30 running pair work produces 15 simultaneous conversations rather than 30 students waiting to speak once each.
Which ENK app should I start with if I'm new to speaking-focused lessons? Spin & Speak (A1–C1) is the simplest entry point — one topic, one student, sixty seconds, no pairing logistics to manage. Once a class is comfortable with solo spoken output, move to pair formats like Dialogue Factory or Ask, Tell, Reveal.
Written by Prof. Vito Schiuma, designer of English, no kidding.