How the paid clipping economy works

For most of its life, clipping was something fans did for nothing. Someone watched a three-hour stream, pulled the ninety seconds worth keeping, captioned it, posted it. The streamer got reach they had not paid for; the clipper got an audience of their own.

That still happens, and it is no longer the interesting part. An industry has grown on top of it (agencies, briefs, rate cards, invoices), and Le Monde's video investigation is the most detailed look inside it published so far. The figures that travelled from it are the enormous ones. The number actually worth reading is small and dull: the price of a thousand views. Almost everything about clipped content follows from it.

The sourceComment le clipping bouleverse nos réseaux sociauxLe Monde · video investigation, July 2026 · in FrenchWatch it here

What is being bought

The shape of a campaign is the same whether the client is a record label, a clothing brand, an e-commerce platform or a film distributor.

A brand arrives at an agency with a budget (Le Monde's example is €10,000), a brief, and the raw material: an interview, a trailer, a set of stems, whatever the clips are to be made of. The agency pushes the brief out to its network. The network is not employees; it is a chat group. The French agency founder Le Monde interviewed described 337 clippers active enough to post daily for a month, more than 600 people in the WhatsApp community behind them, and 2,700 clips shipped in the previous month.

Each of those clippers cuts their own videos and posts them from their own accounts. Then the money moves on one term: a rate per thousand views. €3 per 10,000 views in one example, €2 per 1,000 in a film campaign the same investigation found listed publicly: a distributor's €3,000 budget, a brief, footage supplied, buying about 1.5 million views if the clips deliver and costing nothing if they do not.

That last clause is the whole product. A video editor is paid for 200 videos and delivers 200 videos, whatever they do. A clipper is paid for the views, so the client is buying an outcome instead of a deliverable, at a price per thousand that no ad platform will match. The founder's own summary: it is sponsorship, slightly cheaper.

What paying per view selects for

If your revenue is views × rate, and a clip costs you fifteen minutes, then the arithmetic points one way. Most clips will do nearly nothing. A few will do a hundred thousand. One might do a million, and that one pays for the month. Your expected earnings are dominated entirely by the tail, so the strategy is to buy as many lottery tickets as your evening allows.

Everything visible in the output comes from that:

• Volume over selection. A 30-minute video yields 400 to 500 clips, and the investigation quotes that figure without irony, because at this rate card there is no reason to stop at ten. Choosing carefully has no upside; the algorithm does the selecting.
• Everything spent on the first second. Nothing rewards the second half of a clip. Watch-through matters only through its effect on distribution, and the hook is what decides whether distribution happens at all.
• Convergent style. The same captions, the same zoom, the same twelve moments from the same podcast. When several hundred people optimise the same metric against the same recommender with the same source footage, they arrive at the same video.
• A wage that reads like piecework. €500 to €1,000 a month, per the same agency, which is why nobody clips for one agency. The people doing this work several networks at once: click microworkers, in Le Monde's phrase.

None of that is a criticism of the people doing it. It is a description of a price sheet. Pay per thousand views and you get volume; you would get something else by paying for something else.

Why it works at all

The mechanism is not persuasion, it is repetition. Advertising has a rule of thumb, the rule of seven, that a person needs something like seven exposures before it registers as a thing that exists. Clipping is a way to buy exposures in bulk, from accounts that do not look like advertising, in a feed where the seventh contact arrives the same week as the first.

It also has a hard limit, and the investigation is careful about it: clipping does not create the appetite. A film nobody wants to see does not become wanted because 400 clips of it exist. What the spend buys is contact with an audience that would otherwise never have arrived at the source, which is worth a great deal, and is not the same as manufacturing a hit.

Where it stops looking like advertising

The same machine that promotes a film promotes a person, and there the numbers stop being an amusing curiosity. Le Monde reports one masculinist influencer spending between $500,000 and $700,000 on clipping campaigns across a single month in 2026: roughly 1,500 clippers, 70,000 videos, 2.2 billion cumulative views. Do the division on their figures and those views cost somewhere around a quarter of a dollar per thousand. Andrew Tate's earlier network, which the same investigation traces back to a 2022 recruitment ad, is reported at around 20,000 clippers.

And the sector has been shopping itself to politics. A participant told Le Monde they had been approached by people around two prominent political figures, one of them a presidential candidate. Nothing came of those particular conversations.

The problem is not that a clip is misleading. Most of them are accurate excerpts. It is that a bought distribution is indistinguishable, in the feed, from an unbought one. Seventy thousand people posting the same person in the same month looks like a groundswell and is a line item. Nothing in the format marks the difference, and nothing currently requires anyone to.

Two jobs sharing one word

If you cut clips yourself, the useful thing to take from all this is that "clipping" now names two jobs with almost nothing in common.

One is a media buy. It is measured in views, priced per thousand, and won on throughput: how many hooks you can produce in an evening, how fast you can find the moments, how little you can afford to care about any individual clip.

The other is editing. It is measured by whether someone watches to the end, comes back, and goes to the source, and it is won on judgement about which sixty seconds of an hour actually stand alone.

The tools are the same. The incentives point in opposite directions, and quite a lot of advice about short-form video is really advice about the first job being sold to people doing the second. Worth knowing which one you are being paid for before you take the note.

Sources

Everything factual above rests on one piece of reporting, and it is worth being exact about which parts:

• Le Monde, Comment le clipping bouleverse nos réseaux sociaux, video investigation, July 2026: watch it (in French). The budgets, the rates per thousand views, the headcounts, the monthly clip volumes, the campaign spend and the political approach are all reported there, on the record, by the people running the agencies.
• The cost of roughly a quarter of a dollar per thousand views is ours, not theirs: it is the division of the spend they report by the views they report, and it is stated as such in the text above.
• For a plain definition of the practice and how it spread, the French Wikipedia entry Clipping (réseaux sociaux) is a reasonable starting point and cites the same investigation.

We have added no figure of our own to any of this, and where our reading of the economics goes beyond what Le Monde says, the article says so in the sentence that does it.