I have hired eleven thumbnail designers over the years and fired seven of them. Not because they were bad at Photoshop. Most of them were better at Photoshop than I am. They got cut because the thumbnails did not move the click rate, and after a while I could not explain to myself why I was still paying.
That experience taught me something uncomfortable. The signals that predict a bad hire show up in the first conversation, long before the first file lands. You just have to know what you are listening for. This post is the list I use now, on the studio side and on the hiring side, written down honestly.
The portfolio is beautiful and none of it is a thumbnail
This is the most common one and the easiest to miss. Someone sends a portfolio full of gorgeous poster art, album covers, brand identities, a couple of app screens. It is genuinely good work. Then you scroll and realise there are maybe three actual YouTube thumbnails in there, and they are for channels you have never heard of with no view counts attached.
Thumbnail design is not general graphic design at a smaller size. A poster is viewed at arm's length for ten seconds. A thumbnail is viewed at 210 pixels wide, in a grid with eleven competitors, for roughly a third of a second, by someone who is half deciding whether to go to bed instead. Those are different jobs. A designer who has spent five years on packaging has learned habits that actively hurt in the feed: delicate type, subtle contrast, negative space used for elegance rather than for separating a face from a background.
When I look at a portfolio now, I do not ask whether the work is nice. I ask whether it survives being shrunk. Open any candidate's portfolio, drop the images to thumbnail size on your phone, and see what is left. If the answer is a blur with a mood, that designer is going to cost you money. Most of the craft here lives in contrast and readability at small sizes, and it is a specific skill, not a general one.
They never ask what the video is about
I want to be blunt about this one because it is the clearest tell in the whole list.
If a designer takes your brief, says "got it", and disappears for two days without asking a single question about the video, they are not designing a thumbnail. They are decorating a rectangle. The thumbnail is a promise about the content. You cannot make a promise about content you have not asked about.
The questions I expect to hear from anyone competent: what actually happens in this video, what is the surprising part, who is the viewer, what does the title say, what has worked on this channel before. That last one matters more than people think. A channel has a visual history, and the audience has been trained by it.
A designer who does not ask about the video is guessing. Guessing is fine once. Paying for guessing eight times in a row is a subscription to noise.
The flip side is a real warning sign too. Some designers ask twenty questions and then ignore every answer. The brief goes in one ear. You can only catch that on the first delivery, which is why the first delivery matters more than the interview.
Revisions are capped, and the cap is low
Two or three revisions sounds reasonable when you read it in a proposal. It stops sounding reasonable the first time you are on round three, the thumbnail is close but the eyeline is wrong and the text is competing with the face, and you now have to choose between paying extra or shipping something you do not believe in.
Here is the thing about revision caps. They are not really about protecting the designer's time. They are about hiding the fact that the designer does not expect to get it right through iteration, so they would rather cap the exposure. Good thumbnail work is iterative by nature. You put a version up, you look at it small, you kill the third element, you push the contrast, you try the face two hundred pixels left. That is the job.
We run unlimited revisions on every pack for exactly this reason. One pack is $250 for 8 custom thumbnails, roughly $31 each, one per video, and we will keep going until the thing works. That is not generosity. It is that a capped thumbnail is usually a worse thumbnail, and worse thumbnails make me look bad.
A related tell: ask how they handle a thumbnail that ships and underperforms. If the answer is silence, or "that would be a new order", you have learned something. The feedback cycle after publish is where most of the learning is, and I wrote about how that feedback loop between designer and creator actually works when it is healthy.
Everything they show you looks like the same file
Scroll a portfolio and count the layouts. Face left, text right, yellow arrow, red circle. Face left, text right, yellow arrow, red circle. Twelve times.
Template work is not automatically bad. A consistent visual identity across a channel is a real asset, and some very large channels run a near fixed layout on purpose. The problem is when the template is the designer's, not the channel's. That means every client they have gets the same skeleton with different photos dropped in, and your channel ends up looking like four other channels in the same niche, which is the exact opposite of what you are paying for.
We build every thumbnail custom per video and we do not run templates. That is slower. It is also the only way to make a thumbnail that answers this specific video rather than the general category of video. It is also why we test new styles on our own channels first, before any of it reaches a client. I would rather burn my own click rate learning something than bill someone for the experiment.
