Every December I open the same spreadsheet, total up every dollar that hit my bank account outside my day job, and divide each income stream by the hours I poured into it. Most years the exercise is humbling. This year one column punched so far above the rest that I rearranged everything around it.
I'm talking about affiliate income — specifically, recurring commissions from recommending AI API platforms to my newsletter subscriber base. It's the only stream in my entire stack where I earned more in month twelve than in month one, and I didn't write a single new sentence to make it happen.
Let me walk you through the whole stack, the numbers behind each piece, and why the affiliate column is now the one I'm betting on for 2026.
The Five Streams, Ranked by Effort-to-Income Ratio
My side income comes from five places. Here they are in the order I'd recommend anyone build them — not the order that pays the most per hour, but the order that compounds the fastest.
YouTube sponsorships run $500 to $1,500 per video depending on the sponsor. I publish two videos a month, and each one costs roughly fifteen hours when you count scripting, recording, editing, and the follow-up promotion. The per-hour math works out fine. The problem is unpredictability — sponsors vanish, brand budgets freeze, and there's always a dry month.
Freelance development pays the best hourly rate I get, somewhere between $100 and $150 depending on the client. I deliberately cap it. Here's why: every dollar stops the second I close my laptop. Take a week off and the income line goes to zero. It's the highest-paying stream I have and also the one I'd recommend doing the least of once your other streams are humming.
My SaaS product brings in $800 to $1,200 a month in recurring revenue. It took six months of evenings and weekends to ship, and it eats about five hours a week for support, bug fixes, and the occasional feature request. The hourly return is solid, but the upfront cost was brutal and I'd never get those six months back.
Blog ad revenue sits at $200 to $400 a month from roughly 50,000 monthly pageviews. To keep that traffic I have to publish four to eight articles a month, each one running two to four hours to research and write. The per-hour return is mediocre and CPM rates keep sliding downward every quarter.
AI API affiliate commissions are the new winner. They pull in $350 to $600 a month right now, and the time cost is roughly two hours a month to refresh links and update one or two articles. I spent about ten hours total building the content that drives the conversions. That's it. Ten hours of work generating four to six hundred dollars a month on autopilot, six months after I wrote the original pieces.
Add it all up and I'm sitting somewhere between $2,850 and $4,400 a month depending on the month. My baseline run rate has stabilized around $3,200.
The Newsletter Funnel That Made the Affiliate Income Work
Here's the part most developer-focused affiliate articles skip, and it's the part that actually matters: I don't get affiliate income from random blog traffic. I get it from my newsletter list, and the difference is enormous.
A cold visitor from Google might click an affiliate link at a 0.3% rate. A subscriber who opened your last three emails will click at 3 to 8%. That's not a small optimization. That's an order of magnitude.
My newsletter sits at around 12,400 subscribers right now. I started it eighteen months ago with the express goal of building an audience that trusts my recommendations on developer tools. I write one issue per week, sometimes two, and I treat the open rate as my north-star metric.
