A creator emailed me last spring with 41 thumbnails he had paid for over four months. Same agency, same designer, same three layouts recycled with a different face pasted in. His click through rate had moved 0.2 points in the wrong direction. He was not scammed. Nobody lied to him. He just hired the wrong kind of shop and did not know what to look for until the money was gone.
That is the annoying part about hiring in this category. The bad outcomes rarely look like fraud. They look like competence that quietly does not work. So instead of a generic warning list, here is what I actually watch for, why each one predicts a bad six months, and the exact question that flushes it out before you pay anyone.
Why do most packaging hires fail quietly instead of loudly?
Because the deliverable always arrives. You asked for thumbnails, you got thumbnails, the files open, the resolution is fine. There is no obvious moment where you can say this is broken. The failure only shows up in aggregate, three months later, when your click through rate has not moved and you cannot point at a single file that was the problem.
The channels that get burned worst are usually the ones publishing consistently. They keep feeding the pipeline, they keep paying, and the compounding loss is invisible because there is no counterfactual sitting next to it. If you want the fuller picture of what this work is supposed to include, I broke it down in what a packaging agency actually does.
Red flag one: they will not touch your title
This is the fastest disqualifier I know. If an agency treats the title as your problem and the image as theirs, they are not doing packaging. They are doing decoration.
Thumbnail and title are one unit. The image should not repeat what the title already says, and the title should not repeat what the image already shows. When they duplicate each other you have spent two slots to deliver one idea. When they work together you get a small gap that the viewer resolves by clicking. That gap is the entire mechanism.
I have watched a decent thumbnail lose to a worse one purely because the worse one had a title that completed it instead of echoing it. So when a shop says titles are outside scope, what they are really telling you is that they have never tested the pair. Ask them directly: what would you change about my last three titles. If they have an opinion in ten seconds, good sign. If they go quiet, you have your answer.
Red flag two: a portfolio that looks like one video
Scroll their work. If forty thumbnails share the same arrow, the same glow, the same three quarter face crop and the same two font weights, you are looking at a template shop wearing an agency label.
Templates are not evil in themselves. They are efficient. The problem is that your channel then looks like every other channel that hired the same shop, and YouTube's home page is a comparison surface. You are not competing against a blank canvas, you are competing against the eleven other tiles around you. Sameness is the one thing that guarantees you lose that comparison.
If I can guess which agency made a thumbnail without being told, that thumbnail is already working against the channel it belongs to.
The honest version of a portfolio shows range. Different niches solved differently. A cooking channel should not be packaged like a finance channel, and neither should look like a documentary. Our own work is on the portfolio page precisely so you can check that range before you talk to anyone.
Red flag three: results you cannot verify
Screenshots of analytics with the channel name cropped out. Percentage lifts with no baseline. Case studies about creators who mysteriously cannot be named but also apparently tripled overnight. Treat all of it as decoration.
Here is my uncomfortable position on this. I never show client names either, because creators do not want their audience knowing the packaging is outsourced, and that discretion is part of what they pay for. So I describe work by niche and subscriber band, something like a finance channel around 80k subs, and I do not attach numbers I cannot prove. The difference between discretion and invention is whether the claims get bigger when the names get smaller.
If an agency shows you a 340 percent lift, ask what the sample was, over how many videos, and what else changed in that window. A publishing schedule change, a shorts push, a video that got picked up by a subreddit, all of these move click through rate more than a thumbnail does. Anyone serious will happily tell you what they cannot isolate.
Red flag four: revision policy written like a legal shield
Two revisions included. Additional rounds billed at hourly. Major changes constitute a new brief. Every one of those clauses exists because the shop expects to argue with you.
Think about what a revision actually is. It is you telling me the idea did not land. That is information, not an inconvenience. The rounds where a concept gets torn down and rebuilt are usually where the good version appears, and a policy that taxes those rounds is a policy that pushes both of us toward accepting the mediocre second draft.
We run unlimited revisions on every thumbnail for exactly this reason. It costs us more on some projects. It also means nobody has ever shipped a file they were lukewarm about because the meter was running. If you want the full walkthrough of how that process is structured, it is in our packaging process.
How fast should a packaging agency actually be?
