How do you measure whether a brand film worked?

Key takeaways

• Views are the weakest metric available and the one most often reported. A view can be three seconds of muted autoplay.
• Completion rate is the honest headline. It measures whether the film held, which is the only thing the film controls.
• Attribution genuinely stops somewhere. Naming where it stops is more persuasive than pretending it does not.
• Set the measurement frame before production. A film briefed without a success definition will be judged against whatever number is available afterwards.
• Internal effects are real results. Sales cycle, recruitment quality and pitch conversion are frequently the strongest evidence and are almost never reported.

How do you measure whether a brand film worked?

Measure completion rate first, then qualified engagement, then movement in the specific business behaviour the film was made to change. Brand film cannot be attributed to revenue with a single clean number, so the working method is to define one primary success measure before production, track leading indicators against it, and treat aggregate view counts as diagnostic rather than as a result.

Why does attribution fail for brand film?

Because brand film operates on the part of the buying process that leaves no data trail.

A performance ad and a purchase sit close together in time and in the same tracked environment, so the link is legible. A brand film works on familiarity, credibility and preference, and its effect surfaces weeks or months later, through a channel that has no memory of it — someone types the company name into a search bar, or answers a call they would previously have ignored, or shortlists a supplier they now recognise.

Three specific breaks, worth being able to name in a budget meeting:

• Time lag. The effect arrives outside any reasonable attribution window.
• Channel handoff. The film is seen on one platform and the action happens on another, usually via direct or organic traffic that gets credited elsewhere.
• Offline and dark social. The film gets forwarded internally, shown in a meeting, played at an event. All of it invisible.

None of this means brand film is unmeasurable. It means the measure is not a conversion path.

Which metrics actually tell you something?

Ranked by how much they say about whether the film did its job.

Why is completion rate the honest headline?

Because it isolates the variable you paid for. Media spend buys impressions. Targeting buys the right audience. What the film itself is responsible for is whether the people who started watching kept watching.

It also travels well internally, because it is intuitive. "Two million impressions" invites the question of what that cost. "Ninety-nine per cent of people who started it watched to the end" does not need explaining.

A worked example. On the DFDS campaign, the Connected TV placement returned a 99.5% completion rate, alongside 2.1 million DOOH impressions. Those two numbers do different jobs and it is worth being precise about which is which. The impressions figure describes the media buy. The completion figure describes the film — and on a non-skippable environment like CTV a high completion rate is partly a property of the format, which is exactly the sort of caveat that earns you credibility when you volunteer it rather than waiting to be asked.

Read completion by platform, never in aggregate. A 60% completion on YouTube pre-roll and a 60% completion on CTV are not comparable events.

What does the retention curve tell you?

More than any single number, because it shows you where the film failed rather than that it failed.

Three shapes recur, and each has a different fix:

1. A cliff in the first three seconds. The opening is not earning attention. Almost always a problem with the first frame and the first sound, not with the film's substance.
2. A steady slide throughout. Normal, and the gradient is what matters. A shallow slide is a healthy film. A steep one usually means the film is longer than its idea.
3. A sharp drop mid-film. Something specific lost them. Find the timestamp and watch it. It is frequently the moment the film switches from story to explanation.

This is the single most valuable output of a measurement process, because it is the only one that changes what you make next.

What about effects that never appear in analytics?

They are frequently the strongest evidence a brand film produced anything, and they are almost never collected because nobody owns collecting them.

Four worth tracking deliberately:

• Sales use. Is the film being sent by the sales team unprompted? A film that gets forwarded into live deals is working, whatever the dashboard says.
• Cycle length and objection pattern. Sales teams notice when the "who are you" conversation gets shorter. Ask them directly.
• Recruitment quality. For employer branding work, the measure is not applications received. It is whether candidates arrive already understanding the company — and whether the wrong candidates stop applying, which is a success that looks like a decline.
• Pitch and tender outcomes. If a film is used in a competitive pitch, ask afterwards whether it came up.

A short structured question set sent to sales and recruitment at 30 and 90 days will produce better evidence than most analytics exports, and it takes an afternoon to design.

How do you set the measurement frame before production?

By answering four questions at briefing and writing the answers down. A film without an agreed success definition will be judged after the fact against whichever number happens to be available, and that number is usually views.

1. What behaviour should change? Not "awareness". Something observable — more inbound, shorter cycles, better candidates, fewer explanation calls.
2. Who has to see it for that to happen? Define the audience narrowly enough that reach becomes a plan rather than a hope.
3. What is the one primary measure? One. Secondary measures are fine; a primary one is what stops the goalposts moving.
4. When do we look? Agree the review points upfront — typically 30, 90 and 180 days — because brand effects are slow and a 14-day verdict will always be negative.

This conversation belongs in the brief. It is one of several reasons the brief decides more than the shoot does, which is covered in the brief is where most video projects are already won or lost.

What should you stop reporting?

Three things, all of which actively weaken the case they are meant to support.

• Raw view count as a headline. It invites a cost-per-view calculation that brand film will always lose, because it is not what brand film does.
• Engagement rate without a denominator. "8% engagement" means nothing without knowing engagement on what, against what benchmark, on which platform.
• Anything you cannot defend under one follow-up question. A metric that collapses when someone asks how it was calculated costs more credibility than reporting nothing.

The general principle: a smaller number you can stand behind is worth considerably more in a budget meeting than a large one you cannot.

How does DOT approach this?

DOT Video Agency asks for the success definition at the scoping call, and will push if the answer is a general one. It is not a formality — the answer changes the film. A film made to shorten a sales cycle is structured differently from a film made to be remembered at a conference, and both are legitimate.

Where a client is not in a position to define it, that is worth naming as a finding rather than working around. A film commissioned without an agreed measure will be evaluated against something eventually, and it is better for everyone if that something is chosen deliberately.

Related reading

• The brief is where most video projects are already won or lost
• The ideal length for a brand film
• What is Connected TV advertising?
• Brand film production