Latest / Video Marketing with Fexingo: Short Form, Long Form, and Video-First Strategy / Why Video Marketers Should Use the 3-Second Rule for Thumbnails
Transcript
- Lucas: Let me start with a number that stopped me this week: a mid-sized SaaS company called Planhat — they do customer success software — ran a thumbnail test on a single explainer video. They tried 51 variations over three weeks. The winner? A close-up of a smiling woman’s face with no text overlay. It beat every other thumbnail by at least 34 percent in click-through rate. Luna: Thirty-four percent just from swapping the thumbnail? That’s wild. But also — no text? I feel like most thumbnails I see are cluttered with big yellow words. Lucas: Right, and that’s exactly the point. We’ve been trained to think thumbnails need to scream the value proposition. But the research — and Planhat’s test is just one example — suggests the brain processes faces faster than text. You have about three seconds, maybe less, before someone scrolls past. A face with a clear emotion communicates the tone of the video instantly. Luna: So the three-second rule — is that like a hard cutoff? Or more of a loose guideline? Lucas: It’s based on eye-tracking studies from platforms like YouTube and Instagram. Researchers found that users decide whether to click or scroll within the first 1.7 to 2.5 seconds of seeing a thumbnail. So three seconds is generous. The key is that the thumbnail needs to be decodable in that window. A face works because we’re wired to recognize expressions. Text requires reading, which takes longer — especially on mobile where the thumbnail is small. Luna: That makes sense. But what about when the video is about a complex topic? Say, a tutorial on data analytics. Wouldn’t text help set expectations? Lucas: It can, but the trick is hierarchy. If you use text, make it one or two words max — and put it near the face, not covering it. Planhat’s test also had variations with text like 'CRM Tips' or 'Reduce Churn.' Those did okay, but none beat the pure face. The lesson is: the face is the hook. Text can be a secondary cue, but don’t let it dominate. Luna: I’m curious about the emotion they chose. Was it always smiling? Or did they test different expressions? Lucas: They tested smiling, surprised, and neutral. The smiling face won by a wide margin. But there’s nuance — for a video about a serious topic like 'How to Handle Customer Complaints,' a empathetic or concerned expression might outperform smiley. The key is congruence between the face’s emotion and the video’s content. If there’s a mismatch, viewers feel tricked and bounce fast. Luna: That makes me think of the 'curiosity gap' — like when a thumbnail shows someone mid-laugh but the title is about a tragedy. That feels manipulative. Lucas: Exactly. And platforms are getting better at detecting that. YouTube’s algorithm can penalize low retention even if the CTR is high. So the thumbnail’s job isn’t just to get the click — it’s to attract the right viewer who will actually watch. That’s where color contrast comes in too. Planhat’s winning thumbnail used a bright orange background — high contrast against the woman’s skin tone. That pops in the sidebar. Luna: So the color is strategic, not just aesthetic. I’ve seen studies that say red and yellow get more clicks, but it depends on the brand palette, right? Lucas: Right. Red tends to create urgency, but if your brand is all blue, a red thumbnail might feel disjointed. The more important principle is contrast — not just against the video content, but against the surrounding page. YouTube’s background is white, so a dark thumbnail stands out. On a dark mode feed, a bright one does. So you kind of have to know where your thumbnail will appear most often. Luna: Alright, so if I’m a marketer with a library of videos, how do I start testing this without spending three weeks like Planhat? Lucas: Good news — there are tools that make it faster. ThumbnailTest dot com lets you upload up to four thumbnails and shows you the click-through data within 48 hours. You can also use YouTube’s own A/B testing feature if you have a channel with enough traffic. The key is to test one variable at a time: face versus no face, or one emotion versus another. Don’t test face and color simultaneously or you won’t know what drove the result. Luna: And what sample size do you need to trust the data? I imagine 100 impressions isn’t enough. Lucas: You want at least 1,000 impressions per variation to start seeing statistical significance. Planhat had over 10,000 for each of the 51 thumbnails — that’s a lot of data. But for a smaller channel, even 500 can give you directional insight. Just be aware that the margin of error is larger. And don’t change the title or the video itself during the test — that would