If you want to see the range of approaches rather than one repeated skeleton, our work is the honest version of this test. Judge it the same way I am telling you to judge everyone else.
They talk about aesthetics and never about click rate
Ask a candidate what makes a thumbnail work. If the whole answer is composition, colour theory, and visual hierarchy, they are half a designer. Those things matter. They are not the point.
The point is that a person scanning a feed makes a fast, mostly emotional judgment about whether this video is worth their next eleven minutes. Everything in the frame is either serving that judgment or stealing attention from it. A designer who cannot talk about curiosity, stakes, contrast against neighbouring thumbnails, or the relationship between the title and the image is going to give you pretty work that does nothing.
I will say the honest version of the numbers here, because I dislike how this gets sold. On a finance channel around 80k subs, we saw click rate move from roughly 4 percent to just under 6 percent across a batch of eight, mostly by cutting text down to three words and pushing face size up. That is a real result on a real channel with a specific audience and a specific back catalogue. It is not a law. On a different channel in a different niche, the same changes did close to nothing, and one video got worse. Anyone quoting you a guaranteed percentage lift has stopped being a designer and started being a salesperson.
Some formats are harder than others, and honesty about that is part of the job. Two people at a table is a genuinely difficult image to make interesting, which is most of why podcast thumbnails underperform unless you break the format. Corporate subject matter has its own trap, and getting B2B thumbnails that do not look corporate is mostly a matter of refusing the stock visual language the category expects.
The pricing has no shape
Three flavours of pricing trouble, all worth walking away from.
The first is the vague hourly quote with no ceiling. Thumbnails are not a good hourly product because the value is in the judgment, not the hours, and you will spend the whole engagement quietly resenting the clock.
The second is suspiciously cheap. Five dollars a thumbnail is a real market and it exists for a reason, but at that price nobody is watching your video, asking about your audience, or iterating. You are buying a stock photo with text on it. Sometimes that is genuinely all you need, and I have written honestly about when you do not need a designer, because for a lot of channels at a lot of stages, the answer is that you do not.
The third is a huge retainer for undefined output. If nobody can tell you how many thumbnails you get this month, you will find out at month three, and you will not like the number.
What you want is a shape you can hold in your head. Ours is one pack, $250 for 8 thumbnails, unlimited revisions, dedicated designer, title consultation included, 48 to 72 hours per thumbnail. Add-ons are small and named: an A/B variant is $15, a title pack is $25, rush is $20. If you are weighing a recurring arrangement against buying as you go, the tradeoffs in subscription versus one off thumbnail design are worth reading before you sign anything.
They cannot work inside your process
The last red flag is the least glamorous and the one that kills the most relationships. A designer can be excellent and still be impossible to work with because they do not fit how your team actually operates.
Signs to watch: no clear place where files land, feedback given and received in four different apps, no naming convention, no version history, nothing written down. It works fine for one video a month. At three videos a week with an editor, a producer, and you in the loop, it collapses. If you are past the solo stage, get the pipeline right before you get the designer right, and building a thumbnail workflow for a team covers the version of that which has survived contact with real teams.
We are Dubai based and work async worldwide, which forces this to be tight. There is no hallway to solve things in. Everything has to survive being written down. Honestly that has made the work better, not worse.
The same discipline applies on the video side. Our sister service handles video podcast editing at Pixel Samy, from $199 with clips at $25, and the projects that go smoothly are always the ones where the handoff between edit and thumbnail is a defined step rather than a message that says "can you make something for this".
What I would actually do
Send one paid test. Not free, paid, and not eight of them. One thumbnail for a video you are genuinely publishing. Watch what they ask, watch how long it takes, watch what happens when you say the first version is not it.
You will learn more in that one round than in any portfolio review. Half the red flags in this post only become visible when there is a real deadline and real feedback in the room.
And if the test goes badly, walk away early. The sunk cost on a designer who almost works is the most expensive thing in this whole process. I have paid it seven times. Send me the portfolio if you want a second opinion before you commit, and I will tell you honestly whether I think it holds up small.