CPM explained: what GH₵60 per 1,000 views actually means for your income
CPM sounds technical, but it's simple: GH₵60 per 1,000 real views, no deductions. Here's why that beats what most African creators earn from ad platforms directly.
By Thabo Dlamini · June 22, 2026 · 5 min read
Every creator has the same folder problem: one long, honest, occasionally brilliant video sitting at forty minutes, and no real sense of which three minutes of it a brand would actually pay to sit next to. Clip detection is ClipAd's answer to that problem, and it is worth understanding exactly what it does, because "AI cuts your video" undersells it and slightly misdescribes it too.
When a creator uploads a long-form video to ClipAd, nothing happens to it automatically. Detection is opt-in — a deliberate action the creator takes when they're ready, not a background job that fires the moment a file finishes uploading. That distinction matters more than it sounds like it should, because it means the raw upload just sits there, untouched, until you decide to do something with it.
This is the first of several places in the pipeline where ClipAd could have taken a shortcut and chose not to. The video is yours until you tell the system to look at it.
Once you choose to run detection, ClipAd transcribes the audio word-for-word with timestamps. That transcript is then handed to an AI model that scores and segments it into distinct, complete moments — not arbitrary time slices, but sections the model judges to be self-contained enough to work as a standalone clip.
The output is a set of proposed clips. Proposed is the operative word. At this stage nothing is cut, nothing is published, and nothing is placement-ready. It's a list of suggestions waiting for a human to look at them.
Two design choices in the detection step are easy to skim past but do most of the actual work. First, every proposed clip snaps to real sentence boundaries, so it starts and ends on a complete thought instead of chopping off mid-sentence. Second, clips are deliberately kept in the 30-second-to-3-minute range, and never shorter than 30 seconds.
Both of those constraints exist because of how people actually watch short-form video, not because of an arbitrary preference. In Vidyard's 2024 Video in Business Benchmark Report — based on 943,305 videos created by Vidyard users between January 1 and December 15, 2024 — videos under a minute long retained 65% of viewers to the end, while videos over 20 minutes retained only 20% of viewers to completion. That's a real gap, and it's the gap clip detection is built to close: turning one video with a 20% completion rate into several with a shot at that 65% number.
The complete-thought part matters just as much as the length part. A clip that cuts off before the punchline, the answer, or the point isn't short-form content — it's a broken preview. Sentence-boundary snapping is what stops a technically-short clip from still feeling unfinished.
A clip that's the right length but ends mid-sentence isn't short-form content. It's just an unfinished thought with a timer on it.
The bet on short-form generally is backed by more than intuition. Sprout Social's 2022 Sprout Social Index found that 66% of consumers said short-form video was the most engaging type of social content, and while that's a 2022 data point rather than a fresh one, the trend it flagged has only hardened since. Per Sprout Social's 2026 Social Media Content Strategy Report, consumers are now slightly more likely to interact with short-form video (52%) than long-form video (48%) on YouTube itself — a platform that used to be long-form's home turf. And on the brand side, HubSpot's 2025 State of Marketing Report — surveying more than 1,200 marketing leaders globally — found short-form video is the content format marketers most often cite as delivering the highest ROI, a ranking that held into HubSpot's 2026 data as well.
Here's the part that's easy to gloss over in a product explainer but shouldn't be: the creator reviews every suggested clip individually and approves or rejects it before it becomes anything real. The AI model's job is to propose. Your job is to decide. There's no batch-approve-everything button and no setting that lets detection results go live on their own.
Only once you approve a clip does ClipAd cut that exact segment into its own frame-accurate standalone clip file. That file is the actual deliverable — a real, separate video, not just a marked-up timestamp range inside the original upload.
This is where the phrase "placement-ready" earns its meaning. Only an approved, cut clip can later carry a brand's ad placement. A raw upload can't. A proposed-but-unreviewed clip can't. The gate isn't cosmetic — it's structural, because the standalone clip file is the actual object a brand's ad gets attached to.
And even after a clip clears that gate, the ad side stays manual too. Once a clip is approved and cut, you place a brand's ad creative onto it yourself — before, during, or after the clip — for rendering and eventual publishing. Nothing auto-runs a brand's creative onto your content. You choose the clip, you choose the slot, you approve what goes out.
It would be technically simpler to auto-approve every AI-scored clip and auto-attach whichever creative a brand uploaded. ClipAd doesn't work that way on purpose. Every stage — running detection, approving a clip, placing an ad on it — is a decision you make, not a default you have to opt out of.
That matters doubly in a market like Ghana, where the audience creators are building for keeps growing fast: DataReportal's "Digital 2026: Ghana" report put social media user identities at 8.59 million as of October 2025, up 780,000 year-over-year, alongside 26.3 million internet users at 74.6% penetration. Bigger audiences make each individual placement decision worth more, not less — which is exactly why the review step stays in your hands.
Strip away the AI framing and clip detection is really a labor-saving first pass: it reads your whole video, finds the moments most likely to hold a stranger's attention for their full length, and hands you a shortlist instead of a blank timeline. You still make every real decision — what gets approved, what gets cut, what carries a brand's ad. That's the whole design.
If you're curious how a cut clip actually turns into paid income once it's carrying a placement, the how it works page walks through the rest of the pipeline, and you can see the current rate card for yourself on the pricing page before you upload your first video.
CPM sounds technical, but it's simple: GH₵60 per 1,000 real views, no deductions. Here's why that beats what most African creators earn from ad platforms directly.
Why ClipAd pays creators in cedis straight to Mobile Money instead of a foreign-currency bank rail — and why that choice matters more in Ghana than almost anywhere else.
Global ad money is moving from traditional media to creators — backed by real ROI data and a young, mobile-first African audience. Here's what that shift looks like once it lands in a brand's actual budget.
Upload what you already make, approve your clips, and get paid in your local currency. Free to join.