Local or National Agency for Video: The Decision Grid

Neither, on principle: what has to be local is the shoot, not the agency. If your subject can be shipped, product, packshot, UGC, choose the vertical video specialist, wherever it is. If your subject is fixed in place, point of sale, workshop, team, multi-site network, you need local capture, not a local office: centralised creative direction, execution close to the ground. Framing the choice as “local agency or national agency for video” confuses the provider’s address with the place of the shoot. These are two separate decisions, and only one of them costs kilometres. According to Comarketing-News (September 2025), TikTok has 27.8 million monthly users in France, its biggest market in Europe.

Why is the “local or national” question badly framed?

Because a vertical video production breaks down into six building blocks, and only one of them has a real geographical constraint. The other five have been manageable from anywhere since 2020, and nobody in the industry has seriously challenged that since.

Look at what you are actually buying when you order ten vertical creatives for TikTok and Meta:

• Creative strategy: angles, promises, competitor mapping. No location constraint.
• Writing: scripts, hooks, structures of 15 to 45 seconds. No location constraint.
• Capture: someone, somewhere, presses REC in front of a physical subject. The only localised block.
• Post-production: editing, subtitles, motion, sound design. No location constraint.
• Iteration: reading the data, a new variant, re-editing. No location constraint.
• Media management: account structure, budgets, trade-offs. No location constraint.

One block out of six. That’s all. And yet almost all the pages that cover this subject, written, without notable exception, by local agencies that conclude you should pick a local agency, reason as if the choice applied to all six.

The vocabulary is the problem. People talk about a “local agency” as an identity, a nature. That is false in operational terms. Locality is not a characteristic of a provider, it is a resource rented by the day. A videographer in Clermont-Ferrand, a sound crew in Lille, a UGC creator in Bordeaux: these are resources you can mobilise, not corporate philosophies. You can rent three of them in the same week without changing agency.

This reversal changes the question. You are no longer choosing between two types of structure. You choose who holds the creative direction and the iteration loop, then you buy days of capture separately wherever your subject is. This is the third dimension of a decision most brands frame on only two axes: the what (creative ads, UGC, organic) and the make or buy between agency and in-house team. The third axis is the where. And it is dealt with after the other two, not before.

Distance is not measured in kilometres between you and the agency. It is measured in kilometres per useful deliverable.

Two symmetrical pleas are in circulation, and both are shaky. The first: “local understands your territory, is responsive, supports the regional economy.” The second: “national has the means, the references and the structure.” Neither says who writes the hooks, who reads the campaign data on Tuesday morning, and how much time passes between a creative hypothesis and its going live. Yet those are the only things that determine whether your creatives work.

Can your subject travel?

It is the only question that settles the local/national debate, and it can be answered in thirty seconds. A product under 15 kg, non-perishable, unregulated: it travels. A restaurant, a production workshop, a team of 40 people, a network of 60 franchisees: they do not travel.

Internally we use a three-slider grid, which we set before even opening a tender. We call it the Subject / Capture / Loop grid.

Slider 1, Subject: what has to be filmed?

Four possible answers: a shippable product, a place, a person, a bulky catalogue. The catalogue is the trap case: 300 references to film is a shipping logistics problem before it is an agency geography problem. We cover this in our article on purchase units for e-commerce product video.

Slider 2, Capture: what physical presence?

None, occasional, recurring, multi-site. An occasional presence, two days per quarter, never justifies choosing your provider on an address criterion. A weekly recurring presence does. The tipping point, from what we observe in our own schedules, is around two capture days per month: below that, travel is diluted in the budget; above it, it becomes a line item in its own right that has to be structured.

Slider 3, Loop: how often do you iterate?

Quarterly, monthly, weekly. It is the most ignored slider and the most discriminating. In vertical video, value is not created at the shoot, it is created in the number of laps between a performance data point and a new variant going live. An agency that delivers a film in six weeks and never touches it again has sold you an object, not a system.

Hence the rule we apply: you choose a specialist on the loop, and you rent locality for the capture. Not the other way round. Choosing first on geography is optimising one block out of six.

The decision table by configuration

Out of six configurations, only two make locality structurally necessary. And in those two cases, it is not the agency that has to be local, it is the shooting crew it mobilises.

Who actually shoots when you sign with a national agency?

