Faizan's Portfolio

CASE STUDY > 100 THUMBNAILS

Make it Click-Worthy (on purpose)

Team

Adam Wickens

My Role

Thumbnail strategy & design, testing, and creative direction

Duration

2024 - present

CONTEXT

Make it Clickbaity…

For the past year, I’ve been consistently designing YouTube thumbnails that drive clicks and engagement—partnering with Adam Wickens, a Canadian reptile educator with 390k+ subscribers, while also shaping the visual identity of CallHub’s growing YouTube presence.

140+

Thumbnails Made

6.2%

Highest CTR

3M+

Views total

KEY LEARNINGS

#1 The Curiosity Gap

Unlike reels or short-form content, long-form video requires an active decision from the user to click. You don’t need MrBeast’s budget to earn that click, you need controlled tension.

Patterns I use:

Make one clear promise, hide one piece of information (curiosity gap).
Show dominant subject (face, or in this case, reptile) with extreme scale.
Keep a three-value palette (dark / mid / accent) for instant read at small sizes.

FIND THE CURIOSITY GAP BETWEEN THE VIDEO TOPIC AND WHAT YOUR AUDIENCE ENGAGES WITH

FIND THE CURIOSITY GAP BETWEEN THE VIDEO TOPIC AND WHAT YOUR AUDIENCE ENGAGES WITH

FIND THE CURIOSITY GAP BETWEEN THE VIDEO TOPIC AND WHAT YOUR AUDIENCE ENGAGES WITH

Designing the Thumbnail

I exaggerated the hazard (radioactive tint + stronger rim light) so the “look vs. touch” tension is legible in 0.5s. The reptile became the centerpiece, and all other elements subordinate to that read. This also worked for the reptile-breeder niche, as the reptile tends to be the main focus of this genre of videos.

CONCEPT SKETCH TO FINAL THUMBNAIL

The icons at the top were not so obvious to his American audience.

Designer Bias Detected

My first pass leaned on iconography that felt “obvious” to me; but it didn’t register for a younger US audience.

The fix: remove abstract icons, push the subject larger, and clarify the physical gesture. Audience over designer cleverness.

FINAL 2 VERSIONS

#2 Testing Thumbnails in real-time

Let’s take a look at another channel, CallHub. We will be studying the video "2024, Explained". It went on to become the channel’s best performing video (340,000+ views), but for the first two days, it barely cracked 200 views.

We knew topic demand was broad, and the video itself was sticky; so it was the packaging (thumbnail + title) that wasn’t earning the click. We just needed the right audience to see it for the algorithm to notice and start pushing it.

Here's what I did to revive the video and make it go viral.

TRYING OUT COMPOSITIONS TILL IT FEELS RIGHT

No need to over explain the title. People are smart enough to figure it out.

Having too many elements does not help with skimmability

Something’s not quite right...

The video was doing worse than previous videos. The complete opposite of what I assumed would happen.

Something had to be done...

VERSION 2: ADDING MORE PEOPLE AND INVERTING THE COLORS

The signature yellow used by “Explained” was not as legible.

Contrast is not as strong in the light version.

VERSION 3: FINE-TUNING

Added more famous faces to the thumbnail.

Bring the blue and red more out of the thumbnail.

#3 People hate AI (so be sneaky)

This one hurt...

I had used a fully AI creature image for one of my early thumbnails. The topic of the video was an obscure dinosaur species, and I was unable to find a hi-res image of it. So, I resorted to using an AI generated image of one. However, I was not aware of the strong anti-AI sentiment that our audience had, and hence once they caught on, they immediately flagged it and trust dipped… Lesson learnt.

USING AI BACKFIRES

Version 2: AI with finesse

That hurted. But I learnded.

But this was a year ago, and AI has gotten REALLY REALLY GOOD SINCE THEN. In fact, you’d be surprised at how much of the content you consume is actually AI generated. So, I decided to take a second-go at it, but this time, with more care.

VERSION 2: “BRINGING BACK EXTINCT ANIMALS”

AI images work best when supplementing real images.

AI images are often oversaturated. Fine tune them to match the rest of the image

Closing Notes: Why this matters beyond thumbnails

Thumbnails are micro landing pages: one promise, one action, one metric. Running 100+ of them sharpened my instincts for message–market fit, visual hierarchy, and rapid experimentation, skills that I apply directly to my day job as a visual designer. Thumbnail design may not be directly related to UI/Product Design, but the repetition and instant feedback nature of them make them great testing grounds for design ideas. As designers, we often work hard to show our business value (and rightly so). But sometimes, trying something completely different is just as valuable. Instead of relying on pre-defined Figma components, thumbnails force you to go back to the basic fundementals of design: Composition, color and story. As such, my learnings from making thumbnails have led to me to become a better designer overall.

My thumbnail playbook

Before Design:

One-line promise (≤6 words).

What am I hiding to create the gap?

What should the viewer feel in 0.5s?

Design rules:

One focal point, extreme scale.

Eye/hand lines point to the subject.

Three-value palette; avoid mid-tone soup.

Test at 160px and in greyscale.

Title lock-up and subject never fight.

Live Testing:

Ship 2–3 variants with explicit hypotheses.

Rotate for 24–36h; log CTR + impressions and save screenshots (YouTube’s context shifts).

System assets

Reusable LUTs, copy blocks, safe margins, and a naming convention so the team can move fast without drift.

Go Home

View Resume

Contact Me