Every freelance writer knows that specific kind of tired. You ship an article at 11 p.m., fire off the invoice, then open a fresh browser tab and start pitching the next client. The money only lands when your fingers are moving. Stop writing, stop earning. I lived in that loop for years — chasing per-article rates, locking in the occasional retainer, refreshing my email hoping a publication needed a 1,500-word explainer on some niche framework.
It pays the bills. It does not build anything.
What I wanted — what I had been circling around for months — was a stream of income that did not require me to invoice every single time. Passive income, recurring revenue, whatever you want to call it. I had read enough posts about writers building productized newsletters, selling templates, and launching tiny SaaS tools to know the pattern: you trade hourly effort for compounding returns. Affiliate income fit that mold. I just had to find the right product to pitch.
This is the honest, slightly messy journal of how I spent my first 90 days promoting an AI API service called Global API. Real numbers. Real setbacks. Real calculations. If you are a writer thinking about whether the affiliate lane is worth your time, this one is for you.

The Setup: What I Brought to the Table

I am not a marketer. I am a writer who happens to know how to install a WordPress theme and argue with DNS settings. My only real assets at the start were:

• A small tech blog pulling around 2,000 monthly visitors from years of writing about dev tools and side projects.
• A Twitter account with roughly 800 followers, mostly other developers I had met through client work.
• About a year of hands-on experience using AI APIs for real client projects. I had built chatbots, summarizers, and one very weird poetry generator for a poetry magazine. I knew what I was talking about when I wrote about these tools. I spent a few evenings researching affiliate programs. Most of the AI platforms I had been using either had no affiliate program at all or offered a flat one-time payout. That meant every dollar I earned stopped the moment the signup happened. If the user churned, I did not care — but I also did not benefit if they stayed. Then I found Global API's program. The structure was different: 15% on first orders, 8% recurring on monthly renewals, and 10% on premium tier referrals. Plus, the platform itself had 150+ models available through one unified interface, which gave me a lot of angles to write about. The recurring piece was the part that made me stop and take notes. If someone signed up through my link and stayed subscribed, I would keep getting paid. That is the kind of math that makes a freelance writer's ears perk up. I signed up on a Tuesday night. I had no plan beyond "write some articles and see what happens." # # Days 1–30: The Slow First Month I set a small, almost embarrassingly low goal for month one: get one paying signup. That was it. One conversion would prove the model worked. Everything after that would be upside. Week one. I drafted an outline for a comparison article — the kind of piece I had written plenty of times as a freelancer, except this time I owned the channel instead of invoicing someone else for it. The piece ran about 1,800 words and walked through how I had been using different AI APIs for actual client projects. I embedded Global API as my top recommendation, mostly because the unified access to 150+ models made my workflow noticeably easier when I was juggling multiple clients. I published the article on my own blog and cross-posted it to Dev.to. Then I waited, which is the hardest part of any writing-based income stream. Week two. Numbers started trickling in. The Dev.to version pulled 340 views in the first seven days. My blog version got 120. Three people clicked my affiliate link. Zero conversions. I reminded myself that this was not a retainer gig where I could invoice for drafts — this was a publishing business, and publishing businesses take time. Week three. The Dev.to post started picking up traction on long-tail search terms. Combined views crossed 520. Affiliate clicks climbed to 11 total. One signup. Still no paid conversion, but I had something to look at in the dashboard. Week four. I published my second article — a tutorial walking readers through building a simple chatbot. Again, I positioned Global API as the recommended platform because of how clean the integration was. On day 28 of my experiment, the original signup converted to a paid Pro plan. The 15% first-order commission landed in my account: $3.00. Month one total: two articles published, 750 combined views, 14 affiliate clicks, two signups, one paid conversion. $3.00 in earnings, $0.00 recurring because that does not kick in until month two. Not life-changing. But it was the first dollar I had ever earned from a piece of content that did not require me to invoice a client. That alone felt like proof of concept. # # Days 31–60: When the Math Starts Working Going into month two, I had