THE SHORT ANSWER
Define the risks this campaign cannot accept, then review six to twelve months of public content, pinned posts, sponsorships, disclosure behavior and verifiable public events. Record links, dates, market context and severity, then have a person decide relevance to the product and audience. “No evidence found” is not “zero risk,” and an old incident should not be judged without timing and subsequent behavior.
Write campaign-specific risk boundaries first
Food, finance, children’s products and entertainment campaigns face different risks. Express exclusions as testable conditions, such as unsupported health claims, hateful content, hidden commercial relationships or an active competitor exclusivity term, rather than “positive image.”
Assign each condition to exclude, escalate or record. This keeps personal preference from mixing ordinary disagreement with material brand risk.
Review content and sponsorships in a fixed window
Start with the last 90 days, then expand to six to twelve months to identify repeated patterns. For each platform, record sample size, oldest and newest dates, inaccessible content and whether sponsorships are clearly disclosed.
Separate one incident from a sustained pattern. When a claim comes from a repost or excerpt, return to the original material and full context instead of treating a search snippet as final evidence.
Review competitor and rights conflicts separately
List recent brands, publication dates, product category, possible exclusivity terms and ads that remain active. A creator may naturally discuss several brands, while signed exclusivity or a live competitor ad can directly affect timing.
Carry unknowns into outreach instead of guessing without contract information. Ask about current exclusivity, live periods and advertising authorization rather than a vague question about competitor work.
Give the decision an owner and expiry date
A campaign owner can clear low risk, while medium or high risk should move to brand or legal review. Save reviewer, date, evidence and final reasoning, then set a refresh point.
Creator content continues to change. For long engagements, review before contracting and again before publication, reopening the decision when material new information appears. QuickKOL should show “reviewed through this date,” not a permanent safety badge.
Use evidence levels so search results do not become facts
Classify evidence as original content, platform or brand record, reliable reporting, secondary retelling or unverified claim. Only the first three can support a formal conclusion; the last two trigger further review. Save URL, capture date and the market, product or audience affected.
Brand safety also includes commercial-content quality. TikTok’s commercial-content resources describe expectations for creator work, so review disclosure, product claims and presentation together. For each finalist, inspect at least 20 recent posts and all pinned content, then list sponsored work from the previous six to twelve months.
Risk evidence and action
Campaign owners define levels; every conclusion keeps evidence and date.
| Finding | Action | Record |
|---|---|---|
| Low | Owner clears | Sample and date |
| Medium | Escalate | Context and impact |
| High | Pause candidate | Primary evidence |
| Unknown | Confirm in outreach | Question and due date |
CHECKLIST
Brand-safety human review checklist
Use this review order rather than an automatic score.
- Risk boundary and market defined
- Recent content window covered
- Pinned and high-reach content reviewed
- Sponsorship, competitors and exclusivity listed
- Disclosure and claims checked
- Primary evidence separated from hearsay
- Decision owner and escalation confirmed
- Review expiry set
Sources and further reading
External pages can change. Recheck platform rules, contracts and local law before launch.