Data · Creative durability

The 6% problem: why a handful of creatives eat the whole ad budget

By Orcool · July 9, 2026 · 6 min read

The number everyone is repeating

Motion's 2026 creative benchmark report put a figure on something media buyers already feel: at enterprise scale, a small sliver of the creative library, roughly 6%, ends up carrying most of the working budget. Framed that way, it reads like a targeting or picking problem. Somewhere in the review process, the theory goes, a handful of creatives get chosen well and everything else gets starved.

We wanted to know what was actually driving that number, so we went and measured our own panel instead of arguing about someone else's.

What we measured

We pulled every currently active ad creative for a fixed 36-keyword panel from the Meta Ad Library, US market, fully anonymous: 1,061 live creatives across 394 distinct advertisers. For each one we recorded how long it has already been running, then split the panel by creative format.

Dynamic and DCO creative is the format teams lean on hardest for volume, more than half of everything live in the panel. It is also the format with the shortest median run and the lowest share still alive past the 30-day mark.

ship volume mostly in the fastest-churning format  →  most of it dies young inside a month  →  spend concentrates on whatever is still standing

Why this explains the 6% number better than a picking story

If well over half of what a team ships is built in a format with a 33-day median lifespan, the math does the rest. Most of the fleet is gone within a month by construction, not by a review committee's judgment. Whatever survives past that point, largely the static and video minority in our panel, is what keeps absorbing budget simply because it is still there to absorb it.

That reframes the 6% finding. It is not mainly evidence of sharp picking. It is closer to a survival curve wearing a spend chart's clothing: a small share of creatives were never going to compete for long-term budget, because most of the fleet was never built to last long enough to compete at all.

The concentration is not proof that a few creatives were chosen well. It is proof that most creatives were built in a format that was never going to survive long enough to be chosen from.

What this means if you plan a testing calendar

Volume in the wrong format does not widen your winner pool. Shipping more dynamic/DCO variants adds more short-lived creative to a pile that is already 57% short-lived creative. It raises the count of things you tried, not the count of things that survive long enough to earn real budget.

Durability is a format property, not just a creative property. Our panel shows format alone moves median lifespan by a factor of three, from 33 days to 101. That is a bigger lever than most individual creative decisions, and it is measurable before a single test even runs.

Track days-live by format, the same way you track CPA. A concentration number like Motion's 6% tells you the outcome. A durability breakdown like ours tells you the mechanism, and a mechanism is something you can plan a format mix around instead of just reacting to after budget has already piled up on the same three ads again.

This is the same pattern we have written about from the other direction, in what per-impression pickers can and cannot tell you: a system that selects well among what exists still cannot fix a fleet where most of what exists was never going to last. The fix sits earlier, at the format and angle stage, which is increasingly the actual job description behind the creative producer becoming a GTM engineer.

Frequently asked questions

Methodology note: this panel is a single snapshot of currently active creatives pulled from the Meta Ad Library across a fixed 36-keyword set, US market, fully anonymous and sliced by format only, no advertiser named. Days-live means how long a still-running creative has already been live, a durability floor rather than a completed lifespan, since we do not yet observe creatives after they drop off. The 6% concentration figure is Motion's published 2026 benchmark, cited as external validation, not our own claim. Read both as a signal about format-level durability, not a guarantee for any single campaign.