Write the review your reader needs, not the one that shows off your taste. That means answering the decision in the first paragraph — should this person read, watch, or play this thing, and which version — and then backing the answer with specifics the reader can check. A review that helps is short on mood and long on verifiable detail.
The craft, per Dictionary.com's entry on write, is simply to "express or communicate in writing" — and communication is the whole job here. You are not performing. You are handing a busy person a verdict plus the reasons, in an order they can use. Everything below follows from that.
What should the first paragraph of a review do?
It should answer the question the title implies, in one or two sentences, with the key qualification attached. "Start with the first collected volume — skip the tie-in miniseries until later." "The game is worth your time if you like slow builds; the first three hours test your patience." The reader should be able to stop after paragraph one and still make a good decision.
Then earn the verdict. Give one concrete, checkable detail early: the edition you reviewed, the platform, the runtime, the number of volumes currently in print. Vague openers like "this long-awaited release finally arrives" tell the reader nothing. A date, a price, a format — any one of these anchors the whole piece.
How much of the story can you spoil?
None of it above a clearly marked line. A useful review says what a work is about, never what happens. "A widow investigates her husband's final case" is a premise. "The twist is that the partner did it" is a spoiled story, and no reader forgives it.
The practical habit: draft the piece, then read every sentence and ask whether it changes how the work unfolds. If it does and it isn't in the official description, move it below a spoiler warning or cut it. Critics who handle this well treat the spoiler line as a promise, not a suggestion. Readers reward that promise with return visits. This connects to our earlier piece, The Best Batman Graphic Novels for New Readers, Ranked by Commitment.
What separates a helpful verdict from an opinion dump?
Specifics that a reader can verify or act on. Compare these two sentences:
- "The pacing drags in the middle."
- "The middle third repeats the same investigation loop four times before the plot moves."
The second one tells the reader exactly what they'll be wading through and lets them judge whether that bothers them. Opinion is fine — a review without a stance is a summary — but every stance needs an observable fact under it. If you can't point to the scene, the chapter, or the system that produced your reaction, either find it or soften the claim.
How do you judge a work fairly when you didn't finish it?
You mostly can't, and you should say so. A review of a hundred-hour game based on twelve hours of play is a first-impressions piece, not a review — label it that way. The same goes for a comic judged on one issue of a six-issue arc, or a series reviewed from the pilot alone.
Finish what you can. If a work is genuinely unfinishable — broken, interminable, or simply bad past the point of duty — say where you stopped and why, in plain terms. "I stopped at hour forty because the progression system stopped producing new decisions" is honest and useful. An unexplained abandon is not.
What this means for your process
Here is the working method, in order:
- Consume the whole thing first. Credits, side content, the ending you didn't like. Notes taken mid-experience capture reactions; notes taken after capture patterns. You need both.
- Write down what the work asks of the reader. Time, money, attention, prior knowledge. A reader deciding whether to buy needs this list more than they need your star rating.
- Pick the three facts that drive your verdict. Not ten. Three. If the verdict can't rest on three specifics, the verdict isn't formed yet.
- Answer first, argue second. Verdict in paragraph one. Evidence after. Context last.
- Check the spoiler line. Everything above it survives a skim from someone who hasn't started the work.
- Disclose what you got. A review code, early access, a press ticket — say so, once, near the top or bottom. Disclosure costs nothing and buys credibility that no adjective can.
One more habit worth stealing from the best critics: review the experience over time, not just the launch weekend. How does the game hold up at hour fifty? Does the show's second half justify the first? Does the collected edition hold together on a re-read? Pacing, systems, and value reveal themselves slowly, and the reviews that age well are the ones written after the honeymoon.
How long should a review be?
As long as the decision requires and no longer. A single-volume graphic novel needs maybe 500 words: what it is, who it's for, what works, what doesn't, which edition to buy. A hundred-hour role-playing game may need 1,500. A streaming series season sits somewhere between, depending on how uneven it is.
Length inflation is a real failure mode. Padding a short work to a long review usually means summarizing plot to fill space — which is how spoilers happen. If you're under your target word count and the piece is complete, ship it short. The reader's time is the resource you're spending, and they gave it to you for a reason.
What about scores and star ratings?
If your outlet uses them, treat the number as a conclusion, never an argument. The score is the last thing you decide and the first thing readers will argue about, so the body text has to stand on its own without it. A reader who reads your reasons and disagrees with your number has still been served. A reader who gets a number and no reasons has not.
If you don't use scores, say what you recommend instead: "start here," "wait for a sale," "for fans of the genre only." Directional guidance often serves readers better than a ten-point scale, because it maps onto actual decisions.
Where should a new critic start?
Start with works you know well and can finish cheaply — a standalone graphic novel, a short season, a compact game. Read your draft against the checklist above: answer first, three specifics, no spoilers above the line, disclosures made. Then read published critics whose reviews you trust and reverse-engineer their structure: where the verdict lands, how the evidence is ordered, where the spoiler line sits.
Reading widely also sharpens judgment. Our Reviews of Classic Animated Films: A Starter Canon is one way to build a reference library of works most readers know, which makes your comparisons legible. Comparing a new work to a shared reference point communicates faster than any adjective.
The closing takeaway
A review that helps readers does four things: it answers the decision early, it supports the answer with checkable specifics, it protects the story from spoilers, and it discloses its own conditions — what was finished, what was provided, what was paid. None of this requires talent so much as discipline. The mechanics respect your time when you respect the reader's.
Write with that contract in mind and the rest — voice, style, personality — is decoration on a sound structure. Skip it, and no amount of clever phrasing will save the piece.
