Case studies US natural wellness subscription brand, 3 min read
Eight Years of Influencer Outreach, 186,831 Emails: The Leak Was After the Yes
Role. Took over the brand's ambassador program in October 2026. Before changing anything, I audited every outreach email the team had sent since 2018.
- 186,831
- emails analysed, back to 2018
- 21%
- of creator yeses that became posts
- 47 of 97
- sampled yes threads where the team stopped replying
Context
The brand sends free product to creators in exchange for posts, and had run this program for eight years through an outreach tool. Placements had fallen sharply even as the team sent more email. When I took the program over, the obvious assumption was that the pitches had stopped working.
I wanted to know where people actually dropped out before rewriting anything.
The audit
I pulled the full outreach history (29,538 websites, 186,831 emails) and matched each creator's first pitch to what they replied and whether they ever posted. Then I read 97 real "yes" threads from 2023 to 2026 one by one, and a sample of 115 replies asking to be paid.
Finding 1: the pitch wasn't the problem
32% of creators replied, which is healthy for cold outreach. The drop happened after they said yes: only 21% of yeses became posts.
In 47 of the 97 yes threads I read, we simply stopped replying. Some creators sent their shipping address and never got a confirmation. Some chased us twice. Since 2023, about 16% of yeses never received another email of any kind, up from 5% before.
When the team did reply, speed wasn't the issue (the median was one day in both won and lost threads). Coverage was. The threads that converted got a "you're in, your kit is shipping" email and three check-ins after delivery. The lost ones got zero or one.
Finding 2: volume made it worse
The team roughly doubled outreach volume, and the placement rate collapsed. More pitches meant more yeses than anyone could follow up on.
Finding 3: a third of replies ask to be paid
35% of all replies asked about payment, and 57% of those never got an answer. The ones that converted were handled by accepting a smaller deliverable (a set of stories instead of a reel), offering a higher commission, or naming the commission upfront. Talent agencies sent 24 of the 115 paid asks and none became a post.
Finding 4: what good pitches had in common
Within the same year and the same sender, some subject lines placed at three times the rate of others. The winners named a specific post, made the ask early, and made replying easy. Pitches that offered a free kit without saying what we wanted in return got replies but almost no posts.
Follow-ups worked up to about the third touch. After that, each extra email mostly added "not interested" replies.
The fix plan
- A follow-up sequence after every yes: confirmation, shipping, then check-ins at weeks 3, 5 and 7
- A pitch that answers the three questions creators asked most (paid or gifted, what to post, what's in the box)
- A clear rule for paid asks, and a stop list: agency-managed talent without a budget, shops, off-niche accounts, and anyone outside the US and Canada
- List hygiene: no re-pitching after a clear no, and no new campaigns dropped into old threads
- About 4,300 past contacts flagged as ready to re-pitch
Where it is now
The fixes are only starting, so there are no results to report yet. I'm sharing this one for the method: pulling the whole history, reading real threads rather than trusting averages, and finding that the cheapest fix was replying to people who had already said yes.
Takeaway
- Find the leak before fixing the pitch. Here the funnel broke after the yes, not before it.
- Sending more can lower results when follow-up can't keep up.
- Averages hide behaviour. Reading 97 threads explained what the totals couldn't.