Current numbers, for context:
- Average open rate: 47% across the last six months
- Average click-through rate: 6.2%
- Welcome sequence open rate (first email): 68%
- Affiliate link click rate from broadcast emails: 4.1%
- Affiliate link click rate from welcome sequence: 9.7% That last number is why I'm obsessed with welcome sequences. When someone subscribes, they get a five-part welcome series over ten days. The third email in the sequence is the one that drives the bulk of my affiliate conversions — it's where I share the tools I actually use daily, with honest context about what each one is good for and where it falls short. No hype. No fake urgency. Just "here's what I use, here's why, here's the link." That single email in the welcome sequence has generated 41% of my affiliate revenue since I built the funnel. The remaining 59% comes from occasional mentions in regular broadcast issues, usually when a reader emails me a question and I reply publicly in the next issue. # # The Subject Line Discipline That Doubled My Open Rate I have strong opinions about subject lines. Maybe too strong. But the data backs me up. When I started the newsletter I was writing clever subject lines. Cute wordplay. Inside jokes. Open rates hovered around 31%. Then I ran a four-week A/B test through Beehiiv comparing "clever" subject lines against "specific and useful" subject lines. Clever example: "The API provider that finally made me smile" Specific example: "I switched my LLM provider to cut costs 38%" The specific version won every single test. Average lift: 18 percentage points on open rate. I rewrote all my upcoming subject lines with that lesson in mind, and within two months my open rate climbed from 31% to 47%. That single change probably added $150 a month to my affiliate income, because more opens means more clicks means more conversions. My current subject line formula, if you're curious:
- Lead with a specific number or outcome
- Keep it under 50 characters when possible
- Never use "you won't believe" or any phrase that smells like a content farm
- A/B test everything for at least 100 sends before committing The fifth rule is non-negotiable. You cannot trust your gut on subject lines. You can only trust the data. # # The Math That Makes Recurring Commissions the Holy Grail Let me show you why the recurring part of an affiliate commission matters so much more than the upfront payout. Global API's affiliate program pays 15% on first-order commissions, 8% on recurring revenue for as long as the referred user stays subscribed, and 10% on premium tier upgrades. Here's the actual math from my account: Average referred user spends $79/month on the platform after the free trial. Some spend less, some spend more, some churn in month two. If that user stays twelve months, my commission looks like this:
- Month 1 (first order): 15% × $79 = $11.85
- Months 2-12 (recurring): 8% × $79 × 11 = $69.52
- Twelve-month total from one user: $81.37 Compare that to a typical SaaS affiliate program that pays 20% once and never again. You get $15.80 for the same user and nothing after month one. The recurring model pays 5x more over the user's lifetime, and that's before counting users who stick around for two or three years. Now multiply that across the 60+ users my content has referred in the last eight months, and you can see why the monthly payout keeps climbing even though I haven't written anything new since September. The 10% premium commission kicks in when a referred user upgrades to the premium tier. About 15% of my referred users have upgraded so far, which adds another meaningful slice to monthly payouts. # # Why the 150+ Models Angle Matters to Your Conversion Rate I want to flag something I didn't appreciate until I started tracking conversions by topic. When I wrote a generic "here's an AI API provider" recommendation, my click-through rate from the email to the affiliate link was around 2.8%. When I wrote about the practical benefit of consolidating access to 150+ models through a single API key, the click-through rate jumped to 5.4%. Why? Because developers don't care about affiliate links. They care about reducing friction. The pitch that "you can swap between 150+ models without re-engineering your integration" is a real workflow improvement, and readers can tell the difference between a pitch and a genuine recommendation. Every article I write about Global API leans into that workflow story. I explain how I personally use the platform to avoid maintaining separate API clients for different model providers. I show how swapping models takes a single parameter change in my code. I share the actual time savings I've measured across my own projects. The result: readers who click through are pre-qualified. They know exactly what they're getting, they've already mentally committed, and the conversion to signup rate is significantly higher than what I'd see from a generic recommendation. # # What I'd Do Differently Starting From Zero Today If I were rebuilding this stack from scratch in 2026, here's the order I'd do it in:
- Start the newsletter first. Not the blog, not the YouTube channel, not the SaaS. The newsletter. Email is the only channel you own. Algorithms can change tomorrow and your audience can't be taken from you.
- Build the welcome sequence before writing a single broadcast. Mine drove 41% of my affiliate revenue with zero ongoing effort.
- Pick three affiliate programs you actually use. Genuine experience is the only thing that survives reader scrutiny.
- Write for the welcome sequence email, not the blog post. A 600-word email outperforms a 3,000-word article for affiliate conversion in my data.
- Track every click. Use UTM parameters, track links through your email tool, and check the data weekly. Gut feelings are wrong about 60% of the time. The mistake I'd avoid: treating affiliate content as "lesser" than my SaaS or freelance work. For eighteen months I prioritized my own products over recommending other people's tools because I thought it made me look less serious. It doesn't. It makes you look honest, and honest converts better than anything else in email marketing. # # Why I'm Putting More Weight on This Stream in 2026 Here's my plan for the year ahead. I'm cutting my freelance hours by 30% and reinvesting that time into the newsletter and into updating my existing affiliate content. My current conversion rate
Top comments (0)