The IELTS Writing test runs sixty minutes end to end: Task 1, then Task 2, no break, no bell warning you when time's up unless you set one yourself. Most candidates walk in with the balance backwards. They drill Task 1 charts until describing a trend feels automatic, then find out too late that Task 2, the essay, counts for roughly twice as much of the final Writing band. Twenty minutes on the task that matters less, forty on the task that matters more. That's not a study tip, it's how the score is built.
This page breaks down exactly what each task asks for, how the four band criteria work when an examiner reads your script, what actually separates a 6 from a 7 from an 8 at the sentence level, and where the marks quietly go missing. None of this is theory. It's the same handful of fixable patterns that show up across essay after essay. Candidates plateau not because their English is weak, but because they keep making the same three or four correctable mistakes.
Task 1 and Task 2: what you're actually being asked to do
Both versions of IELTS (Academic and General Training) share the same overall shape: two tasks, sixty minutes, no pause between them. What changes is what Task 1 asks you to produce.
| Task 1 | Task 2 | |
|---|---|---|
| Time | 20 minutes maximum | 40 minutes maximum |
| Minimum words | 150 | 250 |
| Weight in the Writing band | Roughly 1x | Roughly 2x |
| Academic version | Describe a graph, chart, table, or diagram | Discuss an argument, opinion, or problem |
| General Training version | Write a letter (formal, semi-formal, or informal) | Same essay task as Academic |
Academic Task 1: reporting, not arguing
In the Academic test, Task 1 hands you "one or more graphs, charts or tables" or "a diagram of an object, device, process or event" and asks you to describe it in your own words: a bar chart of export figures, a line graph tracking population change, a pie chart comparing energy sources, a flowchart of how something is manufactured, or a pair of maps showing how a town changed over twenty years. Your job is to report, not to explain why something happened or say whether it's good or bad. Giving an opinion in Academic Task 1 is one of the fastest ways to lose marks under Task Achievement. The task is asking you to select and summarise, not to argue.
A strong Task 1 response does three things: opens with a paraphrase of the question (not a copy of it), gives a clear overview of the two or three most significant trends before diving into detail, then supports that overview with specific, selected data, not a sentence-by-sentence recitation of every number on the chart. Examiners are explicit that Task 1 rewards selection. Listing every data point you can see is not the same as identifying what matters.
General Training Task 1: the same clock, a completely different skill
General Training swaps the data-description task for a letter. You're given an everyday situation, a problem with a purchase, a request to a landlord, an application to a local council, and three bullet points telling you what the letter needs to cover. The skill being tested isn't data literacy, it's register: can you tell the difference between how you'd write to a close friend, a landlord you've never met, and a company's customer service department, and can you hold that tone consistently for 150+ words.
A classic example: "You recently bought a piece of equipment for your kitchen but it did not work properly. Write a letter to the shop. In your letter, describe the problem with the equipment, explain what happened when you phoned the shop about it, and say what you would like the shop to do." That's a semi-formal or formal letter: you don't know the manager personally, so "Hey, so this blender you sold me is broken" is a register mismatch that costs Task Achievement marks even if the grammar is flawless. An informal prompt (writing to a friend about a holiday, say) allows contractions, first names, and a much more relaxed opening, and using "Dear Sir/Madam" there would be just as wrong in the other direction.
Task 2: five question types, and they are not interchangeable
Task 2 is a 250-word-minimum discursive essay, and it looks the same on both Academic and General Training. General Training's version is slightly less formal in register, but the underlying task and the four criteria used to grade it are identical. Task 2 questions fall into five recognisable types, and the instruction words in the prompt tell you which one you're looking at. Misreading the type is a Task Response problem before you've written a single sentence, because each type expects a different essay shape.
- Opinion (agree / disagree). You're given a statement and asked how far you agree. Example: "Governments should spend money on public transportation rather than building more roads for cars. To what extent do you agree or disagree?" You need one clear position, held from the introduction through the conclusion, not a summary of both sides with your opinion tacked on at the end.
