How to Measure and Improve Video Marketing ROI

I have sat on plenty of scoping calls where someone tells me their last video had “amazing ROI,” and when I ask how they got to that number, the answer turns vague fast. So let me lead with the formula I actually use, then walk through the parts that make it mean something.

Video marketing ROI = (revenue attributable to video − cost of the video) ÷ cost of the video × 100.

A $2,000 video that generates $10,000 in attributable business is a 400 percent return, and the arithmetic is that simple. What trips people up is that both halves of the fraction usually get calculated wrong, which is how so many glowing ROI numbers turn out to be fiction. I will take both halves in turn.

The cost side, where the number gets fudged

Undercounting cost is the most common way I watch a marketer manufacture a return that looks better than the truth. The denominator has to be fully loaded, and when I price a job the real cost sits in four buckets, of which the production invoice is only one:

• Production: the brief, the shoot day, talent, location.
• Post-production: editing, motion graphics, captions, and every extra versioned cut.
• Distribution: media spend and placement.
• People: project management and the internal team hours nobody logs.

That last bucket is the one teams forget, and it is often the largest single line. The value side gets misread in the opposite direction, because a revenue-only view understates what a good asset does. Video feeds pipeline and shortens the sales cycle long after launch week is over, so counting only closed sales sells the whole thing short. Load the cost in full, then give the value its full credit too.

The metrics ladder: from views to revenue

Most of the metrics people report do not map to money. They stack into a ladder, and only the top rungs touch revenue.

Reach and engagement are the numbers marketers love to drop into a deck, and on their own they carry the least weight. Click-through rate is the rung I care about most, because it is the moment a viewer becomes a trackable lead, and it is the single most improvable lever on the top half of the formula. I report from the top of the ladder down, never from the bottom up.

Attribution: connecting a view to a sale

Attribution is the hard part, and hand-waving it is the reason video keeps getting under-credited on the P&L. This is the piece I watch teams skip, and skipping it is exactly why the channel ends up looking weaker than it performs. The toolkit I lean on:

• UTM parameters on every link out of a video.
• Dedicated landing pages so a visit is unambiguous.
• Hosting analytics (Wistia, Vidyard) piped into your CRM (HubSpot, Salesforce).
• A stated attribution model, picked on purpose instead of inherited by default.

That last point does more work than it looks, because the model you choose changes who gets the credit. The four common ones treat a video in the buying journey like this:

Attribution model	How it credits the video	Best when
First-touch	All the credit goes to the first video a lead saw	You want to know what drives awareness and top-of-funnel discovery
Last-touch	All the credit goes to the final touch before the sale	Short cycles where one video does the closing
Linear	Credit split evenly across every touch	You want a simple, defensible view across a long journey
Multi-touch (weighted)	Credit weighted by influence at each stage	B2B with many touches and an analytics stack that can back it up

Two of these choices matter more for video than for almost any other channel. A lot of video’s influence lands without a click at all, so view-through conversions and multi-touch models keep you from throwing that influence in the bin. The measurement window matters just as much: in B2B a demo watched in January can close in April, and if you credit it too early you undercount the video badly. I keep leading indicators like engagement and CTR on one clock, lagging outcomes like closed revenue on another, and report each against its own timeline.

Seven ways to actually improve ROI

There are only two levers here, grow the top of the fraction or shrink the bottom, and every move below pulls at one or both.

1. Hook hard in the first 3 to 10 seconds. Retention drives everything downstream, and I have watched strong videos die because the opening asked the viewer to be patient.
2. One clear CTA per video. A viewer you never ask to act does nothing.
3. Put the video where the conversion happens: landing, product and pricing pages, and sales emails, not only the social feed.
4. A/B test every thumbnail and hook against the last one, and keep the CTA in that test too.
5. Match length to funnel stage. Under 60 seconds for awareness, 30 seconds to 2 minutes for consideration, and longer only for high-intent late-funnel viewers, because completion collapses as videos stretch out.
6. Cut one shoot into many assets. This is the fastest way I know to drop cost-per-asset, a straight win on the denominator.
7. Set one objective per asset before you measure it, then judge it on the metric that objective actually moves.

That last habit heads off the mistake I see most often: judging a top-of-funnel awareness video on direct sales, which makes good work read like a failure.

A word on the “stats”

Two figures show up in almost every video-marketing pitch: “viewers retain 95 percent of a message from video versus 10 percent from text,” and “video lifts landing-page conversion by up to 80 percent.” Both trace back to small, old studies (roughly 2009 and 2011), and I treat them as marketing folklore rather than settled fact. I would not build a business case on either. Build it on your own attributed numbers instead, and on the reassuring reality that most marketers do report a real return. Wyzowl’s 2026 data has 82 percent citing good ROI, down from 93 percent the year before, which is a useful reminder that the return is earned on each video, not handed to you by the format.

At Moonb I build the attribution into the brief, so the video is designed around the outcome it has to move and nobody is reverse-engineering the ROI a year later. If proving return is where your video keeps falling down, see how Moonb works.