The brief is rarely the request

A buyer writes three words and pays you. The three words say "rate this", and the thing they actually want almost never fits in a number, which is why earners who deliver only a number get short reviews and no repeat work.

This is not a psychology exercise. It is the most direct lever on your review quality and your repeat rate, both of which are worth more to your income than your price is.

Why the brief is thin in the first place

Buyers write short briefs for four practical reasons and none of them is that they do not care.

The subject is awkward to write about, and writing more feels like explaining yourself. They do not know the vocabulary, so they use the platform's category name and stop. They assume you will ask if you need more. And they have seen the automated version, where you submit and receive output with no conversation, so a brief feels like an input field rather than the start of a job.

That last one is worth sitting with, because it shapes the whole encounter. Most buyers have already tried something free and instant before they reach a person - the account of what an automated score actually produces is the clearest description of what they have already seen. They are not coming to a human for a number they can already get in four seconds. They are coming for whatever the automated version failed to give them, and the brief does not say what that was.

The five requests underneath

Nearly every brief in this trade resolves to one of five underlying requests. They are not mutually exclusive, and one usually dominates.

Reassurance. The buyer wants to be told that something is fine, and is braced for it not to be. Tells: a defensive or self-deprecating line in the brief, an apology for the submission quality, an explicit "be honest" that is doing a lot of work.

Curiosity. Genuinely wants to know, has no strong emotional stake, treats it as information. Tells: a neutral brief, sometimes a specific question, often a comparison to a previous result.

Novelty. Here for the experience rather than the output. Tells: brief mentions the format more than the subject, asks about the delivery style, or is a gift.

Specificity. Wants an answer to a particular question that the generic version does not address. Tells: a brief that names one dimension and ignores the rest.

Being taken seriously. The dominant one and the least recognised. The buyer wants evidence that an actual person actually looked, carefully, and did not phone it in. Tells: any brief at all that has been written rather than left blank. The act of writing is itself the signal.

Reading which one you have, in thirty seconds

You are not diagnosing anyone. You are reading three signals off the brief and picking a lean.

Length and effort. A long brief is almost never novelty. A one-word brief is either novelty or a buyer who assumes you will ask.

The presence of a question. A specific question means specificity, and it should be answered first, explicitly, before anything else you deliver.

Emotional temperature. Hedging, apologising, pre-emptive framing: reassurance. Flat and factual: curiosity. Jokes: usually novelty, occasionally reassurance wearing a costume, and the way to tell is whether the joke is about the submission or about the situation.

When you cannot tell, ask one question. One, not three - an interrogation loses the booking, and reading a thin brief covers which single question extracts the most. The best generic version is not "what do you want to know" but "is there anything in particular you want me to focus on", because it gives the specificity buyers permission to name their question and gives everyone else an easy pass.

What changes in the delivery

Same work, same time, different emphasis. This is the part that costs you nothing and changes the review.

For reassurance, lead with context rather than verdict. Put whatever the number is inside a range and a frame before you state it, and be specific about what is unremarkable rather than only about what stands out. The failure mode is bluntness delivered as honesty. The other failure mode is hedging so hard that you have said nothing, which reads as evasion and satisfies nobody.

For curiosity, show the reasoning. The observation is worth less than why you concluded it. Curiosity buyers reliably mention "explained why" in reviews and reliably come back with a follow-up question, which makes them the most repeat-prone of the five.

For novelty, invest in the delivery rather than the depth. The content can be perfectly ordinary; the packaging is the product. This is the one case where a recorded format is worth more than a written one at the same price, which is part of what recorded reviews add.

For specificity, answer the question in the first line. Then do the rest of the job. Burying their question in paragraph four reads as not having read the brief, even if the answer is excellent.

For being taken seriously, include at least two observations that could not apply to anything else. Not adjectives - specifics. This is the entire content of the request, and it is why writing specifically without padding is the highest-return writing skill in the trade.

The one that is worth understanding properly

Being taken seriously is underneath most of the others as well as being its own category, and it is the whole reason human work commands a price at all.