In a majority of cases outside Île-de-France, it is not the agency you met, it is an independent videographer hired for the day in your département. This is not a scandal. It is sometimes even the best possible organisation. But you need to know before you sign, because it changes who owns the rushes and who answers if the shoot goes wrong.

The French audiovisual production sector is highly fragmented: according to INSEE’s SIRENE data on NAF code 59.11B (production of corporate and advertising films), the vast majority of registered structures have fewer than three employees, and supply is heavily concentrated in Île-de-France. Check against the up-to-date database before quoting a precise percentage, but the order of magnitude is known to everyone who recruits crews outside Paris. Direct consequence: “national coverage” almost always means “national address book”. And the reverse is also true: the local agency subcontracts motion design, sound design or subtitling to a provider who is not in its own office. According to the Arcom study on the economic fabric of audiovisual production (7th edition, 2023), Île-de-France concentrated 64% of the sector’s structures in 2021.

Cascading subcontracting is not a flaw. The flaw is discovering it on the day of the shoot.

The 5 questions that reveal who presses REC

• Name the person who will hold the camera. A first name, a link to their work. Not “our network of partners”.
• Are they an employee, on a freelance basis, or an independent invoicing directly? Each status has a different coordination cost, generally 15 to 25% above the freelancer’s direct rate according to the quotes we see come through.
• Who assigns me the rights to the raw rushes, and by what date? Many contracts assign the edited deliverable, not the rushes. A major detail if you want to re-edit yourself in six months.
• What happens if this person is unavailable in November? Test the answer. “We’ll find someone” is not one.
• Who edits, and where? The editor is often the real author of a vertical creative. If they change with each project, your visual identity changes with them.

These five questions are geographical. They complement, without replacing, the 12 questions to ask before signing, which cover the rest of the contractual scope.

How much does proximity really cost, and how much does distance cost?

Logistics is never expensive per kilometre, it is expensive per deliverable. A Paris–Nantes round trip for three creatives is ruinous. The same round trip for forty creatives is an accounting detail. That is the only honest way to compare.

We have already detailed price ranges by sector in the marketing video production price grids. Here we look at just one line: the one almost nobody prices correctly on a quote.

The metric to impose: logistics cost per useful variant

Formula: (travel costs + billed travel time + neutralised days) ÷ number of usable creatives from the shoot. A usable creative is a variant that can go live, not a rush.

Let’s take a quote simulation; these are working assumptions, not published data. One capture day in a region with a €400 travel flat fee and a half-day of travel billed at €350: €750 of logistics. If the shoot produces 6 usable creatives, logistics costs €125 per creative. If the same shoot is structured to produce 24 variants, same sets, same talent, different angles and hooks, which is perfectly feasible over a well-briefed day, you fall to €31 per creative. The logistics cost fell by 75% without a single kilometre being saved.

That is why the “local = cheaper” argument collapses as soon as you look at the right numbers. A local agency that delivers 6 creatives per day will always be more expensive, per deliverable, than a remote team delivering 24. What you save on petrol, you lose in shoot yield. We stopped selling single-creative shoot days in 2024, precisely because the maths no longer held.

Two honest warnings. First, this reasoning assumes your product or venue can handle 20 to 30 variations in a day; that is not true of a corporate shoot with ten speakers to interview. Second, at a weekly rhythm, a videographer 20 minutes from your warehouse beats any remote structure, whatever the calculation. We come back to this below.

Does “local knowledge” have value on TikTok and Meta?

Meta’s algorithm doesn’t know your agency is 12 kilometres from your head office, and it never will. An ad account contains no “provider proximity” variable. Targeting, bidding, delivery: everything is agnostic to the advertiser’s territory.

What follows goes against half the sales pitches in the industry, sometimes including ours when we talk fast. Knowledge of the local economic fabric has no measurable effect on a CPM or a 3-second retention rate. Platform benchmarks are national, even European; we published them for France in the TikTok Ads France 2026 benchmarks. A hook works or it doesn’t; it doesn’t work better because whoever wrote it lives in the same town as you.

