AI Overviews Gutted Affiliate SEO: What Still Gets Clicks

How much traffic have affiliate sites lost?

A lot, and the damage is uneven. If you accept the March 2026 core-update number in the brief, then 71% of affiliate sites lost rankings, and the pattern matches what many operators saw earlier: pages built to intercept generic review traffic were the first to fall. The bigger issue is not one ranking drop. It is the collapse of click supply on queries where AI Overviews answer the searcher's first question.

That is why the 83% no-click figure matters more than the raw ranking count. A page can still hold a visible position and collect less traffic if the Overview resolves the query before the user scrolls. The result is brutal on affiliate sites that depended on a long tail of informational searches, because those searches used to feed the funnel for free.

There is a second hit that operators miss. Even when a page survives, the average click-through rate changes because the searcher sees the answer, a citation strip, a map pack, and sometimes ads before the organic result. The page did not disappear. Its share of attention did.

Short version: the traffic loss is not just ranking loss. It is result-page displacement.

Which query types still send clicks?

Queries with clear buying intent still send clicks, especially when the searcher wants a product page, a price, a local option, or a specific comparison that the Overview cannot fully settle. Branded searches, model-number searches, and “best X for Y” terms with a real decision behind them are still the best survivors. Pure education is where the click bleed is worst.

That lines up with how Google documents Search. AI Overviews are designed to synthesize, not to replace every downstream action, and users still click when the answer only gets them to the next step. If the query implies selection, compatibility, or a transaction, you still have a path.

In practice, the clicks come from four buckets:

• Brand-plus-product searches like “Name brand air fryer review” or “Model X vs Model Y”.
• Problem-solving searches where the fix requires a tool, an app, or a SKU.
• Local intent where the user needs a nearby provider, not a summary.
• Long-tail comparison queries where the user needs pricing, specs, or availability.

One short sentence matters here. Intent wins.

If you are tracking ai overviews affiliate traffic, you should separate informational queries from decision queries before you draw conclusions. Otherwise you will overstate the damage and miss the pockets that still print clicks.

Why did review sites get hit hardest?

Review sites got hit hardest because they were easiest to compress. A page that repeats “best,” lists a handful of products, and copies the same buying cues as everyone else gives the Overview little reason to send a click. The model can summarize the consensus without needing your 1,800-word roundup.

Google Search Central has spent years pushing creators toward helpful, original content, and the Search Quality Rater Guidelines give low marks to pages that feel assembled for ranking rather than usefulness. That is not a moral lecture. It is a ranking constraint. When a review page has thin first-hand testing, generic pros and cons, and affiliate links that all point to the same merchant, it becomes easy to replace with a synthesized answer.

There is also a trust problem. In regulated niches and expensive purchases, readers expect proof, not a listicle assembled from merchant feeds. If your page does not contain fresh screenshots, pricing checks, field notes, or a real comparison framework, the Overview can often cover the same ground faster than you can.

Do not blame only the algorithm. Many review sites were already dependent on archive depth and stale URL volume. That model looked resilient because old posts kept ranking. Once the SERP stopped rewarding breadth for its own sake, the weakness showed up fast.

What broke was not content volume. It was undifferentiated content.

What earns a citation inside the Overview?

Pages that earn citations inside an Overview usually provide something the model can point to with confidence: a direct answer, a clean definition, a unique data point, a policy citation, or a page that resolves ambiguity fast. The citation slot goes to sources that are easy to verify and hard to paraphrase away. If your page only adds adjectives, it will not survive.

In affiliate verticals, the best citation candidates tend to be:

• Original pricing checks with dates.
• Side-by-side spec tables that are actually maintained.
• Unique testing notes from real use.
• Policy or regulatory references from named authorities.
• Clear screenshots of checkout flow, return terms, or plan limits.

