You sell a small YouTuber three genuinely different thumbnails plus three matching title options for one upcoming video, built to run in the platform’s own thumbnail test, using only frames and photos the creator actually owns — AI speeds up the concepts and background work, and you do the checking, the lettering, and the judgment that makes each option honest and readable on a phone.
Who this fits (and who should skip it)
It fits you if you watch a lot of long-form video in one or two niches, you can tell at a glance why one thumbnail reads clearly at phone size and another turns to mush, and you are willing to set the words yourself in an editor instead of trusting an image generator to spell. The buyers who can actually pay are creators who already upload on a schedule — weekly long-form, most weeks — and already get enough views, ad revenue, sponsors, or product sales that packaging is a real line item for them. Their tell is plain: good videos wearing a plain video frame or a wall of tiny text as the thumbnail, and titles that describe the video instead of giving a reason to click.
Skip it if the channel is brand new, posts rarely, or gets a couple of hundred views a video. A better thumbnail cannot fix that math, and a creator at that size has nothing sensible to pay you with. Skip Shorts-only channels too — most Shorts are found by swiping, not by clicking a thumbnail, so the product barely exists there. Skip anyone who wants a thumbnail that lies about the video: a shocked celebrity face in a video that person never appears in, a money screenshot that never happened, a result the video does not show. That is not a style choice, it is the fastest way to burn a channel’s trust and your name with it.
What you deliver for one video is specific: three finished thumbnails at full size, each built on a different idea — say, a face or person frame, a close-up of the key object, and a clean, text-light version — plus three title options written to pair with them, and one plain line under each pair saying what is different about it. One revision round where the creator can mix and adjust. Files only, delivered in a shared folder. The creator uploads them and runs the test on their own channel; you never need, and should never ask for, their channel password.
Startup cost and time
- Money: $0 to start. A free AI chatbot or image tier for concepts and background drafts, a free editor (Canva’s free tier or Photopea) for layout and lettering, and the creator’s own frames and photos as the raw material. That really is the whole stack. Optional paid image tools run about $19–$39 a month, and thumbnail subscription services charge creators about $29–$69 a month — do not buy either before a client pays you. Those prices are also your competition’s price list, which is worth knowing before you quote.
- Time to be pitch-ready: two evenings. One to build a sample pack (below), one to write your intake and your target list. The intake is short: the video topic and rough title, the frames or photos the creator can give you, the one true thing the video delivers, the words or claims that are off-limits, and the channel’s usual colors and style so the pack still looks like their channel.
- Time per paid job: about 60–90 minutes the first few times, closer to 45 once your layout routine settles. Concepts and backgrounds are the fast part now. The slow part is the part worth paying for: checking each design at small size, fixing hands, edges, and garbled details by hand, setting the text yourself so every word is spelled right, and making sure all three options are honestly different — not the same picture in three colors.
Earning scenarios
The market check first, because it keeps this honest. Checked November 10, 2026: Fiverr’s own guide puts the average thumbnail job at about $18; Upwork puts a single custom thumbnail at about $25–$100, a set of two to three test concepts at about $75–$250, and ongoing channel support in the low hundreds a month; entry designers commonly sit around $20–$35 a thumbnail, with seasoned specialists well above that. At the cheap end, gig listings start at $5. You are not selling one $18 picture and you are not a seasoned specialist — you are selling a beginner, direct-sale test pack priced under the going rate for a concept set, with the checking and the title pairing included.
- Low: $0 in the first month. You pitch fifteen creators, most never reply, two say the sample looks fine and then go quiet. Normal. Packaging is easy for a busy creator to postpone, and free and cheap AI tools are one search away, so “no” and silence are the default answers.
- Typical (BogusAPY beginner estimate): $59–$99 for one video’s test pack — three thumbnails, three paired titles, one revision round — sold direct. Two to four packs a month is roughly $120–$400 of evening money. At an hour or so a pack that is modest pay, and the first packs will run slow while you learn what readable actually looks like at phone size.
- High (not promised): one weekly creator takes a four-video month at an estimated $199–$349, and an occasional second creator orders single packs around upload weeks. A good month might reach a few hundred dollars beyond the typical case. That is the ceiling worth picturing at the start — not the five-figure specialist retainers at the top of the market, which belong to people with years of tested results behind them.
- Time to first dollar: one to two weeks is plausible, because the pitch carries a free concept made for the creator’s own recent video — they can judge it in thirty seconds, the job itself takes a day or two, and a flat fee under $100 is an easy first yes. Finding the creator who replies is the slow part, as always.