Where locality is worth money, however

• Drive-to-store. A campaign that pushes towards a physical shop benefits from showing the street, the sign, the car park, the tram stop. That is shot on site. And it is shot better by someone who knows when the light is good on that particular street.
• Employer brand and recruitment. Filming a production team in Saint-Étienne means going there several times, winning people over, coming back. A one-off shoot gives stilted videos. In its factsheet on filming conditions in France, the Ministry for Europe and Foreign Affairs reminds us that broadcasting or publishing someone’s image requires the prior consent of the people filmed.
• Casting and location scouting. Finding six creators in Toulouse who match your positioning, negotiating access to a venue, obtaining a filming permit in a city centre: these are real skills, time-consuming and billable. Nobody does them well from a spreadsheet. In Paris, for example, the City requires the filming permit application to be submitted on the AGATE platform at least three weeks (15 working days) before the shoot.
• Accent, language and cultural references. A creator who speaks with their regional accent, or in Breton, or in Corsican, produces a creative that nobody can make elsewhere. This is not folklore: with regional audiences, it changes the completion rate.
• Events. A trade show is covered on site or not at all.

Note that each of these five cases describes a resource to mobilise on the ground, not an agency to choose. You can buy local casting, local scouting, local creators, without moving the creative direction. That is exactly what a properly organised remote studio does.

What we cannot say with certainty: from how many shoots a year a dedicated local team becomes more profitable than hiring one-off capture. It depends too much on the sector, the number of sites and the stability of the need. On the projects we have run, the tipping point appeared between 15 and 25 capture days a year, but that is an order of magnitude from our schedules, not market data.

Multi-site network, franchise, regional SME: which model for which case?

The franchise network is the only case where the answer is neither local nor national: it is a central kit plus local execution, and the two are inseparable. It is also the least documented model on the French market, even though it concerns thousands of brands.

Pure e-commerce, shippable product

Zero locality. You ship three units of the product, the studio shoots, edits, iterates. Looking for an agency “near you” in this case amounts to deliberately restricting your talent pool to a few square kilometres to save on a €12 parcel. Brands still do it, and the price is paid in hook quality, not in postage.

Single-site shop or venue

A restaurant in Annecy, a gym in Reims. The venue doesn’t move, but the need for capture is low in volume and high in publishing frequency. The right architecture: one grouped shoot day per quarter for the brand material, and a pool of two or three creators from the area who produce content continuously from angles written remotely. The creator costs less than a crew, comes more often, and produces native format. The difference between these profiles is, incidentally, first of all contractual, not creative.

Multi-site network or franchise

Sixty points of sale, sixty local realities, a head office that wants consistency. The classic mistake is to let each franchisee choose their regional videographer: you get sixty visual identities, no creative capitalisation, and a total bill higher than a centralised production would have cost.

The model that holds: head office produces a kit: validated script structures, hooks tested in paid, subtitle templates, framing guidelines, sound library, examples of good and bad edits. The sites execute, with their smartphone or a local videographer, from this kit. Head office keeps validation and data feedback. What flows from national to local is not finished video, it is method.

A hard point to anticipate: franchisee buy-in. A kit nobody uses is worth nothing. Plan a two-hour training session per region and a single channel for feedback. Without that, you have made a PDF.

Industrial SME or B2B with a production site

Here, recurring local capture is a necessity, for a reason that is underestimated: access. Badges, safety instructions, production slots, confidentiality of certain lines. A crew that comes once loses half a day on formalities. A crew that comes every month knows the workshop foreman and saves three hours each visit. The creative direction, for its part, can stay centralised with no loss at all.

Which clauses and red flags on a quote, depending on location?

A quote that contains the phrase “travel costs to be determined” is not a quote, it is a purchase option you are signing blind. To the 11 classic red flags of a video quote, geography adds four, and they all cost money after signing.

The 4 wordings to ban

• “Travel costs to be determined” or “extra”. Demand a priced and capped flat fee, or an explicit calculation formula tied to the mileage scale in force. A serious provider knows what a round trip to your place costs.
• “Shooting crew” without a name. Ask who. If they refuse to name anyone, it’s because they don’t know yet, which means they have no contractually secured local capacity.
• “Delivery within 3 weeks” without a breakdown. A round trip of rushes is not a production lead time. Demand: shoot date, first version date, final version date after one round of feedback.
• “Assignment of rights on delivered videos”. The delivered videos, not the rushes. If the shooter is a local third party, you will never get the source files, and any future re-edit will go through a renegotiation.

The 5 clauses to have written in

• Travel flat fee capped, with the cap in euros incl. VAT and the wording “any overrun subject to prior written agreement”.
• An explicit distinction between shoot days and billed days, with travel appearing on its own line.
• Ownership and handover of the raw rushes, with deadline and delivery medium, including when the capture is subcontracted.
• A continuity clause: if the shooter is unavailable, the agency provides a replacement of equivalent level at no extra cost and with no postponement beyond X days.
• A named contact, with channel and frequency of check-ins contractually fixed; a weekly 30-minute video call is worth more than a quarterly lunch 8 km away.

