Field guide

How to choose a video production company

The showreel is the least informative thing they will send you.

By Paul Joseph · Updated 24 Aug 2026

Choose a video production company on how it interrogates your brief, not on how good its work looks. A showreel shows you what a vendor was permitted to make, for a company with a different budget, a different brand and a different funnel. It is a filter for taste, and taste is the part you are least likely to get wrong. What separates a vendor that produces an asset you can defend from one that produces an expensive object is almost entirely upstream of the camera: whether they ask what the film is for, what a viewer should do next, and how anyone will know it worked.

Before you brief anyone, it is worth knowing which link in your own chain is weak — the Dark-Funnel Video Audit scores that in two minutes, free, and it changes which kind of vendor you should be shortlisting.

Why the showreel misleads

A showreel is a selection of a vendor's best-funded, least-constrained work, cut to music, stripped of context. It cannot tell you what the brief was, what it cost, how long it took, how many rounds it went through, or whether it produced anything. Two films that look identical on a reel can sit on opposite sides of the only question that matters — one was commissioned against a funnel stage and instrumented, the other was commissioned because the previous one was three years old.

This matters more than it sounds, because the reel is doing exactly what it was built to do: it is a sales asset optimised to make a buyer feel confident quickly. Feeling confident quickly is the failure mode. The most common version of this mistake is choosing the vendor whose reel best matches the film you have already imagined — which selects for a vendor that will build the thing in your head rather than the thing your funnel needs.

Use the reel for one narrow purpose: confirming basic craft competence and whether the register suits your market. That takes about four minutes across three vendors. Then set it aside and spend your evaluation effort on the parts that actually vary.

The five questions that separate vendors

Ask all five of every shortlisted vendor, and pay more attention to the shape of the answer than to its content. You are testing whether they think about the work the way you need them to.

Notice what is absent: nothing here asks about equipment, crew size or turnaround. Those are procurement details, they are easy to compare, and they are not where the variance lives.

Brief everyone identically, or you are not comparing

The most common procurement error is sending three vendors a loose description and comparing what comes back. Each fills the gaps differently, each prices its own interpretation, and the resulting numbers describe three different projects. The cheapest then wins by having scoped the least — which is not a negotiation outcome, it is an accident.

Send the same written brief to all three, name the same deliverables, and state the budget range. The five decisions in the 60-Second Brief — funnel stage, viewer action, martech trigger, contact record, pipeline outcome — are the ones that make quotes comparable, because they are what determines scope. A brief that names a runtime and a tone leaves every vendor guessing at the same four unknowns.

A worked example: three quotes, one brief

Illustrative arithmetic rather than a benchmark — the point is the method for normalising, not the figures.

A B2B software company briefs one customer story: a single shoot day at a customer site, a 2-minute master and three social cut-downs, for use on an owned page and in paid social for two years. Identical brief, three vendors. The headline numbers come back at $18,000, $34,000 and $52,000, and the instinct in the room is that the first is a bargain and the third is taking liberties.

Now price the gaps at what it would cost to close them afterwards, which is the only fair basis for comparison. Vendor A's three cut-downs commissioned separately run $2,400 each; a two-year music licence procured alone runs about $900; and the paid-social usage it did not grant is the one that hurts, because the asset was commissioned for paid social — buying that right retrospectively, from a position with no leverage, is quoted at $6,500.

The $16,000 gap between A and B closes to $1,400 — inside the noise of any estimate, and well inside the value of preferring one team over the other on the five questions above. The real decision was never A versus B on price; it was whether C's perpetual rights and project files are worth $18,000 to a company that may want to re-edit this asset in three years. That is a genuine judgement call, and it only becomes visible once the quotes are normalised.

Two cautions. The retrospective prices are estimates, and the honest way to use this method is to ask each vendor to price the missing items up front rather than to guess on their behalf. And a low quote is not evidence of bad faith — Vendor A most likely scoped exactly what it read, which is an argument for a tighter brief rather than against that vendor. Whether the whole programme is sized correctly in the first place is a separate question, worked through in what corporate video costs.

