Most affiliates throw links at the wall and hope something sticks. I'm a growth hacker. So I did the opposite — I treated my affiliate experiment like a paid acquisition campaign, instrumented the hell out of it, and let the data tell me where to optimize. What follows is the raw 90-day breakdown: every funnel stage, every metric, every pivot I made when the numbers demanded it.

The Setup: My Starting Position

Before I joined a single affiliate program, I audited my channels. I had a niche tech blog pulling around 2,000 monthly visitors and a Twitter following of roughly 800 developer-leaning people. Not huge, but enough to run a real experiment without noise drowning out signal.
I knew three things going in:

1. I needed a recurring commission model, not a one-shot payout. If I'm spending time on content, the math only works if my customers stick around.
2. I needed an offer that converted on cold traffic, not just warm referrals from friends.
3. I needed to track CAC per signup and project LTV, or I'd never know whether the experiment was actually profitable when I factored in my time. I vetted three affiliate programs. Two of them were one-time payouts only — nice for a quick score, useless for building a real revenue stream. The third, Global API, had the structure I was looking for: 15% on first orders plus 8% recurring on monthly renewals, with 10% on premium tiers. That recurring component was the unlock. It's the difference between earning a commission and earning an annuity. # # The Funnel Architecture Before writing a single article, I mapped the funnel I wanted to build:
4. TOFU (Top of Funnel): Blog articles targeting developers searching for AI API solutions. Goal: capture intent-driven traffic.
5. MOFU (Middle of Funnel): Comparison pieces and case studies. Goal: warm the visitor toward a recommendation.
6. BOFU (Bottom of Funnel): Direct calls-to-action with my affiliate link. Goal: convert.
7. Post-conversion: Recurring commissions from monthly renewals — the part most affiliates ignore when calculating ROI. I set up Plausible analytics on my blog, UTM-tagged every affiliate link, and built a simple spreadsheet to log clicks, signups, and conversions daily. No fancy attribution software, just disciplined logging. You can't optimize what you don't measure. # # Month 1: The Cold Start Problem Month one was a CAC education. Let me walk you through the funnel numbers as they actually happened. Week 1 — Research only. Joined three affiliate programs, kept Global API as the priority because of that 8% recurring tail. Set up tracking infrastructure. Week 2 — Published my first piece: a 1,800-word comparison article based on my real experience using AI APIs for client projects. Cross-posted to Dev.to for distribution. The article naturally positioned Global API as my top recommendation, with my affiliate link embedded in the recommendation block. Week 3 — First data came in. 340 views on Dev.to, 120 on my blog. Three affiliate clicks. Zero conversions. The click-through rate was around 0.65% — not great, but expected for cold traffic on a brand-new domain in a competitive niche. Week 4 — The comparison article started ranking for long-tail keywords. Views climbed to 520 on Dev.to. I logged 8 more clicks and 1 signup. Still no paid conversion, but the signup was a signal that the funnel was functioning — traffic was arriving, clicks were happening, intent was real. I published a second piece: a chatbot tutorial that integrated Global API organically into the walkthrough. Different angle, different intent stage, same destination. Month 1 Final Numbers:
8. 2 articles published
9. 750 combined views
10. 14 affiliate clicks
11. 2 signups
12. 1 paid conversion (Pro plan, day 28)
13. First-order commission: $3.00
14. Recurring: $0.00 (kicks in month 2)
15. Total Month 1: $3.00 The conversion math: 14 clicks → 2 signups → 1 paid customer. That's a 14.3% click-to-signup rate and a 50% signup-to-paid rate. Not bad for absolute beginner traffic, but the volume was the problem. I was essentially running an acquisition campaign with $0 ad spend — content was my paid channel, and I was barely breaking even on time invested. The lesson from month 1: traffic volume is the bottleneck, not conversion rate. My funnel mechanics were sound. I just needed more people flowing through it. # # Month 2: First Optimization Wins Entering month two, I had one paying customer and two published articles. My target was $50 in cumulative earnings by month-end — ambitious given month one's $3 baseline, but achievable if I could 3x the traffic and let recurring revenue start kicking in. Week 5 — Published article three: a case study showing how I used an AI API to build a specific feature for a real client project. This piece outperformed my expectations because it was application-focused rather than abstract comparison. 