This last clause advantageously replaces all the talk of a “relationship of trust” and “human proximity”. Trust cannot be contractualised. A check-in rhythm, a name, a channel, can.

How can you test a remote provider in 30 days without taking a risk?

A 30-day pilot with a single capture batch and two iteration loops tells you everything a year of references won’t. If you hesitate to look beyond your town, buy a test, not an annual commitment.

The protocol we recommend, and apply to our own starts:

• Week 1: brief, validation of angles, product shipping or remote scouting from a photo file. Criterion: does the provider ask questions about your performance data, or only about your guidelines?
• Week 2: single capture, then delivery of 6 to 10 variants. Criterion: number of genuinely distinct variants, not reframings of the same edit.
• Week 3: going live, reading the first signals, first loop. Criterion: time between your data feedback and the new variant. Beyond 5 working days, the loop is too slow for paid social.
• Week 4: second loop, review. Go/no-go criterion: do you have at least two creatives you want to keep running?

What this pilot measures: loop speed, variant density per capture, quality of data reading. Three things that have nothing to do with distance. What it doesn’t measure: staying power over time and the ability to absorb a peak. For that you need a broader evaluation method, such as a 30-day audit of a creative ads agency.

One last, counter-intuitive point. Many brands outside Île-de-France choose local because they fear being a small client among big Parisian accounts. It is a well-founded worry, but it isn’t solved by geography: it is solved by asking how many accounts your contact manages and writing it into the contract. An agency 6 km away that put you 14th on its priority list will serve you less well than a studio 500 km away that put you 3rd.

Three questions to ask before writing any brief

Choosing between a local agency or a national agency for video means arbitrating a variable that isn’t one. The three real questions fit on one line each.

• Can my subject travel: yes, no, partly?
• How many usable variants do I need to produce per month, and who makes them?
• How many days of physical capture does that really represent over twelve months?

Answer those three and the geography deduces itself, as a line in a quote: “X days of on-site capture, travel flat fee capped at €Y”. Not as a choice of partner.

There is one case where we would frankly tell you to pick someone from your town: a need for more than one capture day per week, all year round, at a single venue. At that rate, travel is no longer an expense line, it is a structural cost, and a videographer fifteen minutes away beats any remote set-up. It happens. It is just far rarer than the market claims.

Frequently asked questions

Do I absolutely need a local video agency to film my shop?

No. You need local capture, which is not the same thing. Creative direction, writing and editing can stay centralised. In practice: one grouped shoot day on site per quarter, supplemented by creators from your catchment area for publishing rhythm. The agency itself is chosen for its vertical video specialisation.

How much do travel costs add to a video production quote?

They must appear as a capped flat fee, never as “to be determined”. The calculation generally relies on the mileage scale and the URSSAF expense reimbursement ceilings, published every year; check the official sources before signing. Also ask whether travel time is billed as a half-day: it is often the heaviest item, more than the transport itself.

How do I know who will really shoot my video if the agency is far away?

Ask for the first name of the person who will hold the camera, their status (employee, freelance, independent) and a link to their work. Then ask who assigns you the rights to the raw rushes and what happens if that person is unavailable. A provider who answers “our network of partners” has no contractually secured local capacity.

Which model should I choose for a franchise network or a multi-site brand?

Neither local nor national: central kit plus local execution. Head office produces the script structures, the hooks validated in paid, the templates and the framing guidelines. Each site executes from this kit, with a smartphone or a nearby videographer. Letting each franchisee choose their own provider produces as many visual identities as there are points of sale.

Is a remote agency less responsive than a nearby one?

Responsiveness is measured in the time between data feedback and a new variant going live, not in distance. Contractualise a named contact, a channel and a weekly check-in frequency. In paid social, an iteration loop beyond five working days is too slow, whether the provider is 6 or 600 kilometres away.

Does knowing the local economic fabric improve ad performance?

Not on a Meta or TikTok account: the algorithm ignores your provider’s location, and platform benchmarks are national. That knowledge has real value elsewhere: drive-to-store, casting regional creators, location scouting, regional accent or language, employer brand. These are resources to rent on the ground, not a criterion for choosing an agency.