US advertisers will spend $13.81 billion on affiliate marketing in 2026, up 11.3% from $12.42 billion the year before, according to eMarketer’s 2026 forecast. That is faster than US retail ecommerce is growing overall, which is forecast at 6.7%. So the headlines claiming AI killed affiliate marketing are wrong on the numbers. The money did not leave. It moved, and it moved away from the exact kind of site most beginner guides teach people to build.
That matters because I run a portfolio of content sites, some of them affiliate-monetized, and I have watched buyer-guide traffic slide across the whole portfolio over the past year while total commissions in the channel kept climbing. This guide is the honest version, updated for a search landscape that looks nothing like it did when the classic “rank a review page and wait for Google” playbook still worked. It covers the real economics, how to pick a niche that survives, which programs still pay, and a 90-day plan grounded in how the channel actually behaves in 2026.
The economics, updated for 2026
Before you build anything, understand the math. These ranges come from industry surveys and publisher data, not from a guru’s screenshot.
Income by experience level:
- Year 1 (beginner): $0 to $500 per month. This is the investment phase, spent building content and learning what converts.
- Year 2 to 3 (intermediate): $1,000 to $4,000 per month if you have been consistent.
- Year 3 to 5 (advanced): $5,000 to $20,000 per month with an established authority site in a profitable niche.
- 5+ years (top tier): $20,000 to $100,000+ per month. The top 1% of affiliates clear over $1 million annually.
The average across all levels lands near $8,000 per month, but that figure is dragged up by a handful of top earners. The median sits closer to $1,200 to $2,500 per month. Startup costs run $500 to $3,000 in the first year for a domain, hosting, and a couple of tools.
The return still justifies the effort. Businesses earn roughly $12 to $15 for every $1 spent on affiliate marketing, which is why more than 80% of brands run programs. But there is a change to the risk section that older guides skip: algorithm dependency stopped being a tail risk and became the main event. A single Google update no longer just trims traffic. It can remove a review page from the results entirely because an AI answer now occupies the space that page used to hold. The rest of this guide is built around that reality.
What AI search actually changed
The damage is concentrated in one place: organic search traffic to review and buyer-guide content.
New York Times Wirecutter, the most cited affiliate publisher in the world, lost more than 60% of its Google search visibility between May and August 2025. Independent trackers measured the same fall from different angles: Sistrix at 69%, Ahrefs at 72%, Semrush at 65%. Wirecutter runs 4,500-word articles built on real product testing. If depth and authority alone were enough to survive AI Overviews, it would have survived. It did not.
The pattern is industry-wide. Google’s March 2026 core update knocked rankings down for 71% of affiliate sites, the hardest hit of any content category, and the July 2026 core update continued in the same direction. Pages built from templated comparisons, reproduced specs, and unattributed reviews saw drops of 60 to 80%. Meanwhile Ahrefs found AI Overviews cut click-through rates for the number-one organic result by 58%. Paul Cunliffe, who ran affiliates at Time Inc UK, described the revenue side bluntly: “a corresponding drop in affiliate revenue from 20-40% in some cases, which is really a massive hit.” Close to seven in ten publishers now say they are worried Google’s changes will hurt their affiliate business.
Here is the reframe that beginner guides miss. Total affiliate spend is up, so if review sites are collecting less of it, someone else is collecting more. Three groups are:
- Creators. On Awin’s network, the share of affiliate revenue flowing to content creators rose from 15.9% to 19.5% in a single year. A creator does not depend on a Google ranking that an AI summary can intercept. They own the relationship, and trust compounds with repeat exposure in a way an AI Overview cannot replicate.
- Cashback and loyalty platforms. These already command roughly 35% of affiliate ad spend, the single largest slice. Their advantage is structural: users have to visit the platform to activate rewards before purchasing, so the click survives and attribution holds even when an AI assistant made the recommendation.
- Coupon and deal publishers. Still large at around 42% of US affiliate revenue in the first half of 2025, but exposed, because an AI answer can surface a working discount code directly and cut the publisher out.
There is one more twist worth pricing in early. Your content may now be worth more as AI training and citation fuel than as a click generator. When ChatGPT discusses some product categories, a large share of the sources it draws on trace back to affiliate content. The review still shapes the purchase. It just does it inside the model, uncredited, instead of on a page where a commission fires. That is why getting your content cited by AI systems has gone from a curiosity to a survival skill. I broke the full version of this shift down in the analysis of why the review-site playbook died, and the broader traffic story in the piece on AI Overviews erasing publisher clicks.
Pick a niche that survives interception
Your niche still sets your commission rates, competition, and income ceiling. But in 2026 there is a fourth test layered on top: can the buying decision be fully resolved by an AI paragraph, or does it need a human who has actually used the thing?
