There is a recurring argument in ad teams about where the real work lives. Media buyers say that bad targeting wastes good creative. Creative teams say that perfect targeting cannot save a bad ad. Both sides are right about something. But they are not equally right.
If you had to put a number on it, creative drives somewhere between 60 and 80 percent of ad performance variance. Targeting, bidding, and budget allocation fight over the rest. Nielsen, Analytic Partners, and several platform-level studies have landed near this range repeatedly over the past decade. The number has moved in creative's direction as platform algorithms have improved, not away from it.
What the algorithm shift actually changed
Five years ago, knowing how to build tight audiences on Facebook was a genuine edge. Custom audiences, lookalikes, interest stacking: a skilled media buyer could find pockets of efficiency that competitors missed. The platform rewarded that precision.
That edge has eroded substantially. Meta's broad targeting push, Advantage+ audiences, and ongoing signal loss from iOS changes have pushed the algorithm toward doing its own audience discovery. You give it creative and budget. It finds the people. The better the creative, the better the people it finds.
Media buying skill still matters. Funnel structure, campaign architecture, budget pacing, bid strategy: those decisions have real consequences. But the leverage point has shifted. The algorithm is better at targeting than most humans now. It is not better at knowing which ad will stop someone mid-scroll.
Where media buying still has leverage
Budget efficiency at scale. When you are spending enough that campaign structure and bid strategies create meaningful cost differences, media buying decisions compound. Knowing when to consolidate campaigns, when to let the algorithm learn versus when to intervene, how to structure prospecting and retargeting, how to pace budgets across a launch window: these decisions are worth real money at volume.
Measurement and attribution also require expertise that the algorithm does not supply. Knowing which signal to trust, how to run incrementality tests, how to read results without fooling yourself: that is media buying work that actually matters and that creative people are often bad at.
But notice what those things have in common. They are operational and analytical. They reduce waste. They do not generate lift. Creative is the only input that creates demand. Media buying allocates it more or less efficiently.
The creative leverage problem most teams ignore
Most ad teams are structured to solve the wrong problem. Significant headcount and tooling goes toward media buying: dashboard access, bidding tools, audience research, reporting. Creative gets a smaller slice. A designer, maybe a video editor, briefs written from gut feel.
The result is that the thing with 70 percent of the leverage gets 20 percent of the attention. And because bad creative feels like a targeting problem ("we just need a better audience"), the diagnosis is often wrong too.
The teams getting consistent creative lift treat it as a research function, not just a production function. Before making anything, they study what is already working in the category. Competitor ads get saved, tagged by hook type and offer structure, and pulled into briefs. Testing runs one variable at a time with enough volume to actually read the result. The question going in is not "what do we want to say" but "what has the market already responded to."
What a creative research habit looks like in practice
The best creative teams spend real time in the Meta Ad Library every week, not just when launching a new campaign. They know what competitors are running, how long those ads have been active (a reliable proxy for profitability), and what patterns repeat across different brands in the category.
Long run time is the signal worth paying most attention to. An ad that has been running for sixty days on a direct-response budget is almost certainly making money. The advertiser has had every incentive to turn it off and has not. That ad is worth studying: the hook, the offer, the format, the proof elements, the call to action.
The teams doing this systematically use tools to save and organize what they find. Screenshots in a Slack channel disappear. A searchable library with tags, notes, and context does not. When it is time to brief a new batch of creative, the library is where the brief comes from: real evidence about what the market has already responded to, not a conversation about what feels right.
The practical split
None of this means ignoring media buying. A great creative running in a broken campaign structure underperforms. Attribution chaos makes it impossible to know what is working. Poor budget pacing wastes the learning phase.
When resources are constrained, the question is where to put more. If your creative research process is "we look at competitor ads occasionally and then brainstorm," that is the gap worth closing first. If your media buying is already competent, squeezing more out of it returns less than building out the research side.
Creative is where demand gets created. Media buying is where it gets allocated efficiently or wasted. That distinction is worth actually building your team around, not just agreeing with in theory.