one paying referral, 14 total clicks to my name, and two articles doing their slow thing on the internet. My new goal: publish three more articles and hit $50 in cumulative earnings by the end of the month. Week five. I shipped article three — a case study about how I used AI APIs to ship a feature for an actual client project. The piece was specific, almost embarrassingly specific. I named the client vertical, the workflow problem, and the API calls that solved it. It pulled 280 views in week one with a noticeably higher click-through rate on my affiliate link, because the audience was developers who recognized themselves in the situation. When a reader sees their own pain point in your writing, they are far more likely to act on the recommendation. Week six. The original comparison post from month one hit 1,200 total views. Google started ranking it for a few keyword variations I had not even targeted. Affiliate clicks rose to four or five per day. Two more readers converted to Pro plans that week. Week seven. Article four went live — a beginner-friendly guide to getting started with AI APIs, clocking in at 2,200 words. It was the most labor-intensive piece I had written, partly because explaining the basics forces you to question every assumption you have been making for a year. Beginners also tend to convert at a higher rate, because they are actively looking for someone to tell them what to use. Week eight. The first recurring commission arrived: $1.60 from my original signup's second month on the platform. It was a tiny number, but I stared at it for an embarrassingly long time. That single line item was recurring revenue. It was going to show up next month, and the month after that, as long as the customer stayed. I had never had an income line that worked like that before. Every other dollar I earned had to be re-pitched, re-written, and re-invoiced. I also published article five, a pricing-focused piece aimed at developers who were nervous about cost. I was starting to learn which angles actually moved the needle. By the end of month two, my running totals looked like this: three new articles published, five total, 2,100 combined views across all my content, and 58 affiliate clicks. The cumulative earnings picture was the part that mattered: first-order commissions on three Pro conversions, plus the first month of recurring. The number was small, but the trajectory was clear. # # What Actually Changed About My Income Let me do the math out loud, because that is the part I wish more writers shared. When I bill per article, I get paid once. If a client wants a follow-up piece, I have to negotiate a new rate, sign a new contract, and write the thing. If I am lucky, I land a retainer — but even retainers are bounded. They usually end. They almost always end. Affiliate revenue does not work like that. The first-order commission on a $20 Pro plan is $3. The recurring commission on that same plan, every month the customer stays, is $1.60. Multiply that across twelve months and the math starts to look like a writer's dream: $3 to acquire a customer, $19.20 over the next year if they stay. If they upgrade to a premium tier, the recurring bumps up to the 10% premium rate. The customer pays for themselves many times over, and you did not have to invoice anyone. Compare that to a per-article gig at $150. You write, you invoice, you wait. The article lives on a publication's domain where you cannot add affiliate links, you cannot update it, and you cannot benefit from it compounding. A blog post on your own site, by contrast, sits there indexing, ranking, and quietly earning you money while you sleep. I am not going to pretend my three-month earnings replaced my freelance income. They did not. But the shape of the income changed, and that matters more than the dollar amount at this stage. For the first time in my career, I had a slice of revenue that did not require me to be at the keyboard. # # The Honest Struggles It was not all clean graphs and rising numbers. A few things I want to call out for anyone thinking about doing this:
• Month one is slow. You are essentially writing on spec, the same way you would pitch a cold publication. The difference is you own the asset afterward. Knowing that helped me keep going when the first two weeks returned almost nothing.
• You need at least five to seven articles before the math starts to feel real. Below that, you are basically waiting for one of your pieces to get lucky. Above that, you have enough surface area for search traffic to find you.
• The wrong commission structure kills the model. I almost signed up for two programs that offered one-time payouts. Had I done that, I would have walked away after month one thinking affiliate marketing was a waste of time. The recurring piece is what turns this into an actual side business.
• It still feels like client work in some ways. You are still pitching, still writing, still trying to earn attention. The difference is you are pitching to readers, not editors,