Reviewing YOUR Retention Graphs (#1)
Understanding why people are clicking away from your videos.
Hey all,
Last week I wrote about 5 scriptwriting mistakes I’ve made over the last year that affected retention on videos I’ve worked on.
The newsletter had a pretty low open rate and minimal interaction, but I converted it into a tweet thread anyway, not expecting much.
It was my most successful thread ever:
During that thread, I asked you to send me your retention curves to review.
And having gone from 888 followers to over 1100 in 24 hours, I had a lot of submissions :)
This week, I’m going to review 5 and give my thoughts on why retention may have dropped.
Huge props to everyone that submitted and gave their permission for me to give some critical feedback in public. That takes some bravery!
Keep in mind, I’m not an expert in every niche I’m reviewing here, but having written dozens of scripts for multiple creators, I’m giving you my best bet as to what I think might have turned viewers off.
Lets go!
1 - Low Stakes
Title: I Spent 48 Hours In Croatia’s Most Expensive City With Only $100
Creator: Ollie Kirk
Average % viewed: 56.9%
Problem: Stakes seem low (and a little arbitrary)
This video was great fun, but the main issue was that there was no real point of tension.
Trying to spend ‘only’ $100 across 48 hours just doesn’t seem that difficult, and he never seems close to failing.
The video is well-made, and Ollie has clear YouTube ‘instincts’…
But think about whether the concept will create intrigue.
My reaction to I Survived on $0.01 for 30 Days → “what, but how?!”
My reaction to this → “I reckon I could manage that”
So my question is…do people actually want Ollie’s videos to have artificially inflated stakes?
To me, the fact that he’s travelling around the world makes the videos interesting enough without making them ‘challenges’.
Instead, could there be more of a focus on storytelling? He created a great ‘character’ out of the zipline guy and almost did the same with the ice cream guy.
Are any of Ollie’s colleagues going to appear regularly who we could get to know a little better?
Are there any little 3-act stories which emerge during the day-to-day?
That’s the video I’d like to see.
2 - Confusing Stakes
Title: If My Wife Wins In Overwatch 2, I Get A TATTOO!
Creator: Catfluent
Average % viewed: 25.4%
Problem: After setting up amazing stakes, the creator makes them confusing.
From low stakes, to confusing ones.
The hook is brilliant 😂 It really shows Catfluent’s personality and sets clear stakes.
The two things I know are:
If she wins the game, he has to get a tattoo.
He doesn’t want this to happen.
But then something confusing happens.
We cut to the gameplay, where his wife is now (theoretically) moments from causing her husband to have to get a tattorr…but he actually seems to be supporting her while she plays!
He’s giving her tips and cheering when she does well.
This affects the tension because we no longer know how he actually feels about having to get the tattoo - seemingly he now doesn’t actually mind too much.
This video has small number of views for now which is why it’s particularly spiky, but the overall downward trend can probably be attributed to the confusing stakes.
3 - Dry Visuals
Title: THE MOST META COUNTER-ATTACKING CUSTOM TACTIC! FIFA 22
Creator: ManLikeRomz
Average % viewed: 36.7%
Problem: Laboured Hook + Dry Visuals
My immediate reaction here is that the hook gives a generous dose of Romz’s personality (which is great!) but we have to wait 50 seconds before we get into the ‘meat’. The first spike shows people are skipping this.
The rest of the graph is like a ski slope. Flat for a while, then a sharp drop, then flat again (etc, etc).
Every sharp drop seems to coincide with when Romz is detailing the specific attributes and play styles the audience needs to assign to their players.
The problem is that this is obviously central to the video.
BUT…
It’s visually flat.
While Romz is giving the information, he could try including some B-Roll from the actual matches where the players demonstrate the attributes he’s describing.
Failing that, some dynamic zooms on the areas of the HUD which are being talked about would just keep the video in motion a little more.
We just need a little something to keep the pace moving and to keep the audience engaged while the information is being delivered.
A Brief Interlude…
🔧 My YouTube Script Template
While no two scripts are the same, it always helps my flow to have a clean space to write and repurpose videos. And it certainly helps to have little reminders about what I’m trying to achieve at each point in the script.
So if you’d like the template I use to write for YouTubers, click here. It’s free!
Back To It…
4 - Perceived Credibility
Title: You're Not A Real Man Until You Unf*ck Your Testosterone Levels
Creator: Jak Lei
Average % viewed: 27.8%
Problem: Perceived lack of credibility
The intro to this video has a few things which could have contributed to the dropoff.
Ignore the spike - it’s a cheeky glitch.
The first shot is handheld which adds a nice personal touch to the video, and sets up a super intriguing premise. Mentioning “balls” right out the gate had my attention! 😂
But I think the first minute features a few too many news clips. We only need a couple of these to give context to the discussion, and the B-Roll overstays its welcome.
It almost feels like it’s padding the video to extend the runtime, rather than spending more time with Jak delivering new information.
Then, by around 3 minutes, we’re at 33% retention and what turned me off can be summed up in two words:
Tucker Carlson 😂
Including B-Roll of Carlson in a discussion about science just doesn’t lend the video credibility, especially when we’re deciding whether to settle in for a 13-minute video about this topic.
I’d have liked to have spent a little more time with Jak, and less with the B-Roll.
5 - Lack of Context
Title: Can 3 Idiots Beat 1 Gang Beasts God?
Creator: DookyFace
Average % viewed: 18.5%
Problem: Intro doesn’t provide context
This video is wild. I want to play this game immediately 😂
The main problem here is a sharp drop-off at the start.
The issue is that the video doesn’t provide any context as to what’s going on.
It seems like we’re being shown a couple of high-energy highlights as a sort of “COMING UP” segment.
But then this goes on for 40s. It turns out we’re watching a highlight reel and it’s already started.
Moreover, the audience is never directly addressed. It feels like we’re a bit outside the joke.
Once I got past that first minute, I was starting to have a little more fun by osmosis. Seemingly, everyone who got that far pretty much stayed til the end too.
But remember, if the audience is confused, they’ll leave. And just a little more interaction with the viewer early on could have helped this.
That’s all for this week!
Thanks to everyone who sent in their retention graphs. There are LOADS more I have to do, so if yours didn’t come up this week then I promise I’ll get around to it soon.
Enjoy your weekend,
George
Thank you for the valuable inputs George. It gives me insights to make the next one even better.