Working out whether video paid for itself

At some point somebody asks whether the video thing is worth it. Here is a defensible way to answer that does not require pretending to more precision than exists.

The costs, including the ones people skip

• Software. The obvious one and usually the smallest.
• Production. Time counts even when no invoice was raised. A day of someone’s month is a real cost.
• Setup and maintenance. Tagging, placing, the monthly pass. Small per video and substantial across a year.
• Performance. If the pages got measurably slower, that has a conversion cost. Usually zero if it was done properly; not always.

For most small stores the honest total is dominated by time, not fees — which is exactly why the batching and the workflow matter.

The return

Start with attributed revenue from video, then adjust downward, because some of those orders would have happened anyway. Somebody who watched the video and bought might well have bought without it.

A workable approach: compare conversion rate on sessions that saw video against comparable sessions that did not, and use the difference rather than the raw attributed total. It is more conservative and it survives scrutiny.

Then take gross margin on the incremental revenue, not revenue. Video that generates £10,000 at thirty per cent margin returned £3,000.

A worked example

A store with 20,000 monthly sessions, converting at 2%, average order £60. Video is on the product pages. Sessions that engage with it convert at 2.6%; the engaged group is 15% of traffic.

That is 3,000 sessions at an extra 0.6 percentage points, so roughly 18 additional orders a month, about £1,080 in revenue, about £324 in margin at 30%.

Against a monthly cost of, say, £29 in software and half a day of time, that is comfortably positive — and it is the kind of number worth recalculating rather than assuming, because it depends heavily on the engagement share.

The second-order return

Two things the calculation misses, both real and both hard to quantify: fewer returns, because shoppers who saw the product in motion had accurate expectations; and less support load, because the question the video answers is a question nobody emails about.

Returns are the larger of the two and worth measuring separately — it is often where the actual money is.

The honest caveats

Attribution overstates. Engaged shoppers were already more likely to buy. Seasonality moves everything. Do not present a number to two decimal places; present a range and say what it assumes.

The useful version of this exercise is not a precise figure. It is knowing whether you are in the territory of “clearly worth it”, “clearly not”, or “too close to tell” — and all three are actionable answers.