From the desk:Listing Content / Social Media

Cost per contract on short form video is $2,300. Volume, or an editor?

Buyer list build, seven months of data, and I'm at the point where the next dollar has to go somewhere specific.

Spend: $1,150 a month all in. That's $650 to a part-time editor at roughly $28 an hour for about 23 hours, $200 in tools and scheduling, $300 for a copywriter who does hooks and captions on a per-batch basis. Seven months, $8,050 total.

Output: 14 to 18 short videos a month, mostly market breakdowns and walkthroughs of what I'm buying. Cumulative 41 qualified leads, meaning people who answered a phone screen and had a price range and a timeline. From those, 9 appointments, 3 under contract, 2 closed so far. So $8,050 across 3 contracts is roughly $2,680 per contract, or $2,300 if the third closes.

My assignment fees run $9,000 to $14,000, so the channel is paying, thinly, and it took five months to produce anything at all.

The fork. Option A, push volume to 30 plus a month by moving first-pass editing to AI tools and keeping the editor for finishing only. Same $650, roughly double the output, quality drops somewhere I can't predict. Option B, keep volume flat and put the $300 copywriter money into more editor hours instead, on the theory that my conversion from view to lead is the weak link rather than reach.

What makes me hesitate on A is that my best three videos accounted for 19 of the 41 leads. That looks like quality carrying the whole thing, and doubling mediocre output may just add noise. But my worst ten videos cost the same to make as my best three, and I couldn't have told you in advance which would be which.

Answer

2jordan_ellis2ActivePRO6

The 19 of 41 from three videos is the whole dataset and you're reading it backwards. That distribution is what you'd expect from a channel where outcomes are heavily skewed, and the way you get more skewed winners is more attempts, because you already told us you can't pick them in advance. That argues for A.

But A as you've written it isn't a volume test, it's a volume-and-quality-drop test with two variables moving. Run it narrower. Keep the editor at 23 hours on 15 videos a month, and separately produce 15 AI-first-pass videos that get 20 minutes of human finishing each, which is roughly 5 extra hours. Tag them. In 90 days you'll know whether the cheap batch produces leads at any rate at all. If the AI batch generates 1 or 2 leads a month, you've bought reach for near nothing and you scale it. If it generates zero across 45 videos, that's a real answer too and you stop.

On dropping the copywriter, I'd keep that money where it is. Your winners were probably winners because of the first three seconds, and the hook is the cheapest thing to buy and the highest variance part of the video. Cutting the hook budget to fund more editing hours on unwatched middles is backwards.

One caution on the $2,300. Two closed out of three contracted, and 41 leads over seven months is a small enough sample that your cost per contract could be half or double that and you couldn't tell. Don't let a number that unstable pick the strategy for you. The 90 day tagged test gives you something you can actually compare.

3 replies

4diane_kowalskiPassivePRO9

pike's split test is right and I'd add one thing about the tagging. Tag by format too, not only by production method. You said market breakdowns and walkthroughs. If your three winners were all one of those, that's a bigger lever than editing hours and it costs nothing to act on. I keep a spreadsheet like this for land content and the format column explained more than the effort column ever did.

1raj_mehta14CapitalPRO6

Mostly out of my depth on the content side, but the part I understood is the lag. Five months before the channel produced anything, and now you're proposing a 90 day test. So you're 10 months into an experiment that has produced 3 contracts, and you'll be 13 months in before the next decision. Do you have the cash and the patience for that timeline, and does the assignment side keep the lights on meanwhile? That's the question I'd want answered before the editing one.