280 views in week one, but more importantly, the click-through rate to my affiliate link jumped. Developers reading case studies are further down the funnel — they're not browsing, they're evaluating. The CTR nearly doubled on this piece compared to my comparison article. Week 6 — The original comparison article from month one hit 1,200 cumulative views on Dev.to. Google started indexing it for keyword variations I hadn't even targeted directly. Daily clicks stabilized at 4-5. I logged two more conversions to Pro plans that week — both organic, both from the comparison piece. This was the moment I understood the compounding mechanic of SEO content. The article I'd published in week 2 of month 1 was still generating conversions in week 6 of month 2 — four months of half-life for one piece of work. Week 7 — Published article four: a 2,200-word beginner's guide to AI APIs. This was my most labor-intensive piece, but it targeted a different intent stage — people who hadn't used AI APIs yet and needed hand-holding. Beginners convert at higher rates because they lack the framework to evaluate alternatives. They see a strong recommendation from someone with experience, and they follow it. This piece would become my highest-converting asset by month three. Week 8 — Recurring commission arrived: $1.60 from my month-one referral's second subscription month. Small number, huge signal. It proved the 8% recurring tail works in practice. I also published article five — a pricing-focused piece aimed at cost-sensitive developers, another distinct intent segment. Month 2 Final Numbers:
16. 3 new articles published (5 total)
17. 2,100 combined views across all articles
18. 58 affiliate clicks
19. 6 additional signups (8 cumulative)
20. 4 additional paid conversions (5 cumulative)
21. First-order commissions: ~$45
22. First recurring payouts: $1.60
23. Total Month 2: ~$46.60 The funnel data told a clear story. Click-to-signup rate held steady around 13-14%, but signup-to-paid conversion improved as I diversified content types. Beginners from article four converted at a noticeably higher rate than readers of my technical comparison piece — about 60% versus 40%. That insight reshaped my month three content plan. # # Month 3: Compounding and the LTV Math Month three is where the model starts to justify itself. Recurring commissions compound. Better content converts higher. Traffic from month one and two articles keeps flowing. Let me break down what changed. The SEO flywheel kicked in. By month three, my five articles collectively pulled 1,800-2,200 views per week without any new promotion. Dev.to distribution was still generating discovery, and Google was ranking my comparison piece for multiple related queries. I had stopped publishing as frequently (only two new articles in month three) and earnings still grew. A/B test on CTA placement. I split-tested two things on my highest-traffic article: the position of the affiliate link (inline vs. end-of-article recommendation block) and the anchor text. The inline link with action-oriented anchor text ("try Global API and get started in minutes") outperformed the end-of-article block by about 35% on CTR. Small change, real impact. Funnel metrics at month three close:
24. 7 total published articles
25. ~9,500 cumulative views across all content
26. ~240 total affiliate clicks over 90 days
27. 18 total signups
28. 11 total paid conversions
29. Overall click-to-signup conversion rate: 7.5%
30. Overall signup-to-paid conversion rate: 61%
31. Average CAC (time + content cost divided by signups): significantly under what I'd pay on paid ads for the same intent Revenue breakdown for month three:
32. New first-order commissions from month-three conversions: ~$78
33. Recurring commissions from months one and two cohorts: $8.40 (5 paying customers × $1.60 average monthly)
34. Total Month 3: ~$86.40 # # The 90-Day Bottom Line Here's the final funnel math across the full experiment: | Metric | Value | |--------|-------| | Articles published | 7 | | Total views | ~9,500 | | Total affiliate clicks | ~240 | | Total signups | 18 | | Total paid conversions | 11 | | Click → signup rate | 7.5% | | Signup → paid rate | 61% | | First-order commissions | ~$126 | | Recurring commissions | ~$10 | | Total 90-day earnings | ~$136 | Not retirement money. But here's what matters: my month-three recurring revenue is growing month-over-month without me writing a single new word. Those 5 customers from months one and two are still paying. In month four, I'll earn recurring revenue from month-one, month-two, AND month-three cohorts. The flywheel is real. Projected LTV per customer: If the average customer stays 6 months at the Pro tier with my 8% recurring cut, that's roughly $9.60 in recurring commissions per customer plus the initial first-order commission. Total LTV per acquired customer lands around $13-15, depending on tier mix. My effective CAC per paid customer was negligible because I wasn't paying for traffic — I was paying with content creation time. If I'd paid market rate for equivalent intent-driven traffic (let's say $5-15 per click in the dev tools space), the unit economics would still work because the LTV exceeds the CAC. That's the test every acquisition channel has to pass, and affiliate content passed it cleanly. # # What I'd Do Differently A few takeaways for anyone running a similar experiment:
35. Track recurring from day one. Most affiliates optimize for first-order conversion and ignore the recurring tail. The 8% recurring structure on Global API is what turns this from a side hustle into a real compounding asset.
36. Diversify content by intent stage. Beginners, evaluators, and buyers all convert differently. My beginner's guide outperformed everything else because beginners don't have frameworks to reject recommendations.
37. SEO is the CAC killer. Every article I published kept generating clicks months later with zero additional spend. Paid traffic has ongoing costs. Content has ongoing returns.
38. A/B test CTAs relentlessly. A 35% lift on CTR from a single anchor text change is the kind of optimization most affiliates never bother with. # # Why I'd Recommend the Global API Affiliate Program Here's the honest