Video content ROI is what a video returns divided by everything it cost, and the second half is where most of these calculations fall apart. Production cost is easy to see because somebody invoiced for it. The rest arrives later, in storage you keep paying for and in the hours someone spends next quarter trying to get back to footage the company already owns. If you want a number you can defend in a budget conversation, the two inputs that matter most are the fully loaded cost of a finished minute and how many times that finished minute gets used again.
Cost per finished minute is mostly labor
Video budgets get read as a line item for the shoot. In practice the shoot is the short part. A ten-minute finished piece can absorb a couple of working days once the edit is counted, and the biggest block inside that edit is often not cutting at all. It is watching everything back to decide which pieces are worth using. That block scales with how much you shot, which is why ordering more coverage rarely makes the next edit cheaper.
If you outsource, this cost is legible. If you have an in-house editor, it is invisible, and invisible cost is the reason in-house video looks cheaper than it is. Take the loaded salary of everyone who touches video, divide by the finished minutes they shipped last quarter, and you have a number that lets you compare your own team to an outside quote. It is usually higher than people expect. Tool choice moves it less than people expect too, since the software that got faster mostly got faster at the assembly, which was never the expensive half. That split is worth understanding before you buy anything, and which parts of the work the current tools actually touch is the shorter version of it.
On the return side, pick something you can observe
Attribution for video is weak, and building a model that pretends otherwise burns a quarter. A brand film does not convert anyone in the session they watched it in. Rather than force video into a last-click frame, choose two proxies you can observe without argument and hold them steady for a year: something about attention (completion rate on a platform you control) and something about internal pull (how often sales, support, or recruiting ask for a video by name). Neither is a return. Both move when the work is good, and a stable imperfect measure beats a precise one you redefine every quarter.
Reuse is the number that moves ROI most
Every additional use divides the original production cost by one more use. Nothing else on the cost side has that leverage. So the question that decides your ROI is not how efficiently you shot, it is what share of last year's footage got used a second time.
Reuse rates stay low for a boring reason. The footage exists, and nobody can get to it. A team with a five-year archive and no way to look inside it is, in practice, a team with no archive, and the next brief turns into a new shoot for something that was already filmed twice. Some teams already buy their way out of this without treating it as an investment: paying per episode to have a long back catalogue transcribed is a purchase of retrievability, made after the fact and one file at a time.
Buying it once instead is the shape of the alternative. Making a library searchable by what is actually in the footage is a cost that lands once and then changes the denominator for everything already stored. Vivu sits at that layer, indexing each file one time after it is uploaded, so the effort of looking through the archive stops growing with the size of the archive. The practical form of this is the hunt described in the search for a video you know exists, repeated a few dozen times a year, priced at whatever an hour of your team is worth.
When this math is not worth doing
If you publish a handful of videos a year and everyone can name all of them, there is no reuse problem to measure and no archive to instrument. The same is true if your video is entirely paid social, where the creative is disposable by design and the ad platform already reports the only numbers that decide anything. Measurement earns its keep at the point where the library outgrew the memory of the person who made it.
What to do with the number
Run the two numbers on last year before you argue about this year's budget. If cost per finished minute is high and reuse is near zero, the fix is not a bigger budget or a faster editor, it is that nothing you own is reachable. If cost per minute is high and reuse is already decent, you have a production problem, and that is a different purchase. If both look fine, the honest answer is that video is working and the reporting exists to prove it, which is a better position than most teams asking this question are in. Teams whose real bottleneck turns out to be reachability usually end up comparing systems built for large libraries, and that comparison goes better once you know which of the two numbers you are trying to move.
FAQ
How do you calculate cost per finished minute of video?
Add the fully loaded cost of everyone who touched the video, including internal salary time, and divide by the minutes that actually shipped. Loaded means salary plus benefits and overhead, not the base number, and it includes the marketer who sat in on reviews, not only the editor.
Count shipped minutes, not delivered assets. A single shoot that produced one two-minute film and six cutdowns from the same footage is eight minutes of output, and treating it as one asset makes the number look far worse than it is.
Should video ROI include the cost of storing footage?
Yes, and the storage bill is the small part. Raw video storage is cheap enough per terabyte that finance rarely questions it, which is exactly why it accumulates without review.
The real carrying cost is the labor of navigating what you kept. If three people spend an hour a week looking for material, that is a larger annual number than the drive, and unlike the drive it grows every time you shoot.
What is a good reuse rate for video assets?
There is no published benchmark worth quoting here, and anyone who gives you one is guessing. Measure your own baseline for one quarter, then try to beat it.
The useful version is a count rather than a rate: pick ten videos you paid for last year and check how many were used again after their launch window. Most teams doing this the first time find the answer is one or two, and that count is more persuasive internally than any industry average.
How long should we wait before measuring a video's ROI?
Measure the attention numbers within a few weeks and the reuse numbers after a year. They answer different questions and they move on completely different clocks.
A launch-window report tells you whether the distribution worked. It says nothing about whether the asset earned back its production cost, because that depends on uses that have not happened yet. Judging a brand asset at thirty days is the most common way teams conclude video does not pay.