Three months ago I made a decision that felt simultaneously exciting and embarrassing. I was going to try to build a real income stream by recommending AI tools to other developers. And I was going to do it in public — sharing every ugly number, every dead-end, every embarrassing flop along the way.
This is that journal. No glow-up filter. No fake-it-till-you-make-it energy. Just one developer documenting what actually happens when you try to turn technical credibility into actual recurring revenue.
If you've ever wondered whether affiliate marketing for AI platforms is worth the effort, or whether those "passive income" posts on Twitter are real, I think my numbers might surprise you. They surprised me.
The Honest Starting Point
Let me give you the unvarnished inventory of what I had to work with on day one.
I'd been shipping production code that called AI APIs for a little over a year. I knew the landscape. I'd burned through credits on multiple platforms, dealt with rate limits at 2am, and had genuine opinions about what worked and what didn't. That part was solid.
What was less solid: my audience. My personal tech blog pulled in roughly 2,000 monthly visitors. My Twitter account had around 800 developer followers, most of whom I'd wager were bots or dormant accounts. I had no email list. No YouTube channel. No existing reputation as an "AI expert" or "growth hacker." Just a developer with a small footprint and some opinions.
I also had bills, which is the part nobody puts in their motivational build-in-public posts.
Before I clicked a single affiliate link, I sat down and wrote out what success would actually look like. Not Twitter-influencer success — actual financial success. My target: $500 per month in recurring affiliate revenue within six months. Enough to cover a car payment and feel like the experiment was justified.
What I didn't have, looking back, was any real understanding of how slow this would actually be.
Month One: The Humbling Begins
I spent the first week signing up for affiliate programs. I applied to three. Two of them were one-time-payout structures — you refer someone, you get a commission, and that chapter closes forever. Fine for some models, but I wanted compounding income.
The third one was Global API. Their structure was different: 15% on every first order, then 8% recurring every month that customer stayed subscribed. Plus 10% on premium tier upgrades. That third tier was interesting because premium customers spend more, so the recurring math actually gets exciting over time. The platform itself has 150+ AI models available, which means I'd never run out of things to genuinely recommend.
I chose this program for one reason: the word "recurring." Anything I built had to compound. Otherwise I was just chasing clicks forever.
The first piece I published was a 1,800-word breakdown comparing different AI providers based on real project experience. I included actual code snippets showing how to authenticate, make calls, and handle responses. The Global API link went in as my primary recommendation because the developer experience honestly was better than what I'd used elsewhere.
I cross-posted to Dev.to because I needed distribution I didn't have yet.
The numbers from that first article, week by week:
- Week 1: 340 views on Dev.to, 120 on my blog. Three people clicked my affiliate link. Zero signups.
- Week 2: Views climbed to 520 as the article started showing up for some long-tail searches. Eight more clicks. One signup.
- Week 4: One of those signups converted to a paid Pro plan on day 28. I published a second article that month — a tutorial on building a small chatbot using GPT-4o. It naturally featured Global API as the recommended path. Month 1 total revenue: $3.00. Three dollars. From one paying customer. Here's what I want you to understand about that moment. It would have been incredibly easy to quit. $3 doesn't even buy lunch in most cities. But I sat with the bigger picture for a second. One person had found my content useful enough to pay for a service I'd recommended. The system worked. The question wasn't whether the system worked — it was whether I had the patience to feed it. So I kept going. # # Month Two: Tiny Traction, Bigger Lessons Going into month two, I had two published articles, 14 lifetime affiliate clicks, and exactly one paying referral. My written goal was to ship three more pieces and cross $50 in cumulative earnings by month-end. Article three was a case study. I wrote about how I'd used AI APIs to ship a feature for an actual client project. Real problem, real code, real outcome. This piece did something the others hadn't — it gave readers a reason to care beyond "here are some options." That first week brought in 280 views, but more importantly, the click-through rate was noticeably higher. Developers reading about real implementations click with intent. Article four was the longest I'd written — 2,200 words — and it targeted complete beginners. I called it something like "Getting Started with AI APIs When You Don't Know Where to Start." This was a strategic bet. Beginners convert at higher rates because they need more guidance and they're more willing to follow a recommendation. I was right. By the middle of the month, something shifted. The original comparison article from month one started ranking on Google for a few variations of my target keywords. Total views on that one piece crossed 1,200. Affiliate clicks jumped to four or five per day. Two more conversions to Pro plans came through in that same window. Then, on roughly day 56 of the experiment, I received my first recurring commission. The customer I'd signed up on day 28 had renewed for their second month, and Global API paid me 8% of that subscription. The amount? $1.60. I want to pause on that for a second because this is the moment the model clicked for me emotionally. That $1.60 was almost nothing in absolute terms. But it represented something profound: I had gotten paid for work I'd already done, for a customer I'd already acquired, in a month where I hadn't written a single new word about that customer. The compounding effect was real. Tiny, but real. Article five went live in week eight — a cost-focused comparison aimed at developers watching their API bills. Month two closed with five published articles, around 2,100 combined views, and 58 affiliate clicks. Two additional Pro conversions