The niches that hold up share three traits: high buyer intent, strong commission rates (20%+ recurring or $50+ per sale), and enough demand to build sustainable traffic. Compare these against the low-single-digit rates on most physical goods.
| Niche | Typical Commission | Example Programs |
|---|---|---|
| SaaS / AI Tools | 20 to 50% recurring | GetResponse, Jasper, Surfer SEO |
| Finance / Credit Cards | $50 to $200 per lead | NerdWallet-style comparison, card issuers |
| Web Hosting | $50 to $200 flat per sale | Kinsta, WP Engine, Cloudways |
| Online Education | 20 to 50% per sale | Teachable, Coursera, Skillshare |
| Health & Wellness | 10 to 40% per sale | Supplement and fitness brands |
| Email Marketing | 20 to 40% recurring | ConvertKit, AWeber, Mailchimp |
Recurring SaaS commissions have become the most defensible corner of the channel, because one converted reader pays for months and the decision benefits from real hands-on evaluation that an AI summary cannot fake. A tool paying 30% recurring on a $99 subscription earns you $29.70 every month the customer stays. The equivalent physical-product commission is often a single-digit percentage paid once.
The niche selection test still holds: can you name 10 products in this niche you have personally used, write 50 articles without running dry, and point to people who spend money solving problems here? If yes to all three, you have a niche. For the deeper breakdown, see the programs that pay 20 to 75% by niche.
Amazon Associates just got harder
If your plan was to blanket a niche with Amazon links, 2026 is the year to rethink it. Two changes landed within weeks of each other.
First, effective April 14, 2026, Amazon changed how onsite commissions are attributed. Commissions are now calculated only on the promoted product (the specific ASIN or its variants). Same-category purchases no longer count. In plain terms, if a reader clicked your kettle review and then bought a different toaster, that toaster used to earn you a commission. It no longer does. That “halo” on related purchases was a meaningful and quiet share of many affiliates’ Amazon income.
Second, in mid-May 2026 Amazon cut Associates commission rates by as much as 50% in several categories, reported by Adweek and confirmed by eMarketer, as part of a multi-month restructuring. Rates already sat low for most goods, roughly 1 to 10% depending on category, with a $100 electronics sale paying around $3. The cuts and the attribution change together mean the same traffic produces less Amazon revenue than it did a year ago.
None of this makes Amazon useless. Its universal trust and enormous catalog still convert cold traffic well, and it remains a sensible default for physical-product recommendations. But it strengthens the case that has been building for years: treat Amazon as one line item, not the whole business, and weight your effort toward direct brand programs and networks that pay real percentages. The Amazon-specific strategy guide covers how the affiliates still earning well from it have adapted.
Choosing programs that actually pay
Not all programs are equal, and rate is the wrong first filter.
Commission structure matters more than headline rate. A 10% recurring commission on a $200 monthly subscription ($20 every month) beats a one-time 50% commission on a $50 product ($25 once) after just two months, and it keeps paying. Cookie duration is the silent revenue driver: a 90-day cookie means you earn if the purchase happens within 90 days of the click, while Amazon’s 24-hour window is famously tight. Network programs on ShareASale, CJ, and Impact commonly offer 30 to 90 day windows.
Prioritize in this order: direct brand programs first (highest commissions and best relationships, though they require application), then affiliate networks for centralized access to thousands of programs and consolidated payments, then Amazon for reach. For a side-by-side on the major networks, see ShareASale vs CJ Affiliate vs Impact.
Build an audience you own, not a library you rent
The single most important strategic change in 2026 is this: build an audience you own before you build a library of pages you rent from Google’s algorithm.
An authority website is still the strongest hub, because you own the asset and organic traffic compounds when it arrives. To launch you need a domain and hosting ($100 to $300 per year on Cloudways, SiteGround, or similar), WordPress with a fast theme (free to $59 on GeneratePress, Kadence, or Astra), a free SEO plugin like Rank Math, and content, which is where almost all your time goes. The niche site blueprint covers the full build.
But a website alone is now a fragile plan. The affiliates growing in 2026 pair the site with channels where the click survives contact with AI: an email list, a creator presence, or both. Your email list is the only traffic source you fully control, and it converts at several times the rate of cold organic traffic because subscribers already trust you. Start collecting emails in month one with a genuinely useful lead magnet, a comparison spreadsheet or a checklist. See why an email list is worth $1 to $3 per subscriber per month.
Content that AI cannot answer in a paragraph
Affiliate income follows content that matches buyer intent, but the bar for what earns has risen. The pages that lost 60 to 80% of their rankings in 2026 were the ones an AI Overview could replace: generic roundups, spec regurgitation, reviews of products the writer never touched. The pages that held on had something a model cannot synthesize.
The highest-converting formats, in rough order:
- “Best X for Y” roundups, but only when built on real testing, not scraped specs.
- Product comparisons, which convert better than standard reviews and are stronger when they reflect side-by-side use.