- Discussion (discuss both views). You're asked to examine two given viewpoints before giving your own. Example: "Some people think that the best way to reduce crime is to give longer prison sentences. Others, however, believe there are better, more effective ways to reduce crime. Discuss both views and give your own opinion." Both views need genuine, balanced treatment. You can't spend four sentences dismissing one side and call that discussion.
- Problem and solution. You're asked to identify causes or problems, then propose solutions. Example: "In many countries around the world, city populations are increasing. What problems does this cause, and what solutions can you suggest to address these problems?" This type rewards essays that pair each problem with a genuinely connected solution, rather than a generic problems paragraph followed by an unrelated solutions paragraph.
- Advantage and disadvantage. Example: "In some countries, more people are choosing to work from home instead of commuting to an office. What are the advantages and disadvantages of this trend?" Depending on the exact wording, you may also need to give a judgement on which outweighs the other. Read carefully for phrases like "do the advantages outweigh the disadvantages," which turns this into an opinion essay wearing an advantage/disadvantage costume.
- Two-part (direct question). One topic, two distinct questions, not necessarily opposing. Example: "Why do many people choose careers that are different from what they studied at university? What can be done to help people make better career choices?" The most common way to lose marks here is answering only one of the two questions in depth and letting the other get a single throwaway sentence.
On word count: there's no upper limit, but going under 150 or 250 words is not just a soft guideline. IELTS quietly removed its automatic word-count penalty around 2018, but writing short still costs you. A response under the minimum almost never has room to fully develop its ideas, so it gets marked down under Task Achievement or Task Response regardless. In practice, aim past the minimum with some margin: something like 170+ words for Task 1 and 270–290 for Task 2 gives you room to develop ideas properly without running so long you have no time left to check your work.
What the four band criteria actually reward
Every script, Task 1 and Task 2 alike, is marked against four criteria, weighted equally within each task: Task Achievement (Task 1) or Task Response (Task 2), Coherence and Cohesion, Lexical Resource, and Grammatical Range and Accuracy. Understanding what moves you from one band to the next matters more than memorising vocabulary lists, because the difference between adjacent bands is rarely about knowing more words. It's about control.
Task Achievement / Task Response
At band 6, the response addresses the task, but some ideas are underdeveloped, repetitive, or only loosely relevant. You can often tell a band 6 essay because it lists points rather than developing them. At band 7, all parts of the task are covered with a clear position and ideas that are extended and supported, even if some support leans generic. At band 8, the response is well-developed, addresses every part of the task fully, and backs ideas with relevant, extended detail; counterarguments get acknowledged rather than ignored. Band 9 is reserved for responses that fully and appropriately satisfy every requirement of the task with only the rarest lapse, a genuinely uncommon result.
- Band 6: "Working from home is good for employees. It saves time and money. This is why many people prefer it."
- Band 8: "Working from home removes the daily commute, which in large cities can mean recovering close to two hours a day: time that goes toward family, exercise, or further study, all of which feed back into long-term productivity."
The band 6 version states three claims and stops. The band 8 version takes one of those claims and actually explains the mechanism behind it. This is the difference between listing an idea and developing it.
Coherence and Cohesion
Band 6 essays use cohesive devices, but often mechanically: a string of "firstly / also / moreover" connectors doing the work that logical sentence flow should be doing, and paragraphing that doesn't always match one idea per paragraph. Band 7 essays organise information logically with clear overall progression, using a wider range of cohesive devices, though not always with full precision. Band 8 essays sequence ideas logically and manage cohesion (referencing, substitution, paragraphing) so well that it stops being noticeable. Band 9 essays use cohesion so skilfully it almost disappears from view.
- Band 6: "Firstly, remote work saves time, also it reduces stress and moreover it saves money, so it is good for employees and also for companies because they don't need big offices."
- Band 8: "Remote work saves employees time and reduces stress, while companies benefit from lower office costs. These combined savings explain why many businesses kept flexible arrangements even after pandemic restrictions ended."
The first is a single run-on sentence stitched together with connector after connector. The second splits the idea into two sentences, drops the connector overload, and uses "these combined savings" to link back. That's cohesion working through reference rather than through a wall of linking words.
