Signals · Creative review

The concept we passed on shipped as a competitor's ad three weeks later.

By Orcool · July 8, 2026 · 5 min read

Two jobs pretending to be one job

Every creative review is doing two different jobs at once, whether anyone names them or not. Job one is taste: does this look like our brand, does it feel premium, would the team be proud to ship it. Job two is mechanics: does this reuse a structural pattern, a format, a hook rhythm, an editing style, that is currently pulling attention on the platform it will run on, regardless of what the ad is actually about.

Most review cycles only ask job one's question out loud, and mechanics gets judged by the same gut reaction that judges taste. That works fine right up until a concept comes along that is mechanically strong and looks nothing like the brand it is meant to represent.

The concept that got passed over

One of the stranger concepts we put together for a client, a global consumer app running creative across dozens of countries, was built directly off a trend format that was live on TikTok at the time: humanlike vegetables, played completely straight, carrying a small dramatic plot that had nothing to do with the product category. On paper it read as borderline insane. The client passed, not because the underlying strategy was wrong, but because the visual did not look like anything the brand would normally put its name on.

What happened three weeks later

A direct competitor in the same market shipped almost exactly that mechanic as a live paid ad. Same trend format, same absurdist premise, adapted to their own product. We were watching the same signal feed that had flagged the trend in the first place, so we saw it land. The mechanic worked because the trend was still fresh enough to read as native content instead of an ad wearing a costume, and it faded from the feed a few weeks after that, on schedule, the way trend-format ads usually do.

Nobody copied anybody. Two teams were reading the same public signal. One team acted while the format was still live. The other team was still in review.

Taste and mechanics do not run on the same clock

Brand taste barely changes month to month. A concept that looks off-brand today will usually still look off-brand next quarter, and that is a legitimate thing to protect. Trend mechanics move on a completely different schedule. A format that is earning attention this week is frequently saturated within a month, once enough advertisers in a category adopt the same pattern or the platform simply stops rewarding it as novel.

A review cycle with multiple stakeholders and a scheduled meeting is built at taste speed. Run a trend-sourced concept through it and the trend is often gone by the time a decision gets made. The concept was not wrong. The clock ran out on it during review, not during performance.

trend flagged signal feed catches the format  →  concept built mechanic applied to your product  →  window closes saturation, weeks not months

The "insane" reaction in a review room is not automatically a red flag. Sometimes it is the tell that a concept is doing its job correctly: borrowing a mechanic that is working right now, instead of trying to look like your last ten ads.

What this means for how you review creative

The fix is not lowering the bar on taste. It is running two separate approval lanes instead of one. Give trend-sourced concepts, the ones that look strange because they borrow a mechanic rather than express brand identity, a fast, time-boxed decision path measured in days. Keep the slower, taste-first process for concepts meant to represent the brand's identity over the long term. Judging both on the same clock is exactly what causes a mechanically sound concept to die of natural causes in a review meeting.

This is part of a broader shift we have written about before: the person approving creative is increasingly reading signal rather than approving on instinct alone, which is the creative producer becoming a GTM engineer. Automated pickers can rank a library you already have, but as we covered in selection is not creation, they cannot flag a trend mechanic before it is obviously trending. That judgment call, and the speed to act on it, still sits with a person watching the feed.

Frequently asked questions

Methodology note: this account describes a real concept and review cycle from our work with a global consumer app client, anonymized here. The competitor's ad is described by category and market rather than named, and the three-week timing is approximate, based on when our signal feed first flagged the trend and when the competitor's ad appeared in that same feed.