Fast enough that packaging never delays a publish, slow enough that somebody thought about it. In practice that means about 48 to 72 hours per thumbnail, which is what we run, with rush available at 20 dollars when a video moves up.
Be suspicious in both directions. A four hour turnaround usually means a template got recolored, because concepting, testing two or three directions, and cutting the weak ones takes longer than that. A two week turnaround means you are in a queue behind bigger accounts and your video will ship with whatever you made yourself at midnight.
Ask what happens when you need something moved. If the answer involves account managers and change requests, you are buying a process, not a designer. The people who work well with creators answer their own messages.
Red flag five: no interest in your back catalogue
This one is subtle and it tells you a lot. An agency that only wants to talk about your upcoming videos is optimizing for their own recurring invoice, not for your channel.
Your old videos are the cheapest growth you have. They are already indexed, already have watch time history, and already carry whatever authority the algorithm assigned them. Repackaging a video from eighteen months ago costs one thumbnail and can pull traffic for years, and it very often beats spending the same money on packaging for a new video that has to earn its impressions from zero.
I ask about the back catalogue on every call because it is usually where the first month of work should go. Six old videos, new packaging, watch what happens over the following four weeks. If somebody never brings this up, they either have not thought about it or they would rather sell you a subscription. Google's own documentation on how YouTube ranks and recommends videos is fairly clear that impressions and click behaviour keep being reassessed, which is why an old video is not a closed case.
What questions actually expose a bad fit in one call?
Six of them, and none require you to know design. I would ask these in this order.
- What would you change about my last three titles, and why those specifically.
- Show me two thumbnails you made for channels in different niches. What was different about the thinking.
- What is the revision policy, and what happens on round four.
- Which of my old videos would you repackage first.
- Who is actually making the file, and will I talk to them directly.
- What have you tested on your own channels.
That last one matters more than it sounds. Every style we offer got tried on our own uploads before it went near a client's channel, which means when I recommend something I am telling you what happened, not what I assume. A shop with no skin in the game is guessing with your impressions.
If you would rather evaluate an individual than a company, the tradeoffs are laid out in this piece on hiring a packaging expert. Neither option is automatically better. They fail in different ways.
Red flag six: pricing you cannot divide, and briefs about colours
Monthly retainers with vague deliverable counts are the easiest place for value to disappear. If you cannot divide the invoice by the number of thumbnails you received, you cannot tell whether you are being served or subscribed.
Our numbers are deliberately boring to calculate. 250 dollars for 8 custom thumbnails, which is about 31 dollars each, one per video, with a dedicated designer, title consultation included and unlimited revisions. Add ons are priced the same way: an A and B variant at 15 dollars, a title pack at 25, rush at 20. You can do that arithmetic in your head, which is the point. The full breakdown lives on our pricing and services page.
Also check what happens if you pause. Anything with a cancellation window longer than a month is designed for the agency's forecasting, not your publishing schedule.
A brief that only covers colours and logo files will produce a thumbnail that only works on colours and logo files. The questions that matter are about the audience. Who is this for, what do they already believe, what would make them stop, and what have they seen eleven times already this week.
This is where niche experience actually earns its keep. A thumbnail for a channel selling to business owners has to do something different from one aimed at hobbyists, because the viewer's suspicion level is different. That gap is why I wrote separately about thumbnail design for coaches and consultants, where the packaging problem is credibility rather than curiosity.
We work from Dubai, async, with creators across most time zones, and the intake conversation is almost entirely about audience rather than aesthetics. If nobody asks you those questions before quoting, they are pricing a file, not a result.
They hire on portfolio beauty instead of thinking. A pretty portfolio proves someone can execute. It does not prove they can decide, and packaging is a decision problem far more than a rendering problem.
So weight the call more than the gallery. Somebody who explains why the arrow is there, why the face is cropped at that angle, and why the title stops where it does is going to be more useful to you over a year than somebody with a slicker Behance. If you want a sense of how we think about creator economies outside the usual hubs, there is a version of this argument in our notes on working with a packaging agency from Chile.
If you have been through one of these hires badly and want a second opinion on what went wrong, email samy@pixelsamy.com and I will tell you honestly whether packaging is even your bottleneck.