Lexical Resource
Band 6 vocabulary is adequate to talk about the topic but shows noticeable errors in word choice, word form, or spelling: errors that don't block meaning but are visible on every page. Band 7 vocabulary is sufficient for flexibility and some precision, with less common words attempted and mostly controlled. Band 8 vocabulary is used fluently and flexibly to convey precise meaning, with only occasional inaccuracy in less common word choices. Band 9 vocabulary reads as natural and sophisticated, with errors occurring only as rare slips.
- Band 6:
Many peoples has a informations about this problem, but government don't do enough action to solve it. - Band 8: "Many people are aware of this problem, but governments have failed to take sufficient action to address it."
Notice what actually changed: "peoples" → "people" (uncountable-noun error, the same family of mistake as writing advices instead of advice), "a informations" → dropped article and pluralised correctly, and "don't do enough action" → "have failed to take sufficient action," which is a collocation fix, not a fancier synonym. Band 8 vocabulary isn't bigger words. It's the right ordinary words used precisely.
Grammatical Range and Accuracy
Band 6 shows a mix of simple and complex sentences, but complex structures are used with limited flexibility and frequent error, usually not enough to block meaning, but enough to be constantly visible. Band 7 shows a variety of complex structures with a good number of error-free sentences and generally sound control of grammar and punctuation. Band 8 uses a wide range of structures flexibly and accurately, with the large majority of sentences error-free. Band 9 shows full flexibility and precision, with errors occurring, again, only as slips.
- Band 6: "The number of cars increased a lot and then it decreased a little and after that it stayed the same until the end of period."
- Band 8: "The number of cars rose sharply before falling slightly, then remained stable for the remainder of the period."
The band 6 version is three clauses bolted together with "and" and "and then," grammatically legal, structurally flat. The band 8 version uses subordination ("before falling") and a participle-style compression to say the same thing in fewer words with more control. This is the single most teachable grammar upgrade in IELTS writing: trading strings of "and" for subordinate clauses.
Where the marks actually go missing
Giving an opinion in Academic Task 1. Task 1 is a reporting task. Writing "this is a worrying trend" or "the government should address this" doesn't just fail to help. It signals a misunderstanding of what the task is asking for, and that's a Task Achievement problem, not a style choice.
Describing everything instead of selecting. Academic Task 1 explicitly asks you to select and report the main features. Candidates who try to mention every number on the chart usually run out of time before writing an overview, and a Task 1 response without a clear overview is capped low no matter how accurate the individual sentences are.
Register mismatches in General Training letters. Starting a formal complaint letter with "Hi there" or ending an informal letter to a friend with "Yours faithfully" reads as a fundamental misunderstanding of the task, and IDP's own guidance on letter writing flags this as one of the most common, and most avoidable, errors in General Training Task 1.
Sitting on the fence in opinion essays. An opinion (agree/disagree) essay needs a position held from the first paragraph to the last. "Both sides have some merit" as a conclusion, with no real stance taken anywhere in the essay, is one of the most reliable ways to get stuck at band 6 on Task Response. The examiner genuinely cannot tell what you think.
Answering only half of a two-part or discussion question. Giving four sentences to one half of a two-part question and one throwaway sentence to the other caps Task Response regardless of how well the covered half is written, because the task explicitly asked for both.
Forcing complexity instead of building it. Many of the ugliest grammar errors in Task 2 scripts come from candidates attempting a long compound-complex sentence that collapses halfway through: a dropped clause, a missing verb, a reference word with nothing to refer back to. A shorter, accurate sentence outscores a longer, broken one every time; Grammatical Range rewards controlled variety, not length.
Treating advanced vocabulary as a shortcut. Swapping an accurate simple word for an impressive-sounding but slightly wrong one (thesaurus-driven substitutions are a common culprit) tends to cost Lexical Resource marks, not earn them, because inaccurate word choice muddies meaning. Precision beats decoration.