Consider what the buyer is actually purchasing. Not information - the free automated version has more consistency than you do and never has an off day. Not speed. What they are buying is the fact that a named person spent real time on their submission and produced something only that could have produced. Research outside this trade shows how much the source matters: in Yin, Jia and Wakslak's 2024 study in PNAS, AI-written responses made recipients feel more heard than human ones, yet recipients "felt less heard when they realized that a message came from AI".

This has a direct pricing consequence. Every element of your delivery that a model could have generated is priced as though a model generated it, which is to say at zero. The parts that are unmistakably human - the specific noticing, the judgement call, the acknowledgement of what is genuinely uncertain - carry the entire fee. What automation took and what it left is the full argument, and this is its operational form: write the parts a model cannot, and be brief about the parts it can.

It also explains a result that surprises new earners. A longer delivery is not a better one. Four sentences of unmistakably specific observation beat six paragraphs of well-organised generality, every time, in reviews and in repeat rate.

Serving the request without overstepping

There is a line, and it matters.

Reading intent means adjusting emphasis and framing. It does not mean telling people what they want to hear, and it does not mean answering questions outside what you were commissioned to do.

Three hard boundaries.

Do not soften into dishonesty. Reassurance is served by context, not by inaccuracy. A buyer who later realises you told them what they wanted has a worse experience than one who got a plain answer, and the review arrives late and lands hard.

Do not stray into territory that is not yours. Health, medical or anatomical questions are not what you were commissioned for and not what you are qualified to answer, regardless of how the brief is phrased. The correct response is one sentence saying it is outside the scope, and then the job you were actually paid for. Staying inside what you actually know is the longer treatment, and this is a boundary worth being unambiguous about before it comes up.

Do not answer a request that was not made. A buyer wanting reassurance did not ask for advice. Unsolicited advice, however kindly meant, reads as a judgement they did not commission, and it produces more complaints than bluntness does.

The distinction throughout is between reading what someone wants from the work and deciding what someone needs in general. The first is professional. The second is not your job.

Where the buyer's own account is better than yours

Everything above is inference from briefs, and inference is second-hand. The buyer-side account of why someone commissions a person, what they hoped for, and what disappointed them is written from the other end of the transaction, and it is worth reading directly rather than reconstructed from your inbox.

Two things it corrects that earners consistently get wrong.

Buyers are comparing risk far more than they are comparing quality. Most of what looks like indecision is a buyer trying to work out whether they will get anything at all, which is why turnaround and responsiveness persuade more effectively than any description of your method does.

And the emotional register of most buyers is milder than earners assume. The picture of a nervous, invested buyer is real and it is a minority. Assuming reassurance is the default request means over-cushioning everybody, and the curiosity buyers find that condescending.

Where your value rests on being methodologically careful rather than emotionally attuned, being explicit about how you arrived at something is what makes that legible, and the conventions worth stating plainly are the reference for saying so without overclaiming.

Turning the reading into a repeatable process

The risk in all of this is that it becomes a thing you do well when you are fresh and forget on job eleven. Three mechanisms stop that.

A rubric with an emphasis dial. Same axes every time, with a note at the top of your template on which section leads. Building a rubric you can reuse is the structure; the intent reading decides the ordering within it, not the content of it.

A first line that names their request back. One sentence at the top of every delivery acknowledging what you understood them to be asking. It costs eight seconds, it catches the cases where you read it wrong while there is still time, and it is disproportionately quoted in reviews.

A record of which requests you get. One column, five values. After forty jobs you will know your own mix, and your mix is what your listing should be written for. If two thirds of your buyers want specificity, a listing selling thoroughness is selling to the wrong third.

That last one connects to pricing. Different requests tolerate different prices - novelty is price-sensitive and volume-driven, specificity is not - and knowing your mix is an input to the rate card you can defend.

The habit that carries most of this

Read the brief twice. Once for what it asks and once for how it is written.

The second reading takes fifteen seconds and it is where the actual request is. Everything in this post is downstream of doing that consistently, and the earners who do not do it are not delivering worse work - they are delivering work aimed slightly to the left of what was wanted, on every job, forever, without ever finding out.

For what buyers are told to expect when they commission on the platform, which sets the baseline you are being read against, Rate Cock's guidance on choosing between an AI and a human judge is the page most of them have seen before they write the brief.