Reviews

An honest read from a stranger

An honest review is a fair, unleaning opinion in plain language, and it is thinner than buyers expect because an opinion about a photo is a small thing.

By 3 min readReviews

Guides on Reviews: SPH as a genre with conventions, Five things a review is mistaken for, The words people use, defined, What happens in a two-minute clip

An honest review gives you one person's plain, fair read of your material, without flattery or added harshness, and it is usually shorter than expected. It is an opinion, not a measurement, which is why "just be honest" is the brief most likely to disappoint a buyer who was quietly hoping for a fact.

What is actually inside one

An honest review is plain-language commentary, delivered without exaggeration in either direction - not inflated into flattery, not sharpened into anything harsher than the judge genuinely means. It responds to your material and your brief the way any other register does, but without a performed tone doing extra work: no worship building it up, no edge cutting it down, just a fair read stated clearly.

That plainness is exactly why an honest clip tends to feel shorter than a buyer expected. A worship clip can sustain a minute by building appreciation across several specifics. An honest clip says what it has to say, states it once, and stops, because padding a genuine opinion with restatement does not make it more genuine, it just makes it longer without adding anything true.

Why it disappoints the people who want a fact

The core misunderstanding is treating an opinion as though it were a measurement. A judge giving an honest read is telling you what one person, looking once, actually thinks - which is real, and can be useful, but it is not data, it does not generalise to anyone else's reaction, and it certainly is not a number you could compare against a population. Two honest judges giving two honest reads on the same material can land in different places without either one being wrong, because an opinion is allowed to vary between honest people in a way a measurement is not. Research that depends on human raters builds this in: McHugh's 2012 review treats interrater reliability, the extent to which raters "assign the same score to the same variable", as something a study has to test rather than assume.

If what a buyer actually wants is a fact rather than a reaction - a real figure, stated plainly - that is a different question with a different, proper answer, and no amount of asking a judge to be "really honest" will turn their opinion into that fact, because it was never built to be one.

What honesty is not

Honesty is not the same thing as harshness, and buyers sometimes brief this register expecting something closer to a harder-edged tone than plain honesty actually delivers. If what you want is a harsh delivery, that is closer to a different register entirely, and asking for "honest" when you mean "hard" tends to produce a clip that undersells what you actually wanted. The register you name and the tone you get should match, and honest specifically means fair and unleaning, not blunt for its own sake.

It is also not a substitute for consistency. An algorithmic score is built to return close to the same read on close to the same input every time, which is a genuinely different property from one person's honest opinion on one occasion - a person is not being inconsistent by giving a different honest answer on a different day to a differently worded brief, because a person's honest opinion was never claiming to be repeatable in the first place.

What it is good for

None of this makes an honest review worthless; it makes it a smaller, more specific thing than buyers sometimes hope for going in. A fair, considered, plainly stated reaction from a real person, addressed to your actual material and brief, is a genuine product - it simply is not the product some buyers accidentally asked for when what they actually wanted was reassurance, or a fact, or permission to feel fine about something. Working out which of those you actually want before you write the brief is the single best way to avoid the letdown this register causes more than any other.

If a score alongside the opinion is useful to you, an automated read gives you a consistent number to sit next to the honest one, and the two together - one stable, one specific to a person's actual reaction - cover more ground than either does alone. Ordering an honest read through a platform that keeps the brief attached to the request is what keeps the judge answering the actual question you asked, rather than the vaguer one "just be honest" leaves open to guesswork.

Read next

Full archive