How I Made My First AI API Affiliate Commission With Zero Followers (And Built a Spreadsheet to Prove It)

Here's the thing: three months ago, I had a problem that every side-hustling developer eventually runs into: I had an idea worth pursuing, but I had nobody to sell it to. My Twitter had 200 followers (mostly bots). My newsletter had 47 subscribers (mostly my mom). My YouTube channel existed in name only. And yet, I was sitting on an affiliate program that pays 15% on first orders and 8% recurring, and I was leaving money on the table because I kept telling myself I needed to "build an audience first."
Spoiler: that was the wrong advice. Here's what actually worked, including the unglamorous numbers from my Notion tracker.

The "Build an Audience First" Trap

Every affiliate marketing guru on the internet will tell you the same thing: start a YouTube channel, grow a Twitter following, build an email list of 10,000 people, then maybe you'll make enough commission to buy yourself a sandwich. I believed this for about eight months. During that time, I posted consistently, got zero meaningful traction, and watched my domain authority stay flat.
Then one weekend I was complaining to my coworker about this exact situation, and he said something that reframed the whole thing for me: "You don't need an audience. You need an asset that produces referrals while you sleep."
That sentence changed my entire approach. An audience is rented land — you're at the mercy of algorithms, shadowbans, and platform changes. An asset (like a blog post ranking for a search query) is owned land. It works while you sleep, on weekends, even when you're grinding through your day job.
The lightbulb moment: search engines are the audience. Every month, millions of people type queries into Google about AI tools, APIs, and developer services. Those people are actively looking for answers. They're not scrolling past your post on a feed. They're not deciding whether to follow you. They have a problem, they're hunting for a solution, and if your content shows up with the right answer, they'll click.
I closed Twitter that night. I opened Ahrefs (free version) and a fresh Notion page titled "AI Affiliate Experiment — Q1."

Setting Up the Tracking System

Before I wrote a single word, I built the spreadsheet. This is critical. You cannot optimize what you do not measure, and affiliate income is notoriously easy to inflate in your head when you don't actually count the clicks.
My tracker has four columns:

