A reel that “performed” is not a strategy. It is a signal. The only question that matters is whether a next step happened: a save, a reply, a profile visit, a booked call. Views without a next step are a hobby. Hobbies are allowed. They should not be in the media budget.
Teams get stuck because they hire for volume. Thirty clips a month, none of them allowed to die. That is a factory. Factories need a product. A test needs a hypothesis and a kill rule. This article is the test.
One claim. One next step.
Name the problem in the first second. Show the result you sell. Tell them what to do. Then measure that action. If you cannot say the action in one verb — book, buy, reply — the reel is a mood board. Mood boards do not pay salaries.
Meta’s creative guidance is practical once the offer is clear. It will not invent an offer for you. Pair it with a kill rule so the algorithm is not promoted to strategist.
The test grid
I run a simple grid: three claims, two proofs, one next step. That is six clips, not sixty. Each clip runs long enough to learn, not long enough to become a brand anthem. Winners graduate. Losers die without a post-mortem novel.
Graduation means the claim moves onto a landing path, a follow-up sequence, and — only then — paid amplification. If you boost a clip that has no path, you have bought more of a dead end. Read the leak essay before you put money behind a winner.
What “good” looks like
Good is not a view-through rate you cannot cash. Good is a booked call, a reply with intent, a product in the cart. Think with Google’s video work is useful for craft. It is not a substitute for your scorecard. If the three numbers do not move, the reel did not work, however pretty the comments.
If you are hiring someone to “do reels,” hire a marketer who can tell you which three deserve spend. Activity is not the job. The talent desk on this site exists for that reason — profiles without public phone numbers, introductions that come through the desk.
Start this week
One phone. One claim you already believe. One link that can take a lead. Publish. Wait. If it works, write the landing line to match. If it does not, you learned cheaply. That is the whole system. The wall of clips on this site is that system, recorded.
If you do not have a site yet, do not start a content factory. Mock the page on the SME desk, pick a colour, write one next step, then film the claim that matches it. A reel that sends people to an Instagram bio is a leak you filmed yourself.
When you are ready to hire, send a brief — not a request for “someone who can post daily.” Daily is not a strategy. Three tests with a kill rule is.
Also read: AI is a multiplier. It is not the strategist.
Also read: A founder scorecard with three numbers