Built immi’s in-house paid influencer program, testing Instagram and YouTube, negotiating lower rates, and using audience research and campaign results to guide investment.
I owned program structure, audience research, sourcing, outreach, negotiations, contracts, creative briefs, performance measurement, makegoods, and rebooking. A virtual assistant supported sourcing, sampling, contract administration, and reporting.
Agency YouTube: December 2021–September 2022 · Internal YouTube: July–December 2023 · Internal Instagram: July–August 2023
YouTube · agency-managed: December 2021–September 2022 (10 months). YouTube · in-house: July–December 2023 (6 months). Instagram Stories · in-house: July–August 2023 (2 months).
The headline revenue and view totals combine all three programs.
Meta was our main acquisition channel, and we wanted to diversify our channel mix at a similar CPA. Earlier at immi, I managed a YouTube influencer marketing agency but paused the program after 10 months because Meta’s CPA was considerably lower.
Several months later, we reformulated the product and received positive early reviews. Given the new product, I proposed restarting paid influencer marketing internally so we could scale it profitably without agency fees.
I used post-purchase surveys to understand how customers discovered immi and what content they enjoyed. I also studied the niches adjacent brands sponsored. Small creator tests helped us evaluate click-through and order rates before pursuing larger creators in promising categories.
I evaluated historical views, audience demographics, existing sponsors, and posting consistency, then used a calculator to estimate a viable fee. I successfully negotiated down rates for 18 of the 29 influencers I worked with in-house, saving 22% off our cost.
Price wasn’t the only lever: I also negotiated pinned comments, extra placements, paid usage rights, and makegoods. I rebooked influencers who met my ROAS threshold. For those who came close, I negotiated another integration at a lower rate to improve the economics of the partnership.
The brief gave creators hook options, product context, personal-experience prompts, and a clear call to action. It encouraged them to use their own voice and share what actually resonated.
I combined discount-code tracking with post-purchase surveys. YouTube sales were updated weekly, and Instagram Stories were evaluated during their 24-hour live window.
At times, I noticed a burst of code uses well after content went live, or Instagram Story codes being used after the 24-hour window. I first checked for a spike in video views, a code in the creator’s bio, or a follow-up Story. I then checked coupon sites, where codes were often leaked.
A virtual assistant monitored coupon sites daily and requested removals to keep attribution as accurate as possible, since it heavily informed long-term partnerships. These checks reduced contamination but could not eliminate it; CPA measured cost per attributed order.
Nerd culture, Asian comedy, and food content initially seemed like strong fits, but often performed far below our ROAS target. A rising camping/van-life creator was a particularly expensive miss: roughly $15K spent for about $2K in attributed revenue. Surprisingly, some of our highest-ROI partnerships came from wine and beer review channels, political commentary, and scientific food channels.
I paused Instagram after two months because its returns were weaker than YouTube’s, allowing me to focus on what was working. Most YouTube purchases occurred in the first two days after posting, then tapered. Evergreen videos sometimes brought additional orders over the following weeks and months.
At three months, CPA was close to Meta’s, so we extended the test and invested further.
Lower rates didn’t guarantee better acquisition economics. I had to evaluate reach, click-through, orders, and audience fit together, rather than treating discounted placements as the main win.
The program gave us a repeatable operating process and clearer evidence about which audiences responded. When overall efficiency became less attractive, we paused paid influencer investment, redirected budget to Meta, and later tested TikTok Shop as a new channel.
A simplified interactive version of my planning calculator. Inputs are illustrative assumptions, not actual campaign results.