I almost didn't write this post.
Not because I had nothing to say, but because the numbers felt embarrassing. My first month promoting AI APIs as an affiliate, I made a grand total of $0. The second month? $47.83. The third month? $312.40.
But here's the thing — I'm a build-in-public person now. I share my monthly income reports because transparency is the whole point. So I'm going to walk you through exactly how I went from zero audience, zero email list, and zero clue to pulling in recurring commissions from an AI API platform. No filter, no cherry-picking, no guru nonsense.
If you're wondering whether affiliate marketing for AI APIs works when you have no audience, this is the diary you're looking for.
Why I'm Telling This Story Publicly
A quick bit of context, because it matters for the rest of the post. I've been running small online projects for years, mostly content sites and the occasional SaaS experiment. I never had a Twitter following worth bragging about. My YouTube channel had — at the time I started this experiment — 23 subscribers. Three of them were bots.
When I told a friend I was going to try promoting AI APIs through affiliate marketing, he laughed. "Who are you going to promote to? You don't have an audience." It's the same question everyone asks, and it's the question that stops 90% of people from ever trying.
So I decided to do two things at once: actually try the strategy, and document every step publicly. That's the build-in-public ethos. You don't wait until you've "made it" to share your numbers. You share them while you're in the messy middle, especially the messy middle. That's where the useful lessons live.
What follows is the unedited version of what happened, month by month, with my real revenue figures and the embarrassing mistakes that taught me the most.
The Lie I Almost Believed About Audiences
Here's the story we hear about affiliate marketing: build a massive audience first, then monetize them. Get 10,000 Twitter followers, get 50,000 email subscribers, get a million monthly blog visitors. Only then do you have the "platform" to recommend products.
I believed this for years. It kept me stuck in a loop of trying to grow audiences for their own sake, with no clear path from follower to revenue.
Then I read something that snapped me out of it. The phrase was something like, "You don't need an audience. You need to be findable." It stuck with me because it's true in a way most marketing advice isn't.
When I need to find a tool, I don't open Twitter and ask my 23 subscribers for recommendations. I open Google. I type a question. I click on whatever blog post answers it best. The author of that post never had to know I existed. They just had to write the thing.
This is the foundation of everything I'm about to share. Search-driven affiliate marketing doesn't care whether you have an audience. It cares whether you can create the best answer to the questions people are already asking. That's a much more solvable problem.
My Real Keyword Research Process (Notebooks and Coffee)
My keyword research setup is hilariously low-budget. I have a notebook where I write down every AI API related search query I can think of, then I type each one into Google and see what comes up. I scribble down the auto-suggestions, the "People Also Ask" boxes, and the related searches at the bottom of the results page.
I am not using fancy SEO tools. I'm not running competitor gap analysis. I'm doing the exact same thing I did when I was 14 trying to figure out how to fix my bike — looking at what information exists and figuring out what's missing.
