I run every side project like a paid acquisition experiment. If I can't measure it, I don't run it. That's the rule.
For the last three months, I've been running an affiliate funnel for an AI API platform, and I tracked everything. Every view, every click, every signup, every paid conversion, every recurring commission. UTM parameters on every link. Spreadsheets with cohort tables. The works.
This isn't a "passive income" story. It's a funnel optimization story. Below is the raw data, the LTV math that made me pick this program over two others, and the three A/B tests I ran that actually moved the conversion rate.
If you're a growth marketer, content creator, or developer with an audience, the underlying mechanics are worth stealing. Even if you couldn't care less about AI APIs.

Why I Picked a Recurring Commission Structure (The Math That Sold Me)

I had three options on the table. Two affiliate programs offered one-time payouts. One — Global API — offered 15% on the first order and 8% recurring on every monthly renewal. The platform also bumps commissions to 10% recurring for their premium tier, and they give affiliates access to 150+ AI models on a single endpoint, which makes promotion easier.
The math wasn't even close. Let me show you why.
If a customer signs up for a $30/month plan through my link, the one-time programs would have paid me somewhere between $6 and $12 once and then nothing. With Global API's recurring model, I get $2.40 every single month that customer stays. By month 6, I'd already match the one-time payout. By month 12, I'd be 4x ahead. By month 24, I'm at $57.60 from a single referral.
The takeaway: LTV is everything in affiliate marketing. Most affiliates chase the highest headline commission rate. Smart affiliates chase the highest projected commission rate over 12-24 months. That's the difference between a one-time bonus and a real income stream.

My Tracking Stack (So You Can Steal It)

Before publishing a single piece of content, I set up:

• UTM-tagged affiliate links for every article, every traffic source, every CTA placement
• Google Analytics 4 with custom events for affiliate clicks (not just pageviews)
• Airtable as my attribution layer, with one row per signup and a "source article" column
• A simple cohort table in Sheets tracking each referral's MRR month over month This cost me about three hours. It saved me from making decisions on vibes, which is how 90% of content creators blow their affiliate income. # # Month 1: The Cold Start Funnel My starting position: a tech blog with ~2,000 monthly visitors and a Twitter following of about 800 developers. Small but relevant. I'd been building with AI APIs for over a year, so I had credibility and actual usage data to draw from. I joined three affiliate programs, wrote two articles in the first month, and tracked every step of the funnel. Month 1 funnel totals:
• 2 articles published
• 750 combined views
• 14 affiliate clicks
• 2 signups
• 1 paid conversion (Pro plan, day 28)
• First-month earnings: $3.00 (one-time commission only; recurring starts month 2) Now let me break that down the way a growth marketer would. View → Click rate: 1.87%. That's a low click rate by content marketing standards, but normal for top-of-funnel educational content where readers are still gathering information, not buying. Click → Signup rate: 14.3%. This is the metric I cared about most. Of the people who clicked my affiliate link, 14% actually created an account. That's strong. Signup → Paid rate: 50%. Only one of the two signups converted to paid. Small sample, but a 50% activation rate is well above the typical 10-20% trial-to-paid benchmark in SaaS. Effective EPC (earnings per click): $0.21. That's $0.21 earned for every affiliate click I generated. Once recurring commissions kick in, this number compounds without me doing anything. The $3.00 didn't exactly fund a vacation. But the funnel was working. The CAC was effectively zero (I wasn't paying for traffic), and I had a paying customer generating monthly recurring revenue for me at $2.40/month going forward. # # The Funnel Decomposition That Changed My Strategy Here's where most affiliates make a mistake: they look at total revenue, not funnel performance by traffic source. I separated my Month 1 data by article: Article 1: A comparison of AI API providers based on real usage (1,800 words, cross-posted to Dev.to)
• 460 combined views
• 11 affiliate clicks (2.4% CTR)
• 1 signup
• 0 paid conversions in Month 1 Article 2: A tutorial on building a chatbot with the GPT-4o API (1,600 words, blog only)
• 290 combined views
• 3 affiliate clicks (1.0% CTR)
• 1 signup
• 1 paid conversion (the $3.00 one) The Dev.to cross-post dramatically outperformed my blog for top-of-funnel reach, but my blog converted better on the back end. The reason was obvious once I saw the data: my blog readers are warmer. They subscribed to my newsletter, they know my voice, and they trust my recommendations. This was my first A/B test hypothesis: Future content should cross-post to Dev.to for top-of-funnel discovery, but include a soft CTA to subscribe to my blog for the conversion moment. # # Month 2: When the Recurring Model Started Compounding Month 2 was where things got interesting. Not because the traffic exploded — it didn't — but because I had three growth levers pulling at once:
• New content bringing in fresh clicks
• Old content from Month 1 still ranking and generating passive clicks
• Recurring commissions from my Month 1 referral Month 2 activity:
• Published 3 new articles (5 total by end of month)
• 2,100 combined views across all articles
• 58 affiliate clicks (a 4x increase from Month 1)
• 2 more paid conversions (Pro plans)
• First recurring commission: $1.60 from my original referral's second month Let me run the numbers again: View → Click rate: 2.76% (up from 1.87%) Click → Signup rate: trending higher EPC: ~$0.30+ and rising The reason the click rate went up wasn't luck. I made three deliberate changes:
• I moved the affiliate CTA from the end of articles to the second paragraph. People who read to the end of a 2,000-word article are the minority. People who read the second paragraph are 100% of readers.
• I added contextual affiliate links inside code blocks, not just in the conclusion. Developers copy code. If my code example uses the platform I'm promoting, the link is right there.
• I started writing for beginners, not experts. My Month 1 articles were written for developers who already knew what an API was. My Month 2 beginner-focused article pulled in readers who needed more hand-holding, and they converted at a higher rate because they were actively looking for a recommendation. # # The A/B Tests That Actually Moved the Needle I ran a few small experiments in Month 2. Most of them were inconclusive (sample size too small to be statistically significant), but two were clear winners: Test 1: CTA placement (end of article vs. mid-article).
• End-of-article CTA: 1.8% click rate
• Mid-article CTA (after the value proposition was established): 3.4% click rate
• Winner: Mid-article, by 88%. Test 2: "Sign up" vs. "Try free" as link anchor text.
• "Sign up for Global API": 2.1% click rate
• "Try Global API free": 3.8% click rate
• Winner: "Try free" by 81%. Lower friction language wins, always. These aren't groundbreaking tests, but they demonstrate the principle: never assume. Test. A 2x improvement on a small conversion rate compounds into massive revenue over 12 months. # # The Cohort Table: Where Affiliate Marketing Becomes a Business The most important thing I built wasn't a landing page or a content calendar. It was a cohort table. For every signup I drove, I tracked:
• Signup date
• Plan (Pro, Premium, etc.)
• First-order commission
• Recurring commission per month
• Status (active, churned, upgraded) After two months, I had a small but real dataset showing that my Month 1 referral was still active, still paying, still earning me 8% recurring. If that customer stays for 12 months, I'll earn roughly $28.80 from a single conversion that cost me maybe four hours of writing. Multiply that by 20 active referrals, and I'm looking at $576/month in passive income from a single piece of evergreen content. That's not a side hustle anymore. That's a business. The 10% premium tier commission makes this even more attractive. If I can guide higher-value customers toward the premium plan, my LTV per referral jumps by 25%. # # What's Next: Month 3 and Beyond Heading into Month 3, I have a clear optimization roadmap:
• Double down on beginner content — highest conversion rate
• Refresh my Month 1 comparison article with 2026 data (a 30