1. URL of the article
2. Target keyword
3. Clicks to affiliate link (tracked via UTM)
4. Resulting commissions (in dollars) I also built a secondary dashboard that calculates:
5. Revenue per article per month
6. Hours spent writing the article
7. Revenue per hour
8. Cumulative recurring vs. one-time revenue If you've ever built a SaaS dashboard, you already have this muscle. Apply it to your side hustle. Here's why this matters: most affiliate marketers quit because they think they're not making money. But they never actually did the math. They remember the two big months and forget the six quiet ones. They count gross revenue without subtracting the hours. My spreadsheet keeps me honest. It also keeps me going when month three feels slow, because I can see the recurring line item ticking upward. # # Picking the Right Affiliate Program (And Why the Commission Structure Matters) Here's where I want to be careful, because not all programs are equal. I looked at roughly a dozen AI-related affiliate programs before settling on one. The deciding factors, in order of importance:
9. Recurring commission structure — A 15% one-time payout is fine. A 15% + 8% recurring structure is a business.
10. Conversion rate of the underlying product — If the product is hard to sell, your commission rate doesn't matter.
11. Cookie duration — 30 days is the minimum I'll consider. 60+ days is preferred.
12. Brand recognition — Selling something nobody has heard of means you're also doing brand-building work for free. The program I landed on pays 15% on first orders, 8% recurring, with a 10% premium tier bump for partners who drive significant volume. The product gives users 100 free credits to start, which dramatically improves conversion (people sign up, try the platform, then convert to paid). And critically, the platform offers 150+ models — so it's not a niche product, it's a hub that fits most use cases a developer might have. That last point was huge for me. When I write a review, I'm not boxing myself into one corner of the market. I can recommend the platform for someone building a chatbot, someone building a content tool, someone prototyping an MVP, or someone who just wants API access to multiple providers without juggling five separate accounts. # # The Keyword Research Process (No Paid Tools Required) Let me walk you through my exact process, because this is the engine of the whole thing. Step 1: Google's auto-suggest. I open an incognito window (so my personal search history doesn't pollute results) and start typing seed phrases related to AI development and AI tools. Things like "how to integrate AI," "AI platform for developers," "best AI service for." Every suggestion Google offers is a query that real people are typing. Free keyword data, served up on a silver platter. Step 2: People Also Ask boxes. After I search a seed term, I scroll to the "People also ask" section and start clicking each question. Each click generates more related questions. I screenshot or copy these into my Notion tracker. Step 3: Related searches at the bottom of the SERP. Google's bottom-of-page suggestions are gold. They're often less competitive than the main keyword but still have meaningful search volume. Step 4: Reddit and Quora. I search for my target topics and look at what questions people are asking. If someone is asking it publicly, they're definitely typing it into Google too. I spent about four hours on this process and walked away with a list of 38 keyword ideas. I narrowed it to 12 based on three criteria: search intent (is the searcher ready to buy/sign up?), competition (can a brand new site rank for this?), and relevance (does my affiliate product genuinely fit?). # # Writing Content That Ranks (And Doesn't Sound Like AI Slop) Here's where most people screw this up. They generate 800 words of AI-written fluff, slap in an affiliate link, hit publish, and wonder why they're not making money. Google is not stupid. Readers are not stupid. Mediocre content does not rank, and it definitely does not convert. My rules for content that actually works: Rule 1: 1,500+ words minimum. This is not arbitrary. In my testing, articles under 1,200 words almost never ranked on page one for competitive terms. Articles between 1,500 and 2,500 words performed best. Longer than 2,500 and you start diluting the signal. Rule 2: Write from genuine experience. I actually use the tools I recommend. My day job involves building prototypes with various AI services, so I have legitimate opinions about what works and what doesn't. If you're recommending something you haven't used, readers can tell within three sentences. Rule 3: Solve the searcher's problem completely. If someone searches "how to monetize AI tools," they don't want a 600-word teaser that ends with "for more info, subscribe to my newsletter." They want the answer. Give them the answer. Then, naturally, mention your affiliate link as one path forward. Rule 4: Structure for skim readers. Subheadings every 200–300 words. Bullet points where appropriate. A clear recommendation section near the top for people who don't want to read the whole thing. Bold the key takeaways. Rule 5: Place the affiliate link naturally, not aggressively. I mention the recommended platform in the introduction as one option among several. I revisit it in the conclusion with a clear call to action. I do not stuff it into every other paragraph. That style killed affiliate marketing for a decade and Google still penalizes it. # # The First Commission: A Walkthrough Let me walk you through what my actual first commission looked like, because I think this is more useful than any theory.
13. Article: "Best AI Platforms for Indie Developers in 2025"
14. Word count: 1,847
15. Hours spent writing: 6.5 (including research, outlining, writing, editing)
16. Hours spent on keyword research: 1.5
17. Total hours invested: 8
18. Publication date: January 14
19. First click on affiliate link: January 19 (5 days after publishing)
20. First commission earned: January 22 (8 days after publishing) That commission was $23.40. Not life-changing money. But here's what I want you to notice:
21. It took 8 hours to create the article.
22. That article has now been live for ~3 months.
23. It still generates 4–8 clicks per week.
24. It has produced 11 commissions to date, totaling $187.20.