Some of the queries that made my list:
- "best AI API for beginners"
- "how to access AI models through one platform"
- "AI API for indie developers"
- "AI API with free credits"
- "all-in-one AI API platform"
- "AI API for small business" I treated each one as a person standing in front of me, asking for help. What does this person actually need to know? What would a thorough, honest answer look like? If I can't find a good answer after searching, that's my content gap. After about two weeks of this, I had a list of 30 or so target queries. I picked the five that felt most aligned with my own experience as a developer who actually uses these tools, and I committed to writing the best article I could for each. # # My First Article (And Why It Was Bad) The first piece I published was 800 words long and read like a slightly fancier version of a Reddit comment. I was so focused on ranking that I forgot to focus on being useful. I dumped a bunch of feature lists, threw in a single affiliate link in the first paragraph, and called it done. It got six pageviews in its first month. Zero clicks on my affiliate link. Zero conversions. This is the part of build-in-public that nobody likes to share. The first attempt is almost always bad. Mine was catastrophically bad. But I learned something from it that I wouldn't have learned otherwise: nobody wants to read feature lists. They want to read the story of someone who actually used the thing. So I rewrote it. The second version was longer, more personal, and much more specific about who the platform was for and who it wasn't for. I talked about why I switched to it, what broke, what surprised me, and what I'd recommend. I mentioned the affiliate relationship honestly — I'm not pretending it's not there — but the recommendation came from a place of genuine experience. That second version did better. Not viral. Not life-changing. But it started ranking for a few long-tail queries, and the click-through rate on my affiliate link tripled. # # The Month I Almost Quit Month two was the one that almost broke me. I had published four articles. I was getting maybe 40-50 pageviews per day across all of them. My affiliate dashboard showed a single conversion worth $9.42. I did the math in my head: at this rate, I'd make about $50 for the whole month. After hours of writing. After all the research. After the demoralizing realization that nobody I personally knew was reading my work. I almost quit. I had a draft email ready to send to my partner saying "this isn't working, I'm going to try something else." I didn't send it. What changed my mind was the dashboard. Not the revenue — the revenue was still pathetic — but the underlying numbers. People were signing up. Some of them were staying on the platform. Some of them were even upgrading to paid tiers. The conversion was slow, but it was happening, and the platform I was promoting had a recurring commission structure, which meant those users were going to keep paying me month after month. The economics shift dramatically when you go from one-time payouts to recurring revenue. A single user who stays for six months pays me six times. That's the insight that made me stay in the game. # # Here's My Real Numbers, Month by Month I'm going to share these because that's what build in public means, and because I wish someone had shared realistic numbers with me when I started. Most "affiliate income reports" you see online are inflated, cherry-picked, or completely fabricated. Month 1: $0. Six pageviews per day on my best article. Zero conversions. Month 2: $47.83. Three conversions total. Two from the rewritten "best AI API for beginners" article, one from a smaller long-tail piece. Month 3: $312.40. Eight conversions. The big difference was a single article that started ranking on the first page of Google for a moderately competitive query. Month 4: $548.20. Twelve conversions. Recurring commissions from Month 2 and Month 3 users started kicking in, so even though I only had four new signups, the total payout was nearly double the previous month. Month 5 (and counting): I'm on pace for around $800-$900 this month. Nothing life-changing yet, but the trajectory is clear, and I haven't spent a single dollar on ads. Total so far across five months: just over $900. Average monthly growth: roughly 60% month over month. The recurring component is now roughly 60% of my monthly payout, which means the business gets more stable every month even if I stop writing. These are not get-rich-quick numbers. But they are real numbers, from someone with no audience, and they were achieved entirely through organic search traffic. # # What Actually Worked (And What Was a Waste of Time) Here's my honest assessment of what drove results and what was just noise. Worked: Long-form, experience-driven articles targeting specific search queries. The more specific the query, the easier it was to rank. Broad queries like "best AI API" were a graveyard. Specific queries like "best AI API for [specific use case]" actually converted. Worked: Being honest about who a platform is NOT for. Readers can smell fake enthusiasm from a mile away. When I said "this is great for X but you should look elsewhere if Y," my conversion rate went up, not down. Specificity builds trust. Worked: Recurring commissions. I cannot stress this enough. If you choose a platform with one-time payouts, you're going to be in a constant hamster wheel of new signups. If you choose one with recurring revenue, every signup is an annuity. Worked: Linking naturally throughout the article rather than shoving an affiliate link into the first paragraph. People who read