Paragon publishes the four Writing dimensions by name — Content/Coherence, Vocabulary, Readability, Task Fulfillment — and almost everyone prepares for the middle two. They build vocabulary lists. They check their grammar.
Task Fulfillment is the one nobody studies, and it is the one with a floor underneath it. A response that gives insufficient information to assess scores 0 or 1 — and 0 and 1 map to no CLB level at all.
That is the imbalance. The two dimensions people work on are the two that cannot rescue the two they ignore.
The failure
You write 180 clean, careful words about the first thing the prompt asked for, run out of time, and never mention the second and third things at all.
Nothing you did wrong is visible at the sentence level. Your vocabulary was fine. Your grammar was fine. The response is simply not an answer to the question that was asked, and the dimension that notices is the one you did not prepare for.
Task Fulfillment — first, deliberately
Paragon's own Task 1 sample gives exactly three bullets of things the email must cover, and the sample instruction asks for about 150–200 words. Those bullets are not suggestions. They are the specification.
- For the email (Task 1): every bullet answered, in the role the prompt gives you — a tenant, a customer, an employee. A complaint to a company and a note to a close friend are not interchangeable.
- For the survey response (Task 2): you are given options and asked to choose. Choose one, say which, and give reasons for that choice. A response that weighs both options beautifully and never commits has not done the task.
What a marker sees is coverage and fit: is each required point actually here, and does this read as the right kind of message from the right person? A missing bullet is not a small deduction — it is a part of the response that was supposed to exist and doesn't.
Content/Coherence
Coverage is Task Fulfillment; depth is here. One sentence per bullet technically covers it, but gives a marker nothing to rate — each point needs a reason, an example, or a consequence attached. Coherence is the other half: ideas in a sensible sequence, each paragraph about one thing, no doubling back to a point you already finished.
Vocabulary
Here is the useful thing rather than the expected thing: this is not a test of rare words.
A precise ordinary word beats an impressive vague one. "The screen is cracked" tells a marker more than "the item is in an unsatisfactory condition." And a wrong "advanced" word costs you more than the simple word it replaced — it loses precision and reads as an error, where the plain word would have read as nothing at all.
Range still counts — but range means having the right word for each different thing you need to say, not decorating one idea with a longer synonym.
Readability
The officially published sub-points are: "Format and paragraphing; Connectors and transitions; Grammar and sentence structure; Spelling and punctuation".
Note what leads that list. Format and paragraphing comes first — a greeting, paragraphs with white space between them, a sign-off. An email delivered as one 200-word block is losing marks here before a single sentence is read. Grammar and spelling come third and fourth, and they are what most people spend all their preparation on.
Three of the four are also your Speaking score
Speaking is rated on Content/Coherence, Vocabulary, Listenability, Task Fulfillment. Writing is rated on Content/Coherence, Vocabulary, Readability, Task Fulfillment.
Only the fourth differs — Readability for text, Listenability for speech. Work on Content/Coherence, Vocabulary and Task Fulfillment and you are preparing for both components at once. Answering the actual question is a transferable skill; typing quickly is not.
The same email, twice
(original AlmiWorld practice prompt)
You ordered a replacement part for your bicycle from an online shop three weeks ago. It arrived damaged. Write an email to the shop's customer service team. In your email: — say what you ordered and when — describe the problem with the item — explain what you would like them to do
Version A
Hello,
I am writing to inform you regarding a lamentable circumstance with my recent acquisition from your esteemed establishment. The product which I obtained was not in a satisfactory condition upon its arrival, which was highly disappointing. I have been a customer for a long time and did not expect this outcome. Please look into this matter at your earliest convenience.
Thank you.
Version B
Dear Customer Service,
I ordered a rear brake cable for a hybrid bicycle on 2 August, order number 41889.
The package arrived on 19 August, but the cable housing was split near one end and the metal ferrule was bent. The outer box was undamaged, so I think the part was already broken when it was packed.
Could you please send a replacement this week? If that is not possible, I would like a full refund instead. I have photographs of the damage and can email them if they would help.
Yours sincerely, Daniel Ortiz
Dimension by dimension:
| Dimension | Version A | Version B |
|---|---|---|
| Task Fulfillment | Bullet 1 missing — no item, no date. Bullet 3 missing — "look into this matter" is not a request. | All three bullets answered explicitly; register fits a company. |
| Content/Coherence | One idea, restated three ways. The loyalty line goes nowhere. | One bullet per paragraph, each with a supporting detail. |
| Vocabulary | "lamentable circumstance", "esteemed establishment" — reaching, and vaguer than plain words. | "split", "bent", "ferrule", "refund" — ordinary, exact. |
| Readability | Single block, no structure, no real sign-off. | Greeting, three paragraphs, sign-off; connectors do real work ("but", "so", "if that is not possible"). |
Version A is not bad writing. It is decent writing answering a different question.
Nobody publishes the weighting
No official source states how the four dimensions are weighted. Nobody can tell you Task Fulfillment is worth N per cent — not us, not any prep site quoting a number at you.
What is published is the floor: a response giving insufficient information to assess scores 0 or 1. A floor does not need a weighting to hurt you. It sits underneath the rubric, not inside it.
What actually costs you
- Answering one bullet in depth and leaving the others unwritten.
- Discussing both survey options without ever choosing one.
- A rare word used slightly wrongly where a plain word was available.
- No paragraph breaks — Readability's first sub-point, lost before you write a sentence.
- Spending your last minutes hunting for a better adjective instead of checking every bullet is actually there.
Try one
[Write a Task 1 email →] — one full task, marked, no card required.