Need Advice Should I Use my cartoon face (part of my branding) on thumbnails as a web designer?

Zanet

New User
TubeBuddy Legend
40
7
I recently ran some A/B tests and because i have low traffic, I'm not convinced there is enough substance

Also I took advice from a live channel audit review on VIDIQ last week where they suggested not having my face on thumbnails but trying something more relevent

All A/B tests are proving different to there advice (though the stats are based on low traffic)

Its hard to explain what i mean without being able to post my channel, but i'm really confused at what point i drop the design and common sense of thumbs for CTR facts and how much traffic i need before the stats are a fair representation?
 

Andrew

Superman
Administrator
5,921
31
youtube.com
Subscriber Goal
5000
I recently ran some A/B tests and because i have low traffic, I'm not convinced there is enough substance

Also I took advice from a live channel audit review on . last week where they suggested not having my face on thumbnails but trying something more relevent

All A/B tests are proving different to there advice (though the stats are based on low traffic)

Its hard to explain what i mean without being able to post my channel, but i'm really confused at what point i drop the design and common sense of thumbs for CTR facts and how much traffic i need before the stats are a fair representation?
Can you post the before and after thumbnail and tell us the difference? Also you can post your link for understanding's sake, but we also need both thumbnails too if possible for the full picture!
 
OP
OP
Zanet

Zanet

New User
TubeBuddy Legend
40
7
Hi Andrew

here are the details - so this is what i was doing - designing every icon with me in (Juian Opie style) clear headline - strong branding etc - what you might expect from a designer - but- see after image


screen1.jpg


So with the above thumbnails it build a kind of consistent brand and for subscribers they will recognise my thumbe - but i only have 80 subscribers, im yet to be discovered although im puttin out useful content (I know everyone says that) though i am giving a value with my 20 years experience.

I went on a different competitor review live channel ans was picked - can i say vid Iq?
They do some nice live reviews on there but suggested my thumbs dont really support the reality of my actual subject

So obviously A/B testing would make sense
in one day i totally redesigned every thumbnail and put them on A/B Testing - this is where i started using the excellent TubeBuddy tools

But the results are a little unusual

See screenshot

screen2.jpg



So i'm left with several questions - which was what i tried to explain on the live discussion last night - but hopefully the screenshot explain it better


Q1. Is a lot of this irreverent because with 80 subs and little traffic any data isnt that relevant?

Q2. looking at the A/B stats - everything is up except the CTR . Is CTR the ultimate decision or could all the other stats supersede the CTR if you look at my stats above you can see why i ask the question

Q3. if i keep some thumbs that say my face logo is a better CTR yet other thumbnails say the non-faced version is better - ultimately it weekends the brand. if someone comes to my channel they could say wow this is consistent - i expect that from a designer, or if i allow only top preforming thumbs to remain and have more of a mixed bag. should i appeal to search results with my thumb or satisfying subscribers?

Thanks Andrew in advance for your great work - if you feel this might make a good subject for a video - you are welcome to use my channel and screenshots
 

Andrew

Superman
Administrator
5,921
31
youtube.com
Subscriber Goal
5000
o i'm left with several questions - which was what i tried to explain on the live discussion last night - but hopefully the screenshot explain it better


Q1. Is a lot of this irreverent because with 80 subs and little traffic any data isnt that relevant?

Q2. looking at the A/B stats - everything is up except the CTR . Is CTR the ultimate decision or could all the other stats supersede the CTR if you look at my stats above you can see why i ask the question

Q3. if i keep some thumbs that say my face logo is a better CTR yet other thumbnails say the non-faced version is better - ultimately it weekends the brand. if someone comes to my channel they could say wow this is consistent - i expect that from a designer, or if i allow only top preforming thumbs to remain and have more of a mixed bag. should i appeal to search results with my thumb or satisfying subscribers?

Thanks Andrew in advance for your great work - if you feel this might make a good subject for a video - you are welcome to use my channel and screenshots
Q1: Depends if you're running for a set time or until statistical significance. I'd say run until you know for sure based on your size.

Q2: CTR Usually does matter, but in your case at your size, you don't have enough data, going back to my first point. CTR usually is the best case, and tends to match the others. However, When did you upload this vs when did you start the A/B test at the same time or different can help here!

Q3: I personally like a combination of the two ideas, keep your face in them and show a bit more of the actual solution you are showing off! Keep in mind your job as a YTer is to appeal both to your current subscribers AND People who don't know you, and finding that point is the sweet spot. This kind of data takes time, and I'd recommend running the tests in the background, as you grow out your content :)
 
OP
OP
Zanet

Zanet

New User
TubeBuddy Legend
40
7
THanks Andrew
My replies below in blue and additional comments


Q1: Depends if you're running for a set time or until statistical significance. I'd say run until you know for sure based on your size.
I sent them all for running on statistical significance - is that better to do?

Q2: CTR Usually does matter, but in your case at your size, you don't have enough data, going back to my first point. CTR usually is the best case, and tends to match the others. However, When did you upload this vs when did you start the A/B test at the same time or different can help here!
THe A/B and thumbnail change was made at the exact same time a week ago on all my thumbnails - is that ok?

Q3: I personally like a combination of the two ideas, keep your face in them and show a bit more of the actual solution you are showing off! Keep in mind your job as a YTer is to appeal both to your current subscribers AND People who don't know you, and finding that point is the sweet spot. This kind of data takes time, and I'd recommend running the tests in the background, as you grow out your content :)
Yes that exactly what i thought - its about finding the sweet spot and balance between the 2 - for now i dont think i can afford to worry about 80 subs as much as being discovered

From what im picking up (please correct me if im wrong)
Your advice would be to stick with letting the CTR dictate the decision and just keep A/B testing and run it until statistical significance
 

Andrew

Superman
Administrator
5,921
31
youtube.com
Subscriber Goal
5000
THanks Andrew
My replies below in blue and additional comments


Q1: Depends if you're running for a set time or until statistical significance. I'd say run until you know for sure based on your size.
I sent them all for running on statistical significance - is that better to do?

Q2: CTR Usually does matter, but in your case at your size, you don't have enough data, going back to my first point. CTR usually is the best case, and tends to match the others. However, When did you upload this vs when did you start the A/B test at the same time or different can help here!
THe A/B and thumbnail change was made at the exact same time a week ago on all my thumbnails - is that ok?

Q3: I personally like a combination of the two ideas, keep your face in them and show a bit more of the actual solution you are showing off! Keep in mind your job as a YTer is to appeal both to your current subscribers AND People who don't know you, and finding that point is the sweet spot. This kind of data takes time, and I'd recommend running the tests in the background, as you grow out your content :)
Yes that exactly what i thought - its about finding the sweet spot and balance between the 2 - for now i dont think i can afford to worry about 80 subs as much as being discovered

From what im picking up (please correct me if im wrong)
Your advice would be to stick with letting the CTR dictate the decision and just keep A/B testing and run it until statistical significance
I would set them until statistical significance, while you build up your catalog of content, since you don't know what will work yet. I prefer statistical significance if you don't know, but day period if you know your audience will react well to it.