- In-depth reviews and long-term ownership reports, the single product deep dive with pros, cons, and an honest verdict.
- How-to tutorials that naturally recommend the tool as part of the solution.
- Problem and solution posts that end in a specific recommendation.
The honest review framework earns trust and, increasingly, citations: state who the product is for and who it is not for, show real screenshots or results from your own use, compare pricing transparently including hidden costs, and include at least one genuine criticism. A review that is 100% positive reads like an ad to both readers and Google. Use the review SEO template and pay attention to where you place the links, which affects click-through more than link count.
One operational habit separates the affiliates who survived from the ones who did not: they measure revenue per article, not traffic. A page with strong earnings per session and weak distribution is a promotion problem worth solving. A page with heavy traffic and near-zero earnings was never a business. When traffic gets scarcer, knowing that difference is what tells you where to spend your remaining hours.
Traffic and attribution in 2026
You need traffic with buying intent, not volume. A thousand visitors evaluating a purchase beat fifty thousand casual browsers.
SEO remains the primary channel, but target commercial and transactional intent (“best,” “review,” “vs,” “alternative to,” “pricing”). Informational queries increasingly get answered inside AI Overviews without a click, so writing “what is affiliate marketing” for traffic is a losing trade now. Email is the highest-conversion channel and the one you control. Social platforms (YouTube, Pinterest, short-form video) work for awareness and creator-led recommendation, where trust compounds across repeat exposure. Paid traffic belongs only after you have proven which offers convert organically.
The attribution point deserves its own line, because it is new. Favor monetization paths where the click cannot be intercepted by an AI answer: creator-led recommendations to an audience you own, cashback-style programs where the transaction routes through a platform, and high-ticket or recurring-commission offers where one converted reader is worth months of income. To rank at all, keep the fundamentals of SEO tight while you build the channels that do not depend on ranking.
The tech stack
You do not need a dozen tools. This is what actually matters:
| Tool | Purpose | Cost |
|---|---|---|
| WordPress + Rank Math | Site + SEO | Free to $59/year |
| GeneratePress or Kadence | Fast theme | Free to $59 |
| Lasso or AAWP | Affiliate link management | $8 to $15/month |
| Google Search Console | Track rankings | Free |
| Ahrefs or Semrush | Keyword research | $99 to $129/month |
| ConvertKit or Mailchimp | Email list | Free to $29/month |
| Google Analytics 4 | Traffic analytics | Free |
That runs $100 to $300 per month with keyword tools, under $50 without them. AI tools like ChatGPT and Claude can speed up research and drafting, but the expertise and honest assessment have to come from you. Google’s helpful content system specifically targets auto-generated content that lacks real experience, and the 2026 core updates enforced that harder than ever.
Mistakes that end affiliate sites in 2026
Promoting products you have never used. Readers and Google both penalize generic reviews, and AI systems cite firsthand experience over summary. If you have not used it, either do not review it or clearly disclose you are summarizing public information.
Ignoring FTC disclosure. The FTC requires clear, conspicuous disclosure of affiliate relationships. Maximum civil penalties for covered violations sit at $53,088 per violation in 2026 (the 2025 figure was held in place for 2026). A simple “This post contains affiliate links” at the top of every post covers you. The full disclosure rules go deeper.
Betting the business on low-rate physical goods. With Amazon’s 2026 cuts and attribution change, the “thousands of Amazon links” model needs more traffic than ever to work. Weight toward higher-rate direct and recurring programs.
Publishing thin, templated content. This is the exact profile the 2026 core updates gutted. A 500-word summary of a product page adds nothing a model cannot generate. Depth, testing, and original media are the moat now.
Skipping the email list. Every affiliate site should collect emails from month one. It is the only audience an algorithm change cannot take from you.
A realistic first 90 days
Days 1 to 14, foundation. Choose your niche with the three-question test. Register a domain and set up WordPress. Apply to three to five programs, weighted toward recurring and direct brand offers rather than Amazon alone. Plan your first 20 topics around buyer-intent keywords.
Days 15 to 45, content sprint. Publish two to three articles per week, aiming for 10 to 15 total. Lead with comparison posts and tested roundups. Set up Search Console and Analytics. Launch a simple lead magnet and start the email list on day one of this window, not later.
Days 46 to 90, optimize and expand. Look at which content ranks and, more importantly, which pages earn per session. Deepen the winners with more firsthand detail. Add internal links across posts. Start a second channel where the click survives AI, a newsletter cadence or a creator presence, so you are not renting your entire audience from one algorithm.
By day 90 you should have 15 to 25 published articles, early rankings, and your first clicks. Revenue typically follows in months 4 to 6. The channel is bigger than it has ever been at $13.81 billion. The comfortable version, where you ranked a review page and waited for Google to send buyers, is the part that died. Build for the version that is still growing: an audience you own, content only a human could write, and a commission path that AI cannot quietly step into.