Creator Marketing Playbooks
What do a creator's reviews tell you?
What one brand's account of a finished job can settle for you, and what was never in it.
You're a few candidates into a shortlist and one of the profiles has writing on it. Other brands have worked with this person before and left short written accounts of how it went, and those accounts sit on the page for the next brand to read.
So you read them, and the reading is the easy part. You're reading an account of a campaign you were not in.
Here a review means what the guide to writing one means by it: a short written account of working with this creator, left where a brand considering them next can read it. That rules out your own notes on a draft. It also rules out what the creator sent you about themselves, and anything your own brand left there. Both of those are boundaries this article adds. Writing one is a separate job and that guide has it. You're at the other end, with somebody else's finished one in front of you.
For the clip: you sell a hair clip, and what it has to prove is that you can twist up wet, heavy hair with one hand and have it hold all day. Three reviews are on the profile you're looking at. None of them was written by anyone who knew that.
What was the person who wrote it looking at?
Their own campaign.
The guide to writing one of these is direct about the asymmetry, and the line it uses is addressed to the brand doing the typing: "you are not the reader." The record it tells them to write from is the one they kept on their own side of the campaign, which you have never seen, and a later section says what the person at your end is missing: "That record was written for a reader who has the campaign sitting next to it, one who already knows what you asked for and when. The person reading the review has neither."
That last line is the guide's judgment about the reader. The surrounding advice is for the writer; applying it here to the next brand's view is this article's extension. Whatever the writer left out didn't travel, and the sentence they left is what you have.
For the clip: the brand that wrote the first review could see their whole campaign while they typed it. You can see one line of it.
How do you tell whether a review carries anything?
Not every one of them is holding something. The same guide asks the brand about to write one: "can somebody holding nothing but your sentence read out of it what you asked for and what came back?" Applying that writer's test to a review already on a profile is this article's extension, and it looks for the same "both halves, the ask and what came back". Run it on what is on the profile.
We asked for the bottle to be in the shot from the first pump to the last comb-through, and the first cut skipped straight to the dry result.
Both halves are in that one. A brand named what they asked for and named what arrived the first time, and you read all of it without having needed to know their campaign first. Whatever else is true about this creator, that sentence is holding something.
Would work with them again.
Neither half is in that one. The writing guide has already ruled on this shape: "There is no ask in it and no outcome, only a verdict, and the next brand cannot do anything with it. Put it in if you like; it costs nothing and it carries nothing. Just don't let it be the whole thing." The guide is not telling brands to stop writing these. It says you can put one in and that nothing comes out of it, so long as it is not the whole of what you leave. So the one on the profile in front of you is not a point for this creator, and it is not a point against them either.
For the clip: two of the three reviews are now sorted, and only one of them has given you anything. The third is the one that's harder to put down.
What about the one that reads like a complaint?
Posted later than we wanted.
That one reads like an outcome, which is why it is harder to put down than the verdict. Later than what, though. The guide to a video that misses its date says: "Something has to have fixed it. Otherwise nothing is late." That rule is for an open campaign where the brand can check what was agreed and whether it was broken. Applying it here, to a finished job described by someone else, means asking what this review lets you verify.
This sentence gives you no date, window, or reference to agreed timing. Wanted might refer to an agreed date or only to the writer's preference; the review alone doesn't tell you which. From this sentence, you cannot confirm what schedule the two sides agreed to or whether the creator delivered late against it.
So it lands with the verdict for this reading task, and not because it proves there was no deadline or no delay. The review-writing guide tests an outcome with no ask behind it from the writer's side and concludes: "It fails the test as written." Applying that test to the sentence on this profile gives the same limit: it looks like evidence, but it doesn't give the next brand enough context to check the complaint.
For the clip: two of the three on this profile named nothing you could check, and one of those two looked like it did.
What has the readable one settled?
One job, from the ask to what came back.
