Video is the easiest marketing spend to approve and the hardest to defend afterwards. A finished film feels like an achievement, and the metrics it produces — views, watch time, engagement — sound like success without describing any. Video production ROI becomes measurable only when you decide, before the brief is written, which commercial job the video is doing and which number would change if it did that job well.
That discipline is unglamorous and it changes budgets dramatically. It usually means fewer hero films and more of the small, specific assets that sit at the points where people actually decide.
Give each video one job
Videos fail commercially when they are asked to do everything: explain the product, establish the brand, answer objections and close the sale, all in ninety seconds, for an audience that ranges from strangers to people holding a quote. The result is a film that is competent and moves nothing.
A useful brief names the viewer’s state and the next step. Someone who has never heard of you needs a reason to care. Someone comparing three suppliers needs a reason to trust. Someone who already asked for a proposal needs an objection removed. These are three videos, and they are cheaper to make separately than one film trying to serve all three.
- Awareness: earn attention from people with no active need.
- Consideration: make the offer concrete — how it works, what it looks like, who it is for.
- Conversion: remove the specific doubt that stalls decisions.
- Retention: reduce support load and increase repeat purchase after the sale.
Where the measurable return usually is
Awareness video is real but slow to prove. The returns that show up fastest are further down: a product demonstration on a pricing page, a short answer to the objection your sales team hears weekly, an onboarding clip that cuts support tickets. These have a defined before-and-after and a small audience whose behaviour you already track.
A conversion video’s return is testable directly: run the page with and without it and compare conversion rate over enough traffic to mean something. That single experiment produces a defensible number, which is more than most brand films ever generate.
Costs people forget
Production is only part of the spend. Distribution frequently costs more than the film — media to put it in front of people, the editing work to cut platform-specific versions, subtitles and localisation, and the re-shoot when the product changes in six months. A budget that funds production and nothing else buys an asset nobody sees.
Shelf life belongs in the calculation too. A film tied to this year’s pricing and interface depreciates quickly; one built around a customer problem can run for years. Where possible, shoot modularly — evergreen core, replaceable specifics — so an update is an edit rather than a new production.
Measuring without fooling yourself
View counts measure delivery, not value. Better questions: did the people who watched convert at a higher rate than those who did not, and can you rule out the obvious confound that interested people watch more? Did conversion move when the video appeared on the page? Did support contacts about a topic fall after the explainer shipped?
- On-page video: A/B test the page, not the video.
- Paid video: measure the downstream action, and pause-test broad campaigns.
- Sales video: track whether opportunities that received it progress differently.
- Support video: count tickets on the topic before and after.
Watch-through data is most useful diagnostically rather than as a score. A sharp drop at fifteen seconds tells you the opening promised something the middle did not deliver — a fixable editing problem, and usually cheaper to fix than to re-shoot.
Production value versus return
Higher production quality raises cost faster than it raises response, past a threshold. That threshold is lower than most briefs assume: clear audio, stable framing, honest lighting and a tight edit cover the majority of commercial video work. Cinematic grading rarely changes a conversion rate.
Spend the premium where it genuinely signals something — brand films, flagship product launches — and keep the workhorse assets fast and repeatable. Teams that can produce a decent video in a week test more ideas than teams that produce one perfect film a quarter, and testing is where the return compounds.
Build a portfolio, not a campaign
Treat video as a portfolio with different risk profiles. Conversion and support assets are low-risk with measurable payback and should be funded continuously. Brand and awareness films are higher-risk investments judged over longer horizons against softer evidence — branded search, direct traffic, sales cycle length.
Keep the two accounted for separately. Mixing them produces the familiar bad outcome: a brand film judged on last-click conversions and declared a failure, or a conversion asset judged on views and declared a success while it sells nothing.
The teams that get a defensible return from video are rarely the ones with the largest production budgets. They are the ones who knew what each film was for, measured the thing it was supposed to move, and were willing to stop making the ones that moved nothing.
Keep reading: Video Production · Graphic Design