Cost comparisons in this industry are mostly unfalsifiable, and that is not an accident. “Save 90% on video production” survives scrutiny only as long as nobody asks against what, over what period, or including whose time.
This post is the method we will use every time we publish one. It has no results in it. Publishing the method first means the next post cannot quietly move the goalposts, and it means you can tell us the method is wrong before we have built a conclusion on top of it.
The unit is a year, not a video
Per-video pricing flatters whoever has the lowest marginal cost, which is always the automated tool. But nobody buys one video. Teams buy a cadence: a launch video every quarter, a changelog clip every sprint, a handful of tutorials when the docs fall behind.
So the unit we compare is twelve months of a stated cadence, and the cadence is written at the top of every comparison. Change the cadence and the ranking changes — that is a real property of the market, not a flaw in the measurement, and hiding it behind a per-video figure is the trick we are trying to avoid.
What goes into the number
- The invoice
What the vendor charges, at their published rate, on the date we checked. Quoted rates from a sales call are marked as such and are worth less as evidence, because we cannot show you the page they came from.
- The revisions
Priced separately, because this is where the agency and freelance numbers actually live. A two-round revision policy and an unlimited one are not the same product at the same price.
- Your own hours
Briefing, review, and the back-and-forth. Costed at a stated blended rate, which we will name every time. This is the line most comparisons omit, and omitting it is what makes DIY tools look free.
- The turnaround
Not a cost, but recorded alongside one. A cheaper video that lands three weeks after the launch is not a cheaper video.
Assumptions we state every time
A comparison is only checkable if you can disagree with its inputs. So each of these appears in the post itself, not in a methodology page nobody opens:
| Assumption | Why it moves the answer |
|---|---|
| Cadence (videos per year) | Shifts the ranking between per-seat and per-video pricing entirely |
| Blended internal hourly rate | Decides whether “free” tools are actually free |
| Revision rounds included | The main hidden cost in agency and freelance quotes |
| Video length and complexity | A 30-second social cut and a 4-minute explainer are different products |
| Date the rates were checked | Rate cards move; a number without a date is a rumour |
Where our arithmetic could be biased
We sell one of the things being compared. That is a structural conflict and no amount of care removes it, so the honest move is to name the specific ways it could show up:
- We know our own tool better than the alternatives. Our timings on CoAnimator are best-case and someone else’s timings on a competitor may be worse than that tool deserves. Where we can, we will use a vendor’s own published figures for their product rather than our attempt at it.
- We pick the cadence. A cadence that suits us is easy to choose accidentally. We will publish the crossover point — the cadence at which the ranking flips — rather than only the cadence where we win.
- We choose what counts as comparable quality. This is the softest input and the hardest to defend. The videos go in the post so you can judge for yourself.
Not that our numbers will favour us. That you will be able to tell when they do.
What happens when we are wrong
The correction goes at the top of the post, dated, with the original claim left visible underneath. If a comparison turns on an assumption that a reader shows us is unreasonable, we rerun it with theirs and publish both.
The first comparison built on this method is being put together now. It covers a quarterly launch cadence with two revision rounds, and it will state every one of the inputs above at the top of the post.