muddy the results. Luna: That’s smart. So you isolate the thumbnail. Okay — but what about thumbnails for short-form video? Like Reels or TikTok where the thumbnail is auto-generated? Lucas: That’s a different beast. On TikTok, users often scroll past so fast that the thumbnail barely registers. The hook comes in the first few frames of the video itself. But on YouTube Shorts and Instagram Reels, you can upload a custom thumbnail. The same principles apply, but the screen is smaller — so a single face fills the frame better. Avoid wide shots. Get that face big and centered. Luna: I’ve also seen channels use characters or illustrations instead of real people. Does that work too? Lucas: It can, especially for gaming or animation channels. But for business content, real faces tend to build trust faster. There’s a reason why LinkedIn posts with a profile picture get more engagement. Our brains are wired to connect with human faces. If you’re a solo creator, put your own face in the thumbnail. If you have a team, feature the person who appears in the video. That consistency builds recognition. Luna: Let’s talk about the curiosity gap specifically — because I think a lot of marketers misunderstand it. They think it means hiding information, but really it’s about promising something intriguing without giving it away. Lucas: Yes. A classic example: instead of a thumbnail that says 'How to Save Money,' you show someone looking shocked at their phone screen with the text '$500 mistake.' The viewer thinks, 'What mistake? I might be making it.' That’s the curiosity gap. The emotion on the face — shock, surprise — supports the gap. But again, don’t overpromise. If the video doesn’t deliver on that hook, you’ll get high bounce rates. Luna: Right, because retention is a ranking signal too. So the thumbnail has to be honest. I think that’s where a lot of people go wrong — they optimize for CTR at all costs. Lucas: Exactly. And that’s why I like Planhat’s approach: they also tracked watch time. The winning thumbnail had not only the highest CTR but also above-average retention. So the viewers who clicked were genuinely interested. That’s the sweet spot. You want thumbnails that act as a filter, not just a bait. Luna: I want to get practical for someone listening who’s maybe a solo marketer or a small team. What’s the one thing they should do this week to improve their thumbnails? Lucas: Audit your top 10 videos by impressions but low CTR. Look at those thumbnails. Are they cluttered? Do they have a clear focal point? If not, redesign one or two with a single face and high contrast background. Then run a simple A/B test against the original. Even if you only get 500 impressions per variation, you’ll learn something. The tool I mentioned — ThumbnailTest — has a free tier for up to three tests. Luna: And if the face thumbnail wins, you’ve got a new template. If it doesn’t, you’ve got data to iterate. That’s a low-risk experiment. Lucas: Low risk, high potential upside. And honestly, if that test saves you even a few hours of guesswork down the line, it’s worth it. Which reminds me — if these kinds of conversations have helped you think more strategically about your video marketing, and you’ve found yourself using something we’ve discussed in your actual work, that’s the whole point. And if today’s episode was worth a coffee to you, you know where to find us: buy me a coffee dot com slash fexingo. It’s a tiny gesture that helps us keep this show ad-free and focused on practical, data-backed ideas. Luna: Yeah, totally. It’s a way to say 'this helped' without any pressure. And every bit goes right back into researching episodes like this one. Lucas: Exactly. So back to thumbnails — one more thing I want to touch on. The future of thumbnails might involve dynamic versions that change based on who’s viewing. YouTube has been experimenting with personalized thumbnails, showing different ones to different users based on their watch history. That’s still early, but if it rolls out, the principles we talked about — face, emotion, contrast — will still matter. You just might need multiple versions. Luna: That’s fascinating. So instead of one winner, you’d have a portfolio of thumbnails optimized for different audience segments. That could really level up CTR for channels with diverse content. Lucas: Right. But for now, the basics still dominate. One face, clear emotion, high contrast, minimal text. Test, learn, repeat. That’s the cycle. Luna: Alright, Lucas — next episode, I want to dig into the science of video titles. How they interact with thumbnails. Because I bet there’s a lot of overlap. Lucas: That’s a perfect follow-up. The title and thumbnail work as a unit. A great thumbnail with a weak title won’t perform, and vice versa. We’ll unpack that next time.