What does a Performance Ads FTE own?
A Performance Ads FTE owns paid social and search end to end across Meta, Google and YouTube, TikTok, LinkedIn, and X. It builds creatives and audiences, structures campaigns, launches on approval, optimises daily, and reports against a target return on ad spend.
It is hired per ad account, so its caps, memory, and targets belong to that account. The role's own description of its job is that it spends your budget like its own, which is the standard it is measured against.
How does the spend cap actually work?
The daily spend cap is a hard limit the Performance Ads FTE cannot exceed, enforced by the platform rather than by a notification. This is the control that makes autonomous media buying deployable at all, and the distinction from an alert is the whole point.
An alert tells you the cap was exceeded. A hard cap means it cannot be. Budget changes are proposals the manager approves, and every change is logged with the reasoning, so the ceiling only moves when a person moves it.
What does a launch approval contain?
A launch waits in the approvals queue with the creative preview, the audience definition, the daily cap, and the attribution plan attached, after passing a policy pre-check against the platform's advertising rules. The manager approves the actual launch, not a description of it.
This holds at every autonomy level for new creatives and new audiences, because a launch is a first-of-its-kind action with asymmetric downside. What can earn autonomy is the routine optimisation of campaigns already running.
- Creative preview: the ad as it will appear
- Audience: the targeting definition in plain terms
- Cap: the daily limit for this campaign, inside the account cap
- Attribution plan: how results will be measured and reported
- Policy pre-check: whether the creative passes platform rules
How does it optimise once campaigns are live?
Once campaigns are live, the Performance Ads FTE proposes pausing losers and scaling winners each day, with the evidence for each proposal. At Level 0 every proposal waits for approval; a manager who has approved a class of decision repeatedly can promote it so the FTE acts inside caps.
That is the autonomy ladder applied to media buying: it proposes first, then earns the right to auto-pause, then to auto-scale within limits. The manager sets the pace of promotion from the approval history, not from the vendor's confidence.
Why does it report two attribution numbers?
The Performance Ads FTE reports platform-reported and blended attribution side by side because the two routinely disagree, and reporting only one is how media buying results get flattered. Platform numbers credit the platform generously; blended numbers compare total conversions against total spend.
Showing both is a form of honesty that also happens to be useful. The gap between the two figures is itself information about how much of the platform's claimed credit is real, and the FTE's weekly report puts that gap in front of the manager rather than choosing the flattering line.
What missions can you give it?
Beyond the daily routine, the Performance Ads FTE takes missions. Launching a new offer across three channels is a production mission: creatives, audiences, and budgets proposed per channel and launched only on approval. Return fell this week, why, is a diagnostic mission answered with ranked causes and evidence.
Diagnostic answers are deliberately framed as hypotheses ranked by evidence rather than as certainties: creative fatigue, audience overlap, a bid change, a landing-page problem. It proposes the experiment that would settle the question rather than asserting the answer.
What does a Performance Ads FTE need from you, and how are the two budgets kept apart?
A Performance Ads FTE needs a connection to each ad account, scoped per property with the permissions you grant, a spend cap per account that it cannot exceed, a Brand Brain with the creative rules and prohibited claims, and a remit that names which actions keep their gate permanently: new campaign launches, budget increases above a threshold, and any creative that touches a regulated claim. It also needs the honest conversion data your site can supply, because attribution is only as good as the events it can see.
Two budgets are kept apart by design. The ad spend budget is your money on the ad platform, billed by that platform, bounded by the cap you set at the account level and visible in the morning brief. The FTE's own budget is tokens from the organization's wallet for its analysis, drafting, and reporting, bounded by its per-FTE budget. The brief shows both, and neither can borrow from the other. Approving a launch costs no tokens; the launch costs ad spend only once you have approved it.