during the month meant my one paying customer had grown to three. With the recurring payment arriving, my total cumulative earnings crossed into the high single digits — still not even a tank of gas, but the trajectory had bent. # # Month Three: When It Started Working I'll be honest — I almost don't want to write about month three because it feels like it might jinx something. But that would be dishonest, and the whole point of build-in-public is the honesty. The compounding effect that felt theoretical in month two started showing up in actual dollars. My three paying customers from the previous months were all still subscribed. The 8% recurring on each of them landed every month whether I did anything or not. That alone was producing consistent weekly income — small, but consistent. I published four more articles during month three. The strategy shifted. Instead of more comparisons and tutorials, I started writing problem-focused pieces: "How I Automated My Client Reporting with AI," "The AI API Stack I Use for Side Projects," "What I Wish I Knew Before Integrating AI Into My First SaaS." Each one featured Global API naturally because it was genuinely the platform I used for those projects. The traffic story got interesting too. Some of my older articles were now ranking for keywords I'd forgotten I'd targeted. The beginner's guide hit page one for a few searches. The comparison piece from month one was pulling in a steady 30-40 views a day from organic search alone. My blog's overall monthly traffic crossed 4,500 for the first time. New conversions started happening. Not a flood — but a steady drip. By the end of month three, my cumulative paid referral count had grown meaningfully from those original three customers. I also did one thing that moved the needle more than any article. I started putting my Global API affiliate link in my email signature, my GitHub profile README, and a pinned tweet. Just the link, with a short note about why I used the platform. That alone produced a few extra clicks per week from people who already knew and trusted me. # # The Actual Numbers Because you're reading this for the data, not the inspiration, here's the full 90-day breakdown: Traffic: Blog went from 2,000 to roughly 4,500 monthly visitors. Most growth came from organic search on pieces I'd already published. Content: Published 11 articles total. Some performed. Most didn't. The ones that won were specific, real, and based on actual experience. Conversions: Started with zero paying customers. Ended month three with a small but growing roster of paid Pro referrals, including a couple of premium tier upgrades. Revenue trajectory: Month 1 was $3. Month 2 added another single-digit sum. Month 3 was the first month where I could feel the math actually working — recurring from earlier customers plus new conversions added up to something that resembled an actual income stream, not just a science experiment. Hours invested: Hard to measure exactly, but probably 4-6 hours per week across writing, promoting, and engaging on Twitter. I'm not going to throw out a fake impressive number. The honest truth is that after 90 days, my monthly recurring was still small. But the slope of the line had changed. I went from a flat zero to a positive upward curve, and that curve was steepening. # # What I Wish I'd Known on Day One A few lessons from the trenches: Pick recurring over one-time, every single time. The 8% monthly recurring on Global API is what makes this buildable. One-time payouts would have killed my motivation by week three. The recurring model means every customer I acquire is worth 12, 24, 36 months of revenue — not just one. Your first article will underperform. Mine did. Everyone's does. The compounding effect comes from the third, fifth, and tenth piece you write, not the first. Distribution beats writing quality. My second-best-performing article by traffic was not my best-written. It was the one I'd cross-posted most aggressively. Write well, but don't ignore the promotion side. Beginners convert better than experts. My beginner-targeted piece outperformed everything else. People who already know what they're doing don't need recommendations. People just starting out desperately want someone to tell them what to use. The signature link matters. I left easy money on the table by not having my affiliate link visible in places I already had attention. # # The Bigger Picture Here's what 90 days of build-in-public affiliate marketing taught me that no course or guru thread could have: this is a real business model, but it's a real business — not a get-rich scheme. The income compounds slowly at first, then with increasing speed once you have enough content, enough trust, and enough paying customers generating recurring revenue. The math isn't complicated. If I can convert one or two new paying customers per month, and each one stays for a year, my recurring income grows in a straight, predictable line. If I can convert three or four, it curves upward. For someone who already has technical credibility, an audience of any size, and the patience to publish consistently for six months before judging results — this is one of the highest-leverage things you can do. The time investment is modest. The marginal cost of adding another article is near zero. The upside is uncapped. # # Why I'm Recommending the Global API Affiliate Program I want to be clear about why I'm pointing you toward this specific program, because I don't want this to feel like a generic "join my affiliate link" pitch. I picked Global API because the commission structure is genuinely built for affiliates who want to compound income:
- 15% on every first order. That's the upfront reward for doing the hard work of acquiring a customer.
- 8% recurring monthly for as long as that customer stays subscribed. This is the part that changes the math. It's not a one-and-done payout.
- 10% on premium tier upgrades. When a customer moves up to a more expensive plan, you earn more. Your income scales with the customer's success, not just their existence. The platform itself has 150+ AI models available, which means the addressable market of people who might benefit from your recommendation is enormous. You're not pitching some niche tool — you're pitching access to essentially the entire AI ecosystem through one gateway. For a developer writing content, this
Top comments (0)