How to Vet a YouTube Creator Before a Sponsorship (2026 Checklist)

Sponsoring the wrong creator rarely fails loudly. The video goes out, the views look fine, and nothing happens: no sign-ups, no sales, no comments from anyone who looks like a customer. Most of the time the creator wasn’t fraudulent. They just weren’t the audience you thought you were buying.

This guide is for founders, marketers and agencies, especially in software and AI tools, who are choosing YouTube creators to sponsor or recruit as affiliates. It covers what you can check from the outside, what to ask the creator for, the red flags that suggest an inflated audience (and their innocent explanations), what the law says about fake followers, and how to write a deal whose numbers can’t drift. The sources are YouTube’s own help pages, the US Federal Trade Commission, and named industry reports, with their dates.

Why subscriber counts mislead

The subscriber number is the first thing everyone looks at and the least useful. YouTube says so itself in its analytics help: subscriber count “is not the most accurate way to estimate the size of your active audience”, and it’s common for viewers to stay subscribed to channels they no longer watch at all.

There’s also less precision than it looks. Above 1,000 subscribers, the public count is a shortened version: a channel with 1,234,567 subscribers shows “1.23M”. YouTube switched to these abbreviated counts in 2019, and its Data API returns the same rounded figure, so every third-party tool works from it too. And since November 2021, dislike counts are hidden from the public, though creators still see theirs in YouTube Studio.

A channel’s recent videos tell you far more. That’s why our own methodology judges creators on average views of their last 10 posts, not lifetime followers.

What you can see publicly, and what only the creator can see

Everything in the top half you can check today without contacting anyone. Everything in the bottom half you have to ask for.

Step 1: Look at the last 10 videos, not the best one

Open the channel’s videos tab, sort by latest, and look at the views on the last 10 long-form uploads. Take the median, not the average, so one viral video doesn’t skew the picture. Compare it with channels of similar size in the same niche, not with a universal benchmark.

There isn’t a reliable universal benchmark to compare against. We couldn’t find an authoritative, dated views-to-subscribers ratio from YouTube or a regulator, and the widely repeated ones don’t trace back to a source. The closest measured data we found comes from SponsorRadar, a sponsorship-tracking vendor. It publishes median views of sponsored videos by channel size, from about 1.5 million detected deals over the last 12 months (figures as of 7 October 2026):

Treat these as a rough sense of scale from one vendor, not a target to hold creators to.

Engagement rates need the same care, because tools calculate them differently. Iqfluence’s 2026 study divides likes and comments by subscribers and finds a median of 0.74% for channels with 1K–5K subscribers, falling to 0.08% above 5 million. CreatorDB’s June 2026 report measures against active followers and finds 3.2% for mid-size channels and 2.5% above 1 million. Both are reasonable; they just aren’t comparable. Pick one tool, use it for every creator, and compare like with like. What holds across every dataset is that smaller channels engage more per follower, which is one reason mid-sized creators are often the better buy.

Step 2: Read the comments

Comments are the hardest metric to fake well and the most revealing. Scroll the comments on three or four recent videos and ask:

• Are people asking real questions about the topic, the tool or the creator’s method?
• Does the creator answer them?
• Do the commenters sound like your customers? A software brand wants people asking about setup and pricing, not just “great video”.

Generic praise on its own isn’t proof of anything; plenty of real viewers write “nice video”. The red flag is a pattern: a wall of interchangeable compliments, emoji-only replies posted within minutes of upload, or comments in languages and from places that don’t match the channel. An industry fraud guideline from the Influencer Marketing Council, as summarised by law firm Frankfurt Kurnit, puts it simply: comments that “seem odd, generic, or off-putting” may suggest they’re fake. Watch for engagement pods as well: groups of creators who agree to like and comment on each other’s posts, which inflate engagement without adding customers.

Step 3: Check the subscriber history

Third-party trackers such as Social Blade chart a channel’s daily subscriber and view history. You’re looking for a sudden jump that doesn’t line up with any video, collaboration or press mention, which can mean subscribers were bought.

Two caveats prevent false alarms. Because public subscriber counts are abbreviated, gains on larger channels show up in steps, sometimes weeks apart, and a step can look like a spike. And a sudden drop is often a good sign: YouTube removes subscribers it identifies as spam, and its help pages say subscribers gained by artificial means, such as buying them from a third-party service, don’t count towards the total.

YouTube’s fake engagement policy bans anything that artificially increases views, likes, comments or other metrics, and says artificial traffic won’t be counted and can lead to strikes against the channel. Repeated or severe abuse can get a channel terminated.

Step 4: Check how they handled past sponsors

Look at the creator’s last few sponsored videos. Do they say the sponsor out loud near the start, show it on screen and carry YouTube’s paid-promotion label? The Data API exposes that label as a field on each video, so it can be checked at scale, though it only shows creators who ticked the box.

This isn’t just good manners. Under the FTC’s Endorsement Guides, “An advertiser may be liable for a deceptive endorsement even when the endorser is not liable”, and advertisers are expected to monitor their endorsers’ compliance. The FTC’s answers to common questions add that its enforcement focus “usually will be on advertisers or their ad agencies and public relations firms”. A creator who discloses well protects you. Our guide to disclosing affiliate links and sponsorships is a useful thing to send creators you work with.