Getting number punctuation wrong in Task 1. Candidates whose first language uses a comma as a decimal separator sometimes write 1.000 to mean one thousand, or flip the roles of commas and periods in figures generally. IDP flags this specifically in its guidance on Task 1. It's a small habit, but it can make an accurate description of the data read as inaccurate to an examiner, which drags down Task Achievement for a mistake that has nothing to do with understanding the chart.
A preparation plan that isn't "practice every day"
Writing twenty average, unreviewed essays over a month teaches you almost nothing. You're rehearsing the same mistakes twenty times. Three essays a week, each one actually reviewed against the four criteria and then partially rewritten with the fixes applied, moves your band further than daily unstructured writing. The rewrite step matters more than the first draft: taking a graded essay and reworking the two or three weakest sentences, rather than abandoning it for a fresh topic, is what actually cements a correction.
Build a paraphrasing habit for Task 1 openings specifically. Copying the question's wording into your opening sentence is one of the most common Task 1 habits, and it signals limited lexical control. Keep a running list of swaps you can reach for automatically: "The chart shows" → "The graph illustrates," "between 2000 and 2020" → "over the two decades from 2000 to 2020," "the percentage of" → "the proportion of." Practice this in isolation: five minutes rewriting just the opening sentence of ten different Task 1 prompts is more useful early on than writing five full Task 1 responses.
Structure Task 2 around four paragraphs, not five. Introduction (a paraphrase of the question plus a one-sentence outline of your position, two sentences is enough, a third rarely earns anything), two body paragraphs (one main idea each, explained and supported with a specific example, roughly 80–100 words per paragraph), and a short conclusion that restates your position without introducing anything new. Discussion essays sometimes tempt writers into five paragraphs (one for each view plus a separate opinion paragraph), but folding your opinion into the second body paragraph or the conclusion is usually tighter and easier to control under time pressure.
Practice the time split, not just the total. Forty minutes for Task 2 breaks down roughly into five minutes planning, thirty to thirty-two minutes writing, and a few minutes checking. Most weak essays come from skipping the planning step entirely and writing into a corner: starting a paragraph without knowing how it ends. Time the stages separately in practice, not just the whole task, so the planning habit becomes automatic rather than something that gets dropped the moment you're under pressure.
Do a handful of essays under full exam conditions before test day. Comfortable, untimed writing at home and sixty minutes of continuous, no-pause writing under a countdown are different skills. The last two weeks before a test should include at least a few full 60-minute attempts, Task 1 straight into Task 2, exactly as it happens in the real exam.
Track your mistakes as categories, not as one-off corrections. After a handful of graded essays, stop reading feedback essay-by-essay and instead lay three or four scripts side by side. If the same tag keeps showing up (subject-verb agreement, missing overview, weak topic sentences), that's not bad luck, it's a pattern, and patterns are what actually cap a band score. Fixing the pattern once is worth more than fixing five individual instances of it across five separate essays.
What SmoothMock's grading does differently
Most feedback loops in IELTS prep are slow: write an essay, wait days for a tutor, get a single number back with a paragraph of comments. SmoothMock grades every attempt against the same four official criteria above and returns sentence-level inline corrections mapped to a mistake taxonomy, so an error like a informations gets tagged as an article-agreement issue under Grammatical Range, not just marked wrong. The goal isn't a faster number. It's seeing which specific pattern is costing you bands, attempt after attempt, so you can fix the pattern instead of guessing at it.
Sources
- IELTS Academic: Writing test format
- IELTS General Training: Writing test format
- IELTS scoring in detail
- IELTS Writing Band Descriptors (public version)
- IELTS Writing Penalty for Being Under Words — ieltsliz.com
- Difference Between Band 6 and Band 7 — IELTS Charlie
- IELTS Writing Task 2 Question Types — Magoosh
- Common Mistakes in IELTS — British Council
- Common IELTS mistakes: dealing with figures — IDP
- Common errors to avoid in IELTS letter writing — IDP
- Avoid these common grammar mistakes in IELTS Writing — IDP
- Task 1 and Task 2: similarities and differences — IDP
- How many paragraphs for an IELTS essay — ieltsliz.com