How Crossway Influence vets creators.
Subscriber counts don’t decide anything. Every recommended creator passes these checks, or the shortlist says which ones they miss and why they’re still worth booking.
Last reviewed: 21 October 2026
Vetting
YouTube · AI coding channel
One pick from the sample shortlist, checked against this page.
- Median views, last 10 videosabout 45,000Pass
- Comments per 1,000 views2.4Pass
- Likes per 1,000 views33Pass
- Latest long videoOctober 2026Pass
- Who’s commentingAI coding tool usersPass
- Past sponsored posts3 watchedPass
- Growth patternNo unexplained jumpsPass
- Price per 1,000 views$40Pass
Book. Every check passed.
YouTube
| Check | Threshold | Why |
|---|---|---|
| Median views over the last 10 long videos (not Shorts) | 2,000+ for a niche infrastructure channel, 5,000+ for general developer topics | An average hides one viral video. The median shows a normal week. |
| Comments per view | 2+ per 1,000 views | Practitioners ask questions. Very few comments suggests a passive audience; combined with sudden jumps in views, bought ones. |
| Likes per view | 20+ per 1,000 views | A second sign that real people are watching. |
| Latest long video | A long video in the last 90 days | A channel that has gone quiet isn’t booked until it posts again. |
| Who’s commenting | Engineers discussing code, errors and setups | “Great video” comments don’t tell you who’s watching. Technical questions do. |
| Past sponsored posts | The last three, watched in full | Does the creator explain the product, or read a script nobody hears? |
| Growth pattern | No unexplained jumps in subscribers or views | Sudden jumps can mean bought growth. |
| Price per 1,000 median views | $20–55 per 1,000 views: the performance-based rate HyperGrowth Partners tested with its portfolio companies. | Above that range, only an exact audience match justifies a booking, and the shortlist says so. |
PostHog publishes its own bar: 5,000+ views per video, and it rates 2–5 comments and 30–50 likes per 1,000 views as good. For smaller, targeted campaigns, smaller channels qualify when the comment section shows exactly the right audience.
Why the median, not the average
One viral video can make a channel look far bigger than it is. The median shows what a normal video gets, which is what a sponsored video will most likely get too.
Newsletters
| Check | Threshold |
|---|---|
| Open rate | 30%+ as published. If not published: past sponsors and clicks per sponsor, asked for every time. |
| Reader location | Matches where your customers are |
| Past sponsors | Developer tools you’d be proud to sit next to |
| Slot | Top slot or secondary, priced accordingly |
Open rates are reported by the publisher and inflated by Apple Mail, so reports count clicks on your own link.
| Check | Threshold |
|---|---|
| Median comments on the last 10 posts | 20+ |
| Who comments | Engineers and engineering leads, not recruiters or the same small group within minutes of posting |
| What they post | Technical content they’d write anyway, not engagement bait |
Who Crossway Influence doesn’t book
The product categories in the terms:
- Gambling (including prediction markets)
- Crypto of any kind (tokens, trading, wallets, blockchain infrastructure)
- Trading, investing and forex apps
- Interest-based lending and buy-now-pay-later
- Alcohol, tobacco and cannabis
- Adult products (including NSFW image generators and romantic AI companions)
- Astrology
- Channels whose recent sponsors include any of these.
And these creators:
- Channels mainly about getting hired (interview prep, job hunting): their audience is looking for work, not tools.
- “Get rich with AI” channels and anything built on hype rather than how-to.
- Channels that won’t share their real analytics when asked.
Before a sponsored post goes live
Draft check
For every sponsored post whose draft the creator shares (the shortlist notes this once each creator confirms):
- Technical accuracy, checked against your docs.
- A clear sponsored label at the start.
- Every link and code tested.
- No claims your team hasn’t approved.
- A copy archived, so there’s a record of exactly what ran.
Sponsored posts without a shared draft get the same checks on the day they go live.