Step 5: Ask for the private numbers

Once a creator looks promising, ask for their YouTube Studio data. Serious creators expect this and usually have a media kit ready. Ask for the last 28 and 90 days of:

• Engaged views and average view duration, for the channel and for their recent videos
• Audience retention on two or three recent videos, especially the “intro” figure: how many people are still watching after 30 seconds
• Monthly audience, and the split between new, casual and regular viewers
• Top geographies, age and gender, to check the audience matches your customers
• Traffic sources, including any views from “YouTube advertising”, which means the creator paid to promote the video
• The same figures for past sponsored videos, which predict your results better than anything else

Screenshots can be edited, so the cleaner option is temporary read-only access. YouTube’s channel permissions include a “Viewer (Limited)” role that can see analytics but not revenue, and YouTube describes it as safer than sharing a password.

One benchmark is worth knowing when you read the data. YouTube says half of all channels and videos have an impressions click-through rate of 2% to 10%. A very high click-through rate paired with low average view duration and falling impressions is, by YouTube’s own description, a sign of misleading titles or thumbnails: people click, feel misled and leave.

Red flags, and their innocent explanations

One flag rarely proves anything. Three together usually do.

What the law says about fake followers

Since October 21, 2024, the FTC’s Consumer Review and Testimonial Rule has made it unlawful to buy or sell fake indicators of social media influence, such as followers, subscribers, views, likes and comments generated by bots, fake accounts or hijacked accounts. It applies when the buyer “knew or should have known” they were fake and they misrepresent influence “for a commercial purpose”. The full text is in 16 CFR Part 465.

For brands, the FTC’s own Q&A on the rule is reassuring: if a campaign you paid for was secretly padded with fake followers, “you wouldn’t be liable unless there were some red flags that should have tipped you off.” Hiring a creator who happens to have fake followers doesn’t make you liable under the rule either. Courts can impose civil penalties for knowing violations, currently up to $53,088 per violation; the FTC kept the 2025 figure unchanged for 2026.

The FTC has acted on this before. In 2019 it settled with Devumi, a company that, according to the FTC, had more than 4,000 sales of fake YouTube subscribers and over 32,000 sales of fake YouTube views.

What third-party tools can and can’t tell you

Vetting tools save time, but know what they’re working from. Outsiders can’t see who subscribes to a channel: YouTube’s Data API only lists a channel’s subscribers in a request authorised by the channel itself. Any “fake subscriber percentage” a tool shows you for YouTube is a model’s estimate, not a count.

• Social Blade tracks daily statistics and growth across platforms, and is useful for subscriber and view history.
• HypeAuditor offers YouTube channel audits with a quality score that weighs subscriber authenticity and view drops.
• Modash and Upfluence are creator search and vetting platforms covering YouTube, Instagram and TikTok.
• The YouTube Data API gives you public figures (views, likes, comment counts, abbreviated subscribers and the paid-promotion flag) straight from YouTube.

Use tools to build a shortlist, then use the creator’s own Studio data to decide.

Structure the deal so the numbers can’t drift

Most sponsorships are paid in one of four ways: a flat fee per video, a price per thousand views (CPM), affiliate or revenue share, or a guaranteed number of views over a period. Published price benchmarks vary so widely, and are so rarely sourced, that we won’t repeat them; SponsorRadar notes that its own CPM ranges come from published deals and agency rate cards rather than measured fees, because creators don’t publish what they’re paid. Ask the creator for their rate card and compare it with the median views you measured in Step 1.

Whatever the model, put these in writing:

1. Which metric counts. Public views or engaged views, over what period, and whether views from YouTube advertising are included.
2. Disclosure. Spoken and on-screen near the start, YouTube’s paid-promotion box ticked, and a disclosure next to any links.
3. Data access. The right to see the video’s analytics after publication. The UK advertiser body ISBA’s influencer code of conduct has creators agree to reasonable audit requests and to fraud-checking technology.
4. Usage rights. Whether you can reuse the video in your own ads, and for how long.
5. What happens if fake engagement turns up. A clear exit and refund clause.

A 15-minute vetting checklist

1. Median views on the last 10 long-form videos, compared with similar channels.
2. Upload consistency over the last 90 days.
3. Comments on three or four recent videos: real questions, creator replies, the right audience.
4. Subscriber history for unexplained spikes.
5. Past sponsored videos: disclosed out loud, on screen and with YouTube’s label?
6. Studio data for 28 and 90 days: engaged views, average view duration, geography, traffic sources.
7. The same data for past sponsored videos.
8. A written agreement naming the metric, the disclosure and your data access.

What to do next

If you’re looking for AI tool reviewers specifically, our list of AI tool reviewers on YouTube worth following has already done the first half of this checklist: every creator on it uploads consistently and tests tools on screen. Before you approach a reviewer, read how to tell if a YouTube AI tool review is honest, because the qualities that make a review trustworthy are the ones that make it convert. And when you reach out, our methodology is a fair model for what creators expect: no strings, no paid placement, and disclosure from both sides.