Red Flags has turned into a Power Word now and it implies a lot of meaning and relatability to the audience.
There are Red-flags in every niche and making a video on them positions you as an expert
e.g Red Flags in an Investment Opportunity or Social Media Advice
Powerful + Relatable
A hook that opens on two spoken words, a category plus "red flags", said flat with no verb and no run up, then immediately names the flags in second person as things the viewer already does or fears. In the example the opening words are "Retinol red flags," and what follows is three of the viewer's own fears about retinol delivered back to back.
The phrase arrives pre loaded. The audience already uses "red flags" for dating, hiring and friendships, so bolting it onto retinol or an investment or a used car imports an entire judgement system in two words with nothing to explain. Before the second sentence starts the viewer knows the shape of what is coming: a list, someone is about to get caught, and there is a correct answer at the end of it.
What holds them is the turn into second person. The flags are not abstract faults in an industry, they are the viewer's own behaviour, "You don't use retinol because you're scared of peeling." Being described accurately in the first three seconds is uncomfortable in the way that keeps a thumb still, and because each fear is short and stacked with no pause, the list accelerates instead of explaining.
One thing the transcripts make clear: the flag list is not the payoff, the pivot is. The clip spends its opening naming three fears and then turns hard on "But instead of avoiding retinol, you just need to find the right one," after which the whole video is the green flag counterpart, flipping the same colour language and answering irritation, wrinkles and texture in the order it raised them. So the flags have to be objections you are able to resolve, which is a tighter constraint than simply picking the red flags in your niche.
"[Category] red flags. You [thing the viewer avoids or does] because you're scared of [fear one], or you're scared of [fear two], and you're probably scared of [fear three]."
It runs about six or seven seconds. The label is two words spoken as a title, usually mirrored word for word in the on screen text, with no "here are the" or "let's talk about" in front of it. The camera sits straight on the face for the label, then the fears come fast with no beat between them so they read as one accelerating list rather than three separate points. The word "probably" on the final fear is doing real work, because it tells the viewer you know them well enough to guess the one they did not admit to.
Slots to swap per idea: the category, the behaviour or avoidance being diagnosed, and the two or three fears sitting underneath it.