That is why a page can win a citation without winning the click. The Overview may quote the fact and satisfy the searcher, which is useful for brand authority but weak for affiliate traffic. To get the click, your page has to offer a next step the Overview cannot compress: a live comparison, a calculator, a current deal, or a workflow.

The FTC’s endorsement guides matter here because disclosure and claim discipline shape what you can safely publish around affiliate offers. If your page mixes testimonial language, revenue language, and product claims too loosely, you increase compliance risk while gaining nothing in search. The safer path is plain reporting: what the offer says, what the page shows, and what you observed.

One useful test: if the page could be fully summarized in 5 bullets, the Overview probably can too.

Should affiliates shift budget from SEO to paid?

For many affiliates, yes, at least partially. Paid traffic is not a magic replacement, but it gives you control over placement at the exact moment SEO gets squeezed by AI Overviews. If your organic clicks fell on high-intent queries, paid search can buy back visibility faster than waiting for the next core update.

This is the part people argue about. The common reflex is to defend SEO as the “long game,” but a long game that no longer returns the click on your highest-value queries is just deferred loss. If the query still converts and the auction is viable, shifting some budget into paid search can be the rational move.

The evidence is structural, not sentimental. Search pages now contain more answer layers, more sponsored inventory, and more AI-generated synthesis. Meta's advertising policies are not the point here, but they show the same broader market truth: platforms are making paid distribution more explicit, not less. If your business depends on intercepting demand, you need at least one channel you can buy on demand.

That does not mean burning the SEO budget. It means resizing it. Keep SEO for pages that can still win demand capture, then use paid search to defend the exact queries that moved from “free traffic” to “expensive traffic.”

Practical rule: if a keyword used to pay for itself through organic alone, re-test it in paid before you kill it.

What does a post-Overview content plan look like?

A working post-Overview plan starts with fewer pages and better page types. Build around queries that need current facts, live proof, or a choice architecture the Overview cannot fully flatten. Then stop treating every article as a standalone traffic asset. Some pages should support paid search, email capture, or retargeting instead.

Here is the shape that still works.

• Comparison pages with original criteria and dated checks.
• Buyer guides tied to one specific problem, not a generic category.
• Brand pages that explain real differences, not recycled affiliate prose.
• Price and offer pages updated on a schedule.
• Support pages that collect leads when the click does not convert immediately.

For the desk, the main adjustment is operational. Monitor the queries that still send clicks every week, not every quarter. Google Search Console gives you the raw click and impression data. Pair it with manual SERP checks so you can see whether an Overview, a shopping box, or a citation strip is taking the slot.

The manual process is slower, but it keeps you honest. Most affiliates do not run it consistently, which is why they miss the exact moment a query flips from traffic source to dead end. If you only look at aggregate traffic, you will keep producing pages for a SERP that stopped paying you.

Start with 20 queries, not 2,000. Check them every week. Kill what the Overview already owns.

There is still money in affiliate SEO, but it is no longer broad, and it is definitely not passive. The pages that survive are the ones that answer a live commercial question better than the Overview can, or the ones that move the reader into a channel you control. Everything else is inventory with a shrinking click rate.

Frequently asked questions

Are AI Overviews killing all affiliate traffic?

No. They are cutting traffic hardest on informational and generic review queries. Clicks still show up on branded searches, model comparisons, problem-solving queries, and pages that add live proof or current pricing the Overview cannot fully replace.

What kind of affiliate page still has a chance to rank?

Pages with a clear job still have a chance. A page that compares two products with current specs, dates its pricing checks, or explains a specific buying decision is harder to compress than a generic listicle.

Should affiliates stop doing SEO now?

No. They should narrow it. SEO still matters for queries that convert and for pages that earn citations or direct clicks, but paid search may now be the faster way to defend the highest-value terms.

Sources

Named rather than linked — verify before relying on any figure below.

• Google Search Central
• Google Search Quality Rater Guidelines
• FTC Endorsement Guides
• Meta Advertising Policies