Notes from running campaigns at scale.

Most reshoots don't happen because a creator is bad at their job. They happen because the brief left too much open to interpretation, and the creator filled the gaps with their own best guess — which may or may not match what the brand actually wanted.

A brief that actually works has a handful of non-negotiables: the single message the piece needs to land, the format and length, what's mandatory (a product shot, a claim, a hashtag) versus what's flexible (tone, setting, wardrobe), and at least one reference piece of content that shows the vibe rather than describing it in adjectives.

The part teams skip most often is telling creators what "good" looks like in their specific niche — a beauty creator and a finance creator succeed on completely different signals, and a generic brief treats them the same. The fix isn't a longer brief. It's a sharper one, built per content category rather than copy-pasted across every creator on a campaign.

This is also where a lot of budget quietly leaks: every round of "this isn't quite it" costs time, goodwill with the creator, and often a paid revision. A brief that removes ambiguity up front is the cheapest quality-control step available to any campaign — cheaper than any review process downstream of it.

"Which creator tier should we use" is the wrong first question. The right first question is what stage of the funnel the campaign is actually trying to move — awareness, consideration, or conversion — because each tier is built to do a different job well.

Nano creators (roughly 1K–15K followers) win on trust. Their audience is small enough that the creator still feels like a peer, not a media outlet, which makes their recommendations read as genuine rather than sponsored. They're the right call for authentic UGC and category-level social proof at scale, especially when you need volume of believable content rather than one big splash.

Micro creators (15K–100K) sit in the sweet spot of reach and engagement. They've usually developed a consistent content style and a loyal audience without losing the personal-feeling connection nano creators have — which makes them strong for consideration-stage content that needs to actually explain a product, not just show it.

Macro creators (100K+) buy you reach and cultural salience fast. They're the right tool for a launch moment or a top-of-funnel push where the goal is "everyone sees this exists," not necessarily deep persuasion — the trade-off is usually lower relative engagement and higher cost per creator.

Most campaigns that actually work don't pick one tier — they layer them: macro for initial reach, micro for credible explanation, nano for volume and social proof at the comment-section level. The mix is the strategy, not any single tier in isolation.

When a campaign slips, the instinct is to blame the creator. In our experience running campaigns at volume, that's rarely where the real problem sits. The real problem is almost always a missing checkpoint somewhere between "the creator sent a draft" and "the brand actually saw it."

At low volume — a handful of creators, one campaign — an account manager can eyeball every draft and catch issues before they become deadline problems. That doesn't survive contact with scale. Once you're running fifty pieces a month, and definitely once you're running thousands, informal review breaks down exactly when you need it most.

The fix we run internally is a three-gate structure: a concept gate before anyone films anything, a draft gate before a piece is considered "done," and a final gate before it ships to the brand. Each gate has one job and one owner, which means a problem gets caught at the cheapest possible point to fix it — a wrong concept costs an email, a wrong final cut costs a reshoot.

The unglamorous truth is that quality control in influencer marketing isn't a creative skill — it's an operations problem. Brands that treat it that way, with real checkpoints and real owners, stop missing deadlines. Brands that treat it as "the creative team will catch it" usually don't.