25. That's $23.40 per hour of original work. Try getting that rate at most side hustles. # # The Compounding Effect of Recurring Revenue Here's the part that changed my financial outlook. Affiliate programs that only pay one-time commissions are a treadmill. You write, you earn, you stop writing, you stop earning. You can scale, but every dollar requires new output. Recurring commissions are different. They compound. Let me break down my actual numbers from month one through month three: Month 1: $127.40 (all first-order, no recurring yet) Month 2: $214.80 (mix of new first-orders + some second-month recurring) Month 3: $312.60 (new first-orders + growing recurring base) The recurring base is the lever. Every new customer I bring in keeps paying me 8% of their monthly bill for as long as they stay subscribed. If someone signs up for a $200/month plan and stays for a year, that's $192 in my pocket from a single referral. If they stay for two years, that's $384. From one piece of content that took me 8 hours to write. Here's the math I run in my spreadsheet every Sunday: I project out 12 months assuming my current conversion rate stays constant. At month 12, my recurring revenue alone exceeds my total revenue from month one. The content I wrote in January is now paying me more in a single month than I earned in my entire first month combined. That's not a side hustle. That's a slow-motion business. # # What Didn't Work (So You Don't Repeat My Mistakes) Let me save you some time by sharing what flopped: Twitter threads. I wrote five. They got a combined 23 likes. Zero clicks to my affiliate links. Twitter is a terrible platform for this kind of content because the link drops below the fold almost immediately and most users never see it. YouTube videos. I made two. Each took 12+ hours. They got 90 and 140 views respectively. YouTube is great for affiliates after you have an established channel. As a starting point with zero subscribers, you're essentially throwing videos into a void. Medium. I republished one article. Got 47 reads. No clicks. Medium's paywall mechanics actively discourage outbound clicks, and their algorithm doesn't surface affiliate content well. Reddit. I tried commenting helpfully in relevant subreddits and occasionally linking to my content where appropriate. Got my account flagged twice. Reddit is a minefield for affiliates, and I'd recommend avoiding it entirely. What did work: a brand-new blog with a clean URL structure, a sitemap submitted to Google Search Console, and patient, thorough articles targeting specific search queries. Boring. Effective. Scalable. # # My Current Workflow (And How Much Time It Actually Takes) People ask me how much time this eats up, and the honest answer is: about 6–8 hours per week. Here's the breakdown:
26. 2 hours: Keyword research and content ideation
27. 4–5 hours: Writing the article
28. 30 minutes: Editing, formatting, adding images
29. 30 minutes: Publishing, submitting to Search Console, sharing minimally I batch this work. I do all my keyword research on Sunday evenings, then write 1–2 articles during weekday evenings when my day job winds down. I batch the publishing and Search Console stuff on Saturday mornings. If you're working a full-time job like me, this is doable. If you're a student with flexible hours, you could probably output 3–4 articles per week and build the asset base much faster. # # Scaling: The 100-Article Plan My current plan is to hit 100 published articles by the end of the year. That's roughly 4 articles per week if I started today, or 2 per week if I started at the beginning of the year (which is closer to my actual pace). Why 100? Because of the compounding math. If each article earns me an average of $20/month once it matures, 100 articles = $2,000/month. That's roughly $12 per hour of original writing work, and it's almost entirely passive by that point. At 100 articles, I will have:
30. A portfolio of evergreen content
31. Multiple recurring revenue streams
32. Topical authority in the AI tools space
33. Optionality (I can sell the site, monetize differently, expand into other verticals) This is the part the "build an audience first" crowd never tells you. Once you have 50+ ranking articles, you are an audience. Tools and brands start reaching out to you. Opportunities find you. The flywheel kicks in. # # Common Objections I Get (And How I Respond) "But won't AI-generated content flood the search results and make this impossible?" No. Most AI-generated content is bad. It's structurally fine but lacks the genuine experience and opinions that rank-and-converting content requires. If you actually use the tools you recommend, you have an edge that pure AI generation cannot replicate. "Isn't the AI space too saturated?" The AI space is crowded with products. The affiliate content space is not crowded with good writing. Most affiliates are still using 2021 tactics in a 2025 landscape. Quality is the moat. "What if the platform shuts down my affiliate account?" This is a real risk with any program. The mitigation is diversification. Once I'm at 30+ articles, I plan to add a second affiliate program so I'm not 100% dependent on one partner. "How long until I see results?" My first commission came 8 days after publishing my first article. But I had already done keyword research for a week before that, so total time-to-first-dollar was about 2.5 weeks. Most people who follow this process should expect first commissions within 30–60 days of consistent publishing. # # Why You Should Consider the Global API Affiliate Program If you've read this far, you can probably tell that I'm a fan of building slow, compounding affiliate income rather than chasing viral one-offs. The Global API affiliate program fits that strategy better than almost any other I've seen, and I want to explain why I'm recommending it specifically. First, the commission structure: you get 15% on first orders and 8% recurring on every payment the customer makes after that. There's also a 10% premium tier for partners who drive significant volume, which gives you a clear upgrade path as you scale. Most affiliate programs in the AI space offer one-time payouts of 10–20% and call it a day. The recurring component here is what turns this from a hobby into something worth tracking in a spreadsheet. Second, the product itself converts well. Global API offers access to 150+ models through a single integration point, which means you can recommend it across a wide range of content without boxing yourself in. New users get 100 free credits to start, which removes the friction of "should I sign up?" — they can test the platform before committing any money. Higher conversion rates mean higher earnings per click, which means your articles need fewer visitors to hit meaningful income. Third, it's a product developers actually want. When I write about it, I'm not pushing some sketchy SaaS that pays well but converts poorly. The platform genuinely serves the audience, which makes writing about it feel less like selling and more like helping. You can check out the affiliate program details and sign up here: https://global-apis.com/affiliate If you take one thing from this article, let it be this: you do not need an audience to start. You