When you freeze at the start of a sentence, the problem is not always vocabulary. You may be choosing content, finding words, planning an opening and monitoring errors at the same time.
Pre-task planning moves some of those decisions earlier. It should answer a few practical questions: What must the listener understand? What comes first? Where am I likely to get stuck? It is not a request to write the perfect English answer.
What research does—and does not—support
Goh's review treats pre-task planning as a scaffold that can cover content, language, organisation and communication strategies. It also makes an important distinction: studies use different tasks, planning times and levels of guidance, and short planning before speaking tests has not produced uniform results.
“Take time to prepare” is not a guarantee of fluency. A more useful lesson is to try guided planning: direct attention to a few relevant decisions instead of leaving yourself with ten minutes and a blank page.
A 30-second planning sheet with four boxes
Imagine asking a colleague to postpone a review. Write only:
- Purpose: move the review from Thursday to Friday.
- Order: request → delayed data → available time.
- Keywords:
data / delayed / Friday morning. - Repair:
What I mean is...orCould we look at another time?
Use English, your first language or symbols. Prompts should help you continue, not invite you to look down and read.
Put the sheet away and record 45–60 seconds. Missing one detail is not failure if your purpose remains clear. For round two, change one box only—perhaps move the result to the first sentence—and speak again.
Plan differently for different tasks
A story: people, problem, turning point, outcome. Do not spend all your preparation on First... Then... while forgetting why the story matters.
An opinion: one position, two reasons and one example. If you have no relevant example, acknowledge the limit rather than inventing one.
A request: the action, reason, time boundary and an alternative. Prepare the polite request itself, not only the explanation.
An interview answer: the question's keyword, a specific experience, its result and its relevance. A memorised “model answer” that cannot survive a follow-up question is too rigid.
Gradually remove the planning support
Start with 60 seconds and four boxes, reduce to 30 seconds, then keep only the purpose and one structural prompt. Compare whether you can organise a similar task with less support, not whether the sheet looks more polished.
If you depend on full English sentences, retain only their opening keywords next time. If keywords make you more anxious, temporarily restore one useful frame such as The main reason is..., then fade it later.
Three common traps
Planning becomes translation. A full translated script consumes preparation time and leaves you trying to recall text while speaking.
Planning only advanced vocabulary. Information order, shared context and an alternative way to explain a missing word often matter more. One familiar expression is worth more than five you cannot retrieve.
Planning replaces speaking. The scaffold must be followed by an unscripted attempt. In an unexpected conversation, Give me a second to think. can also buy a short, natural planning moment.
A small MirrorSay planning loop
Speak one unscripted version and mark the sentence that was hardest to organise. Say your intended meaning to MirrorSay, compare a natural version and repeat it with the audio. Then return to the whole task without reading.
On another day, use task repetition to see whether the message survives without the planning sheet. To reduce word-by-word assembly, add formulaic language practice.
References
- Goh, C. C. M. (2017). Research into practice: Scaffolding learning processes to improve speaking performance. *Language Teaching*, 50(2), 247–260. DOI · Institutional manuscript
- Ellis, R. (2005). *Planning and Task Performance in a Second Language*. John Benjamins. Book DOI
*MirrorSay is a free AI speaking-practice app: say a sentence, see a natural way to express it, then shadow the model. No sign-up, and speech is processed on-device whenever possible.*