Another brand named an ask, the first thing that came back didn't have it, and somebody wrote both of those down in one sentence. The writing guide describes what that leaves a reader holding: both halves are there and "neither one leans on the campaign".
That's worth having. It's also where the sentence stops. You know what one brand asked this person for and what the first cut did with it. You don't know what else the brief fixed, how many rounds of changes they had bought, what else came back after that, or what either side thought of the finished video, because none of that is in the sentence and the sentence is all that travelled.
For the clip: the sentence you're holding is about somebody else's bottle.
Why doesn't that tell you whether they fit your campaign?
Because fit is a question about your campaign, and your campaign was not the one being written about.
Look at what the readable one is made of. Another brand asked for their bottle to stay in frame through an application, and the first cut skipped to the end of it. That lands close to your shelf — same category, same wet hair, a product that goes on in the same bathroom — and it still doesn't settle your question. Yours is whether this person can twist up heavy wet hair one-handed and leave it holding all day, and whether the people watching them would recognize it as a thing worth watching. Neither of those is in the sentence, because neither of those is what the other brand asked for.
That's what a readable review is: one job, under somebody else's ask. Where the ask resembles yours you have a lead worth following rather than a check you have run.
That matters for where reading reviews goes in your week. The guide to choosing creators for a campaign owns the checks that settle fit, and it has a rule for when one of them gets to go early: "A check earns an early slot by being cheap and able to settle whether someone fits." Reading the reviews is as cheap as a check gets. The writing is on the page already, and getting to it involves nobody but you. It's the second condition that reading them misses, and the rule joins the two with an and.
That guide doesn't list review-reading among its checks. Applying its scheduling rule here is this article's extension, and only to decide where the task belongs in your week. Fit remains open because your campaign was not the one being reviewed.
What does the number of reviews count?
How many of them are on the page.
One competitor's creator training on building a review history tells creators to make asking for one a habit, during a campaign or after it: "Make it common practice, to always message campaign managers while either actively working on a campaign, or after finishing one to ask for a review." That's what the training teaches. Whether the person whose profile you have open ever did any of it is not something the profile says.
The profile doesn't say whether this creator or anyone else followed that advice, or how the count got to where it is. This article therefore treats the count only as the number of reviews available to read, not as proof of how many jobs the creator completed.
For the clip: the profile you're reading has three. That's three to run the test on, and it is not a score.
What if there are no reviews at all?
Then you haven't learned anything about how other brands' asks came back, which is different from having learned something bad.
The choosing guide says of an empty branded content history: "An empty branded content history is something you don't know, not something you've learned. It isn't a mark against anyone." That judgment is about branded content, not reviews. This article applies the same limit to a blank review section because both records depend on somebody having created them, and their absence doesn't reveal why they are missing.
The same answer in that guide also limits what the blank costs you. "What stays unknown is narrow", it says of the empty history, and it is narrow here too. What you don't know is how other brands' asks came back. That was never what decided your shortlist, and everything that does decide it is still sitting there.
For the clip: the next name on your shortlist has a blank where this one had three, and you have the same work ahead of you on both.
Frequently asked questions
One of the reviews is negative. Is that a reason to drop them?
Not on its own. Run the same test on it: a negative review with both halves in it has settled one job, which is what a readable positive one settles too.
What a negative one doesn't carry is the reason the brand wrote it. One competitor's creator training on building a review history lists what draws one, and the list is mostly about how somebody conducted themselves on the job — "being unresponsive with brands" is on it, and so is "posting low-quality content", which is the one item on the list about the video rather than the conduct. That list is what the training tells creators to avoid. It is not a count of why brands write negative reviews, and nothing here has measured that.
Two of the reviews don't agree with each other. Which one do you go with?
Don't choose between them before checking what each review describes. Identify the ask, the result, and any context each one supplies, then decide how either account relates to your campaign. They may describe the same job or different jobs; even reviews from different jobs can carry conflicting signals, and a profile may not give you enough context to resolve them.