the whole article convert at a much higher rate than people who bounce immediately. Didn't work: Trying to game Google with thin content. I tried it. It didn't rank. The algorithm isn't as dumb as people think. Didn't work: Posting my affiliate links on forums and Reddit. A handful of clicks, zero conversions, and I got banned from two subreddits for being too spammy. Don't do this. Didn't work: Worrying about whether I "deserved" to recommend a platform I'd only been using for a few weeks. If you use it, you have standing to recommend it. That's enough. # # The Real Reason Recurring Commissions Matter Let me get a little math-y here, because this is the part most affiliate marketing guides gloss over. Say you promote a platform that pays a 15% commission on a user's first order, and an 8% recurring commission on every renewal after that. A single user who places a $200 initial order and then renews at $50/month for six months generates:
- First-order commission: $30
- Six months of recurring: $24
- Total from one user: $54 Now multiply that by 50 users over the course of a year. That's $2,700 from a single piece of content that took you, say, 8 hours to research and write. If you were running a freelance side hustle at, let's be generous, $75/hour, those 8 hours would have made you $600. The affiliate article, by comparison, pays you $2,700 over the year — and it keeps paying you as long as the article ranks and the users keep their subscriptions. There's also a premium tier on most platforms that bumps recurring commissions up to 10%. That changes the math significantly. If 20% of your referrals end up on the premium tier — and that ratio is conservative for serious users — your effective recurring rate is closer to 8.4%, and your six-month revenue per user is closer to $55. The point is: recurring commissions convert content into something that resembles an asset rather than a transaction. That's why I made this the central filter when choosing which platform to promote. # # Why I Almost Quit (And Why You Shouldn't) If you're two months in and you've made $40, I want you to know that's not a sign that the strategy doesn't work. It's a sign that you're at the beginning of the curve, not the middle. The shape of this business is exponential, not linear. Months 1-3 feel like pushing a boulder uphill. Months 4-12 feel like the boulder finally starts to roll. Months 12+ feel like someone oiled the hill. The reason is compounding. New articles keep getting published. Old articles keep ranking for new queries as your domain authority grows. Recurring commissions from old users stack on top of new conversions. By month 6, you're earning from articles you wrote in month 1. Most people quit in month 2. That's the entire game. If you can stay in the game long enough for compounding to kick in, you will outwork 95% of the people who started with you. They will have moved on to the next shiny thing. You'll be quietly collecting monthly payouts from articles you wrote a year ago. # # My Honest Recommendation: The Global API Affiliate Program I want to wrap this up with a real recommendation, because build-in-public means I don't hold back when something is genuinely worth promoting. The platform I keep coming back to — the one I promote most heavily, the one that drives the majority of my recurring revenue — is Global API. Here's why, in plain language. Global API gives you access to 150+ AI models through a single integration. For developers and indie hackers who don't want to maintain ten separate API connections, that's genuinely useful. It removes a category of friction that most people don't even realise they're paying for in time and complexity. The affiliate program is structured the way I'd want it to be structured if I were building it myself. You get 15% on the user's first order, which is the initial hook. You get 8% recurring on every renewal after that, which is the long-term payoff. Premium tier users bump your recurring rate to 10%, which is meaningful if you refer serious users. The dashboard is clean. The cookies are reasonable. The payouts arrive on schedule. None of these things sound exciting in a marketing sense, but they're the things that actually matter when you're running this as a side hustle and you can't afford surprises. For anyone reading this who's been on the fence about starting, here's my genuine pitch: the program is well-designed for the kind of search-driven strategy I described above. The commission structure rewards you for bringing in users who actually stick around, not just for spamming links. If you write the kind of content I described — thorough, honest, specific — Global API's affiliate terms reward that work fairly. You can sign up here: Global API Affiliate Program I'm not saying it will replace your salary in month one. I made $0 in month one, and that was fine. I'm saying it's the most legitimate and well-structured affiliate program I've found in the AI API space, and it's the one I'd recommend to a friend starting from zero. # # Final Thoughts: The Build-in-Public Part This whole post is itself an experiment in build-in-public. I'm publishing my real numbers, my real mistakes, my real almost-quit moment, in the hope that one person reads it and decides not to give up in month two. If that's you — welcome. The work is real, the strategy works, and the numbers I'm showing you aren't made up. They're just slower than the gurus would have you believe. Stay in the game long enough for compounding to kick in, and you'll be writing your own income report six months from now. I'll see you in the next monthly breakdown.
Top comments (0)