Claude Code Meta Ads Test: What the 7.5x Return Really Shows
By Matt Cullerton, founder of Mavric and Apso
A TikTok from @mr.paidsocial claims a creator gave Claude Code her video library, a Meta API key and her credit card, and that over seven days the agent spent $2,019.18 and produced $15,235 in gross sales. The 7.5x figure is correct arithmetic. It is a platform-attributed revenue multiple on media spend, and the repository the video recommends creates every ad paused for a human to review.
Does the 7.5x return add up?
15,235 divided by 2,019.18 is 7.545, which rounds to 7.5. Subtracting spend from sales leaves $13,215.82 above media cost, a 6.5x net multiple on the budget.
The problem is the word "return." Ads Manager reports return on ad spend, which counts one input. The product here is physical hardware, so cost of goods, shipping, payment processing and refunds all come out of that $15,235 before anyone reaches profit. The creator's own prompt, shown on screen, says "if you need any additional context about COGS or anything like that, let me know." Cost of goods never appears in the result the video reports.
How Meta counts those sales
Meta Ads Manager reports the conversions its own attribution model assigns to your ads. The default window is 7-day click and 1-day view. Under the view half, a purchase counts for Meta when the buyer saw the ad without clicking and then bought through any other path within 24 hours.
That matters for this creator. The clip shows an Instagram post with 183K likes, and her prompt says she has "always posted organic content" and holds "a library of high-performing organic content." Organic demand that converts during a paid window lands in the paid column. The video mentions no holdout group, geo test or baseline week, so there is no way to separate what the ads caused from sales that were already happening.
Is it really hands-off?
The repository is krusemediallc/arcads-claude-code. Its meta-ad-builder skill says: "Every ad is created PAUSED. Tell the user to review and un-pause in Meta Ads Manager." The same file says the skill creates ads only and does not create or edit campaigns, ad sets, budgets or audiences.
The video says the setup "made it so I didn't have to click a single button in Ads Manager." In practice someone has to build the campaign and ad set, set the budget, choose the audience, then open Ads Manager and unpause each ad. The agent writes copy and assembles creative into ad objects.
The credit card framing is also off. The repo asks for two environment variables, META_ACCESS_TOKEN with the ads_management scope and META_AD_ACCOUNT_ID. The card stays with Meta as the ad account's payment method, and the agent never sees it. What the agent gets is authority to create objects inside an account that bills a card already on file.
Who owns the repo
@mr.paidsocial is Caleb Kruse of Kruse Media LLC, and the GitHub account krusemediallc is the same person. The video says "go to this github repo" without noting that it is his own, and he sells a paid Skool community built around these workflows. That does not make the code bad, but the recommendation has an interest behind it.
The video says the repo has 1,100 stars. On 2026-09-29, two days after the video posted, the GitHub API reported 1,552 stars and 376 forks on a repo created 2026-04-01. Stars measure bookmarking and say nothing about whether the code produces profitable campaigns.
What it costs beyond ad spend
The repo describes itself as an "Arcads external API" workspace and requires an Arcads API key, with Meta credentials listed as optional and needed only for publishing. Arcads is a paid subscription with plans running from roughly $29 to several hundred dollars a month. Neither that nor a Claude Code subscription is inside the $2,019.18.
Is seven days enough to judge it?
$2,019.18 over seven days averages about $288 a day. Meta's learning phase targets roughly 50 optimization events per week per ad set before delivery stabilizes, and performance swings inside that window by design. A seven-day test is one learning phase, the noisiest week a campaign will have.
The video shows no control, no repeat run and no losing run. The narrator's own Skool-sale screenshot shows five ad sets, two of which spent money and returned a conversion value of $0.00. The rest of that table is filmed at an angle and cannot be read, so the $30,000 figure attached to it cannot be checked. His twelve years running Meta ads and over $150 million in spend are self-reported, and Meta publishes no per-manager figure to check them against.
How to run this safely
One frame shows the question "did you give it a limit on what it could spend?" and the video moves past it. Daily and lifetime budget caps live at the campaign and ad set level, which this repo does not touch, so you set them. The firmer stop is an account spend limit on the ad account, a manual setting configured once.
Meta's Platform Terms require you to protect access tokens and not transfer or share them. Running an agent on your own machine with a token for your own ad account fits those terms. Pasting that token into a hosted service you do not control does not. Creating a batch of paused ads through the Marketing API is supported use. Launching dozens of campaigns an hour on a young account, or driving a logged-in Ads Manager session through a browser, is the pattern associated with accounts disabled for unusual activity.
Bottom line
The 7.5x multiple is real arithmetic on Meta-attributed gross sales, not profit. The recommended repo creates paused ads and leaves campaigns, budgets and audiences to a person. Set an account spend limit before handing any agent an ads_management token, and judge the setup on more than one seven-day learning phase.
Sources
- krusemediallc/arcads-claude-code: the repository the video points to, owned by the narrator, MIT licensed.
- meta-ad-builder SKILL.md: states that every ad is created paused and that the skill does not touch campaigns, ad sets, budgets or audiences.
- Meta Platform Terms: Section 6 covers safeguarding data and not sharing access tokens.
- About the Learning Phase: Meta's explanation of why delivery stays unstable early on.
- Marketing API Access Tier update: Meta's post on the tiers that gate API-driven ad management.
- Arcads: the paid AI video ad product whose API key the repo requires.
Published September 27, 2026. Based on a TikTok by @mr.paidsocial. Full video notes (markdown)