How long do you have to hook an audience with a trailer? I wish I had an exact answer, but 5-10 seconds is an oft repeated figure from sources like Steam Dev Days. It doesn't sound like a lot of time to work with, but count to 10; think about how much you can know about a video in that time.
10 seconds is more than enough time for:
"HEY GUYS WHAT'S UP, WELCOME TO ANOTHER EXCITING VIDEO..."
"In this video, we're going to explore this world... You will learn about the things in it, the people, and many other exciting things..."
[Extended logo animation]
How do people decide whether or not to continue watching after so little time, and how do you use that knowledge to get keep them engaged?
I think the crux of why people stop watching comes down to:
When I say "Mediocrity" I just mean the first few seconds of the trailer are not striking in any particular way. Whether it's the art style, narration, title card language, or gameplay mechanics, there's just a sense that what you're seeing is not unique in any way. If this is the case, the only way you can expect the audience to continue watching is if they're a captive audience in a theater or livestream event, or they're a loyal fan who will give you the benefit of the doubt.
One easy way to fall into a mediocrity trap (even if there are unique things about your game!) is when you have a game similar to others, but you only use focus on the qualities that overlap instead of the unique ones.
For example, look at these descriptors:
• Procedurally generated levels
• Permanent Death
• Power Ups & Upgrades
• High Difficulty
These can be used to describe the following games:
The Binding of Isaac
FTL: Faster Than Light
Crypt of the Necrodancer
Enter the Gungeon
If the trailer for either of those games started with those descriptors, at best an informed audience would continue watching to see what is unique, but at worst they would stop watching because it all feels too generic or familiar.
Previous Experience is also related to familiarity. This is when a video evokes a feeling similar to something you've seen before which you didn't like. Think about every time you looked for a tutorial on YouTube for a very simple thing, but the one you found started off with a guy enthusiastically saying: "HEY GUYS WHAT'S UP, MY NAME IS [PERSON'S NAME] AND I HAVE AN AMAZING TUTORIAL FOR YOU TODAY." Then the person proceeds to tell you about their day or something totally unrelated to the tutorial, you scrub through to find where it actually begins, but you can't find it, and you go off to search for another one.
Trailers can fall into traps like this especially in movies that have very similar plots to ones which came before it. No trailers are made in isolation, so it's important to consider what else is out there, and how it compares favorably or unfavorably.
Lastly is Pattern Recognition.
We are very good at picking out patterns, and we also crave variety. Trailers can fall into patterns very easily if they are structured like bullet point lists
UNIQUE ENVIRONMENTS! [Montage of environments]
DOZENS OF WEAPONS! [Montage of weapons]
6 BOSS BATTLES! [Montage of boss battles]
After the second bullet point, the audience will probably understand where things are going, but by the third they'll be very sure. And if the editing, music or art aren't interesting enough, it becomes rote very quickly, and if nothing else about the game holds their attention, you have nothing left.
Even editing styles can create patterns. For me, a common sign of a less experienced editor is a reliance on cross fade transitions or fades to and from black. This is because fades are kind of like a palate cleanser for the eyes which give permission to show something completely unrelated to the previous shot. Whereas cuts generally demand some sort of direct association with the previous shot (which is more difficult to do). So when I see a trailer where all the transitions are fades, this gets tiresome to my eyes, and I lose interest quickly.
There's nothing inherently wrong with patterns and repetition, but what you're repeating has to be interesting for the audience to crave more, whether it's via the editing, gameplay mechanics, art or music.
The audience craves variety in all things:
It's up to the editor to mix and vary these elements as much as possible while still making it coherent enough to understand. Most game trailers especially tend to lack variety in pacing. Just think of watching a person run a marathon; you're going to figure out pretty quickly that what they're doing isn't going to look very different for a long time. But what if they had to leap over obstacles occasionally, they got injured, or someone tried to thwart them. THAT would make everything far more harrowing and engaging.
So to sum up with some more positive language. To combat the things that cause people to tune out:
Start with what is unique
Consider how your trailer compares to others out there