Social Media Video Costs: Agency, In-house or UGC?

What a short-form video costs: benchmarks and their limits

Publicly verifiable pricing data for short-form video from the DACH region barely exists. Every citable cost source on this topic is American, calculates in dollars and usually comes from production companies or agencies with a stake in the answer. Treat the figures below as an order of magnitude and as a structure for your own calculation, not as a quote. All prices in this article are as of August 2026. For a Vienna or Munich budget, get two or three local quotes with an identical scope of work and put them next to these ranges.

The US production company D-MAK Productions puts agency work at 500 to over 3,000 US dollars per short-form video. The same source names 100 to 500 US dollars per clip for UGC from small creators and 1,000 to 3,000 US dollars for production companies shooting in UGC style, cheaper per asset as soon as they shoot in batches. At the top end sits classic agency creative. The US industry blog Influencers Time puts manually produced original assets at 3,500 to 18,000 US dollars with 12 to 25 working days of lead time, while AI-assisted editing comes in at 120 to 600 US dollars per usable output with one to four working days for derived assets.

That puts a factor of more than a hundred between the cheapest and the most expensive route. This is not a quality ranking. It is a sign that five different products are being sold under one label.

Usage rights are not a detail, they are a cost block of their own

The most common costing error happens after production. A creator clip you post organically is covered by the base fee. As soon as the same clip runs as an ad or is served through the creator account, a licence comes on top. D-MAK Productions names an uplift of 20 to 100 per cent on the base fee for that, or a monthly licence per video. Other guides name narrower bands inside that range, the order of magnitude stays the same.

That shifts the maths at the decisive point. A clip for 300 US dollars that sinks organically costs 300 US dollars. A clip for 300 US dollars that works and is therefore meant to run as an ad for six months costs a multiple of that. Influencers Time therefore works with a full-cost range of 150 to 1,200 US dollars per clip including usage rights for creator UGC. Negotiate rights, term, channels and markets in the first conversation, not in the moment of success, because that is when the bargaining position flips.

Social media video costs at break-even: when does in-house pay off?

The model question is decided by volume. Green Frog Labs calculates that a budget of 5,000 US dollars per month buys 12 to 20 videos at an agency and 30 to 40 from an in-house creator. The same source credits agency content with 2.3 times higher engagement rates on average. The agency cites a third-party analysis that has not been independently checked, and the figure describes the very service it sells. It works as an argument, not as evidence.

The counter-calculation comes from the fixed-cost side. Vidico sets the threshold at 4 or more videos per month and puts an in-house team at costs from 150,000 US dollars per year with a break-even against outsourcing after 12 or more months, against agency packages for short-form of 3,000 to 15,000 US dollars per month. BLARE Media is stricter and arrives at an effective 20,000 to 25,000 US dollars per video at 12 videos a year once salaries, equipment and overhead are included, and sees in-house ahead only from 30 to 40 or more videos per year.

The table is derived from the two thresholds named above, not a survey of our own. If you want to make the decision cleanly, calculate with your own figures: the fully loaded annual cost of a producer including payroll overheads, equipment, software and downtime, divided by the assets that team realistically produces. Realistic here means what comes out in bad months, not what is in the business plan.

AI production lowers the unit price, not the review effort

The strongest cost lever currently sits between the original and its derivatives. If a manually produced original costs 3,500 US dollars and up while AI-assisted editing costs 120 to 600 US dollars, then any structure that turns one expensive shoot into many cheap derivatives pays off. That is exactly the logic behind content repurposing: the cost of the original spreads across a whole set of derivatives instead of a single asset.

With fully generated clips the unit cost falls further, depending on model and resolution down to cents or low single-digit dollar amounts per short clip, with native 4K costing a multiple of 1080p. These prices change fast. Which model is good for what and where the limits sit is covered in the overview of AI video generators. Two points hold regardless of price: from 2 August 2026, generated content needs transparency labelling under Article 50 of the EU AI Act in the EU, and it needs a human review gate. Both cost time, and that time belongs in the calculation.

Where the benchmarks diverge

You should know three caveats before you write any of these figures into a budget template.

Where the sources come from: D-MAK, Vidico, Green Frog Labs and BLARE Media are production companies and agencies, not independent market researchers. Their figures are industry benchmarks, not study results, and they tend to describe the model that each company sells.

Contradictory thresholds: Vidico names 4 videos per month as the in-house threshold, BLARE Media 30 to 40 per year, which works out at roughly 3 per month, and the Green Frog Labs calculation implicitly assumes far higher volumes. Treat this as a corridor, not as a cut-off.

Currency and market: All values are US dollars from the US market. Labour costs, agency structures and the usual scopes of work differ in the DACH region. Converting to euros produces false precision, not reliability.

The metric that actually counts: cost per usable asset

Price per video is the wrong control metric because it ignores the rejection rate. A package of 20 clips at 200 US dollars each looks cheaper than 8 clips at 600 US dollars each. But if only 6 of the 20 clips clear internal approval and the retention threshold, the effective price per usable asset is around 667 US dollars against 600. This calculation regularly reverses the result of a quote comparison.

For it to work you need two things. First, a defined threshold for when an asset counts as usable, usually via hook rate and completion rate, covered in hook rate, retention and watch time. Second, the discipline to log that rate per supplier for at least a quarter. After three months you know which provider is actually cheaper, and the argument about unit prices settles itself.

Common budgeting mistakes

Budgeting production only: Without a media budget, even the best asset stays inside your own follower bubble. With organic reach declining, production without distribution is half an investment.

Negotiating rights after the fact: The uplift of 20 to 100 per cent hits you exactly when a clip performs and you want to scale it.

Comparing unequal quotes: Concept, casting, location, edit versions, subtitles, vertical variants and revision rounds are cut differently by every provider. Only an identical scope list makes prices comparable.

Ignoring the pricing model: Retainer, project price and outcome-based models shift risk and predictability in different ways. How these three models compare is described in detail for other services, see pricing models for agent infrastructure: retainer, project, outcome.

Building in-house too early: A full-time role is a fixed-cost decision for at least a year. It only pays for itself once the cadence holds through holidays and sick leave as well.

The decision in three steps

Set the cadence you can sustain for twelve months first, not the one you would like. Then run two scenarios for exactly that volume, external and internal, each with fully loaded costs and with usage rights. And measure the rejection rate per supplier from month one, because it moves the unit price more than any negotiation over fees. If you stay below four assets a month, you have already answered the question.