Specialist or full-service

The prior question — whether to buy at all, or to build the capability internally — is worked through in in-house video team versus agency. Taking the buy decision as settled, the category question is usually asked the wrong way round. It is not “which type of vendor is better” but “which link in my chain is weak, and can I fill it internally?” — which is a question about your own organisation, and the reason the Audit is worth running before the shortlist rather than after.

If the brief is sound, the funnel stage is decided and someone owns distribution, a production specialist gives better craft per unit of spend, because you are buying the thing they are best at and nothing you do not need. If nobody internally owns where the asset sits or what fires when someone watches it, a partner who will do that thinking earns the premium — but ask for evidence rather than a capability slide. “Show me a project where you chose the funnel stage and what you measured” separates the two immediately.

The failure mode to avoid is buying strategy you already have. Paying a full-service premium and then handing over a fully-formed brief means paying twice for the same thinking, and it frequently produces friction, because the partner's process assumes a blank page you did not give them.

Warning signs

None of these is disqualifying alone. Two or more together is a pattern.

• A quote with no questions attached. The strongest single signal. It means the vendor priced its assumptions rather than your requirement, and you will discover which assumptions in the edit.
• A price before scope. A number offered in the first call, before deliverables are agreed, is an anchoring device. It will be revised upward once scope is real, and by then you have a relationship and a deadline.
• No named people. “Our team” is not an answer. You are buying a director's judgement and an editor's craft, and both are individuals with names.
• Silence on rights and revisions. The two clauses most likely to cost you later are the two most often left vague. Absence here is not an oversight; it is optionality retained by the vendor.
• A reel with nothing measured behind it. Ask what any two films on the reel were for. A vendor that cannot answer for its own portfolio will not start asking on yours.

Buy a pilot before you buy a partner

Where the relationship is intended to be ongoing, one small paid project is the cheapest information available. A pitch shows you a vendor's sales process. A pilot shows you how they scope, how they communicate when something slips, how they take a note they disagree with, and what they do in the last forty-eight hours before a deadline — which is where the difference between vendors actually lives.

Two conditions make it work. Pay properly, because an underpaid pilot tells you how a vendor behaves when underpaid, which is not the question. And make it a genuinely useful asset rather than a test exercise, so the money is not wasted if the answer is no. A single explainer for a step you already send traffic to is a good candidate: small, real, and measurable within a quarter.

Judge the process at least as seriously as the output. A vendor that delivered something merely good through a calm, well-run process is a better multi-year bet than one that delivered something excellent through chaos you would have to absorb every time.

Settle these before signing

Six items, none contentious if raised early and all expensive if raised late. Put them in the contract rather than the email thread.

1. 1. Named deliverables. Formats, aspect ratios, durations, captions, and the specific number of cut-downs. “Social versions” is not a deliverable.
2. 2. Usage rights. Which channels, which territories, how long, and whether paid media is included. Match this to where the asset will actually live, not to where you expect it to live today.
3. 3. Revisions. How many rounds, what constitutes one, and what is chargeable. The distinction between a revision and a re-scope should be written down while everyone is still relaxed.
4. 4. Music and stock licences. Who buys them, for how long, and what happens at expiry. An expired licence can quietly make a working asset unusable.
5. 5. Project files and raw footage. Whether you get them, in what form, and for how long they are archived. This is what determines whether a re-edit in year three is cheap or impossible.
6. 6. Who signs off, and by when. Approval delay is the commonest cause of overrun, and it is almost always the client's fault. Name the approver and agree the turnaround before the shoot.

Item six is the one buyers skip and then pay for. A vendor cannot hold a schedule against an approval chain that was never named, and the resulting overrun is usually charged back — fairly.

Related: the Buyer's Guide covers the commissioning decision that comes before vendor selection, how to write a video brief is the document you send them, the Frame to Funnel Method names the frameworks a good vendor will recognise, and the glossary defines every term used here.