Side Hustle & Income
Does Affiliate Content Still Work Now That Search Answers Questions?
By Jim Vernon, Editor, AI Intelligence International · Published 10 February 2026 · Reviewed against our editorial standards · About the author
The affiliate playbook of thin comparison pages ranking for buying keywords stopped working when search results began answering the question directly. Sites built on that model lost the majority of their traffic in a short window.
Affiliate income itself did not die. What died was the assumption that generic informational content would keep sending free buyers forever