AI Overviews and Affiliate Sites: What Google's AI Answers Do to Buyer-Guide Traffic and What Still Gets the Click
An AI Overview is the generated paragraph at the top of some Google results that answers the query before any link, with citations to the pages it drew from.
For an affiliate site it changes two things: some informational posts now lose the click to the answer box, and some buyer guides now get cited inside it. Which happens to which page depends on what the page says, and that's the part you control.
This post covers which queries trigger overviews and which don't, what the click data shows so far (with the honest caveat that it varies), which elements of a page get cited, why a verdict with a reason still earns the click where a spec summary doesn't, what to stop writing, and where to move the hours.
The snippets post covers the answer-first format; overviews reward the same format for different reasons.
The short answer
Overviews appear mostly on informational queries and take the click when the answer is a fact. They appear less on "best X" queries and, when they do, cite and link product pages and guides, which still sends traffic.
Write informational posts only where the answer needs judgment; write guides with verdicts, reasons and caveats an overview can cite but can't replace; stop writing one-fact posts. Expect the informational half of the site to lose more than the guides.
Which queries trigger an overview
| Query shape | Overview likelihood | What the overview does | Effect on the page |
|---|---|---|---|
| What is / how does X work | High | Answers it, cites two or three pages | Click lost unless the page adds something |
| How many / how long (a number) | High | Gives the number | Click mostly lost |
| Can you / is it safe to | High | Yes or no with a caveat | Click lost unless the caveat needs the page |
| X vs Y | Medium | Summarises differences, links both | Click partly kept; the verdict still needs the page |
| Best X for Y | Lower, rising | Lists products with reasons, cites guides | Click kept when the guide is cited; lost to the listing when it isn't |
| X review / is X worth it | Medium | Summarises sentiment, cites reviews | Click kept for a verdict with a reason |
The pattern: the more the answer is a fact, the more the overview keeps the click; the more it's a judgment, the more it has to send the reader somewhere.
What the data shows, with caveats
Studies from SEO tool vendors report lower click-through on queries where an overview appears, with reported drops ranging widely depending on the query set and how the study counted; the one consistent finding is that informational queries lose more than commercial ones. Google's own statements say overviews send clicks to a wider range of sites.
Both can be true: fewer clicks per query, spread differently. For an affiliate site, the practical reading is in Search Console: compare impressions and clicks on your informational pages over the last year against your guides. Most sites see the first line fall and the second hold.
What gets cited
- Self-contained passages that answer the query directly: the same 40 to 60 word format that wins snippets.
- Specific claims: a number, a named product with a reason, a trade-off stated plainly. "The Metro's cut-off beam is why it's the commuter pick" is citable; "there are many great options" isn't.
- Pages that already rank on the first page for the query. Overviews draw from the result set, not from the whole web.
- Tables and lists for comparisons and "best" queries, where the overview is assembling a short list.
What doesn't get cited: pages that restate the listing, pages with the answer buried under a preamble, and pages whose "answer" is to tell the reader it depends. The E-E-A-T post covers why a stated basis and a real opinion matter to Google's systems; overviews are one more place that shows.
Why a verdict still earns the click
An overview can tell a searcher that commuter lights need 200 to 400 lumens and a cut-off beam. It can't tell them which one to buy for their commute, stand behind the choice, or explain why the popular pick is wrong for them, because it has no stake and no experience; it summarises what pages said and links to them.
The page that said it, with the reason, is the one the searcher opens next. That's the guide's verdict, the product section's caveat, the comparison's deciding difference. A guide built from those earns the click after the overview; a guide that's a spec list has been summarised and is done.
What to stop writing
- One-fact posts. "How many lumens for a bike light", "how long does a bike light battery last". The overview answers them and keeps the click. If they feed a guide, fold them into its FAQ.
- Definition posts for terms a reader could get in one sentence.
- "Both are great" comparisons. The overview already does the neutral summary; a comparison without a verdict adds nothing.
- Guides without a reason per pick. A list of ten products with specs is exactly what an overview assembles itself.
Keep: informational posts where the answer needs judgment ("is a 1,000-lumen light too bright for the road"), everything that feeds the guides through internal links, and every guide with a real verdict. The intent post covers the split; the overview shifts the line toward the commercial side.
Where to move the hours
- Guides with a verdict, a reason per pick and a caveat per product: the content an overview cites and can't replace.
- Comparisons that pick a winner for a named reader.
- The FAQ at the end of each guide, in answer-first form: it collects the one-fact queries as citations rather than as separate pages that lose.
- A refresh pass on existing informational posts: add the judgment, or fold them into a guide.
Doing it with AffSERP
The elements that get cited and still earn the click are the structure of a guide written by AffSERP's X Best Review tool: a verdict with a reason in the first screen, a caveat per product drawn from the review pattern, a comparison table, and an answer-first FAQ that absorbs the one-fact queries.
The Informational Article tool is for the questions that need judgment; the one-fact questions go in a guide's FAQ instead, which is a choice made in the keyword list rather than in the tool.
FAQ
Do AI Overviews reduce traffic to affiliate sites?
On the queries where they appear, yes, and they appear mostly on informational queries: 'how long does', 'what is', 'can you'. Click-through on those drops, with the size varying a lot by study and by query. On 'best X' queries overviews appear less often and, when they do, tend to list products with links, which still sends clicks to the sites cited. The informational half of a site loses more than the guides.
How do I get cited in a Google AI Overview?
Rank on the first page for the query, answer it in a self-contained passage (the same format that wins featured snippets), and say something specific the overview can attribute: a number, a verdict, a named trade-off. Overviews cite pages they drew from; a page that restates the listing has nothing to draw. There's no markup or opt-in that earns a citation; the passage does.
Should I stop writing informational posts because of AI Overviews?
Stop writing the ones whose whole value was a one-line fact ('how many lumens is bright enough'); the overview answers those now and keeps the click. Keep writing the ones that need judgment or a comparison, and keep the ones that feed the guides through internal links, because their job was never the click alone. Shift the hours toward guides and comparisons, where the verdict still needs a page.
What content survives zero-click search?
Content where the answer is a decision rather than a fact: which one to buy for a specific reader, what you give up at each price step, why the popular pick is wrong for some people. An overview can summarise a fact; it can't stand behind a recommendation, and it links to the page that does. Verdicts with reasons, caveats from real review patterns and comparisons between two specific products are what still earn the visit.
AffSERP guides lead with a pick and a reason, carry a caveat per product and end with an answer-first FAQ, which is the shape that gets cited and still gets the click.
Start with free words