Launch steps
- Build the sample first, and label it a sample. Pick a public video in a niche you actually watch and rebuild its packaging as a concept pack: three thumbnails, three titles, one line each on what changed. You are not claiming the creator hired you or used them — it is a demonstration, marked as one, in your portfolio folder.
- Learn the small-size test before you sell anything. Shrink every design to the size it appears in a phone feed and look again. If the words, the face, or the object do not read in two seconds at that size, the design is not finished, whatever it looks like full-screen. Three to four words of text, set by you, beats a paragraph rendered by a generator.
- Make a target list of fifteen scheduled uploaders. Regular long-form, real views, and a way the channel earns (ads, a sponsor, a product, a service) — and packaging that is visibly an afterthought. Note the last video’s current thumbnail in one line. That note is your opening evidence.
- Pitch with one free concept, for their video. “Your last video does the hard part — here is one thumbnail concept for it, free, no strings. I sell a three-option test pack per video, flat fee, so the platform can pick a winner with real data instead of a guess.” One concept, not the whole pack. The pack is what costs money.
- Deliver files, titles, and the difference notes — nothing else. Full-size images in a shared folder, the three titles in a short doc, one revision round in writing. The creator runs the test and owns the result, good or bad. Ask once, a week or so after publish, how the test went; that answer, with permission, is the only result you may ever quote.
- Offer the four-video month once, after a pack lands well. One line, one price, no pressure tour. Repeat work is where this idea stops being pocket money, and it is earned by packs that were accurate, early, and easy to use — not by a sales script.
Red flags
Never promise clicks, views, or growth. A thumbnail is one input among the topic, the title, the channel’s history, and the video itself. You sell three honest, testable options and a fair test. Anyone guaranteeing a click-through rate is selling fiction, and a creator who has been burned by that pitch will assume you are too.
Fake proof is disqualifying. No invented revenue screenshots, no fake before-and-after, no logos of brands the video never mentions, no faces of people who are not in the video. AI will generate all of it happily if you let it. Every image element has to trace back to the creator’s own frames, photos, or properly licensed material, and the thumbnail has to describe the video that actually exists.
AI details fail quietly. Extra fingers, melted object edges, warped product shapes, and misspelled words are the standard failure set, and they hide at full size and shout at small size. Zoom in, then shrink it down, then look once more before anything ships. A pack with a garbled word in it tells the creator exactly how much checking they paid for.
The $5 floor is real, and so is the do-it-yourself button. Gig sites race single thumbnails to $5, AI generators offer unlimited concepts for a low monthly fee, and subscription services undercut per-video pricing at volume. If your only offer is “a picture, fast,” you lose to all three. The pack survives only as judgment: channel fit, honest claims, readable lettering, and three options different enough that testing them means something.
Put usage in writing. Before the first paid pack, agree in one short message what the fee covers — the three designs, the titles, one revision round, and the creator’s right to use the final thumbnails for that video. “And the channel banner, and a logo, and next week’s too” is a new quote, not a revision.
First move tonight
Spend about 40 minutes on the sample that sells this whole idea. Pick one public video you watched recently in a niche you like, save its current title and a mental picture of its current thumbnail, and build exactly one replacement concept: pull a strong frame or build a clean layout from the video’s real subject, draft the background or cleanup with AI, then set three or four words of text yourself in a free editor. Shrink it to phone-feed size and sit it next to the original. If yours reads faster and says nothing the video does not deliver, you have the free concept you will send your first creator tomorrow — and you have felt the checking work that is the actual product.
Sources: thumbnail pricing checked November 10, 2026 — Fiverr’s thumbnail hiring guide average of about $18 per job; Upwork’s thumbnail designer guidance at about $25–$100 per single thumbnail, $75–$250 for a set of two to three test concepts, and low hundreds a month for ongoing support; entry designer rates around $20–$35 per thumbnail; thumbnail subscription tiers at about $29–$69 a month, and AI thumbnail tools at about $19–$39 a month. Per-pack and monthly fees above are BogusAPY beginner estimates for direct sales, labeled as estimates in the text.
Get posts like this by email: Subscribe to BogusAPY
Follow BogusAPY on X: @BogusAPY
New idea every evening — filed under AI Money Ideas on BogusAPY. This is an idea with honest numbers, not a promise of income, and it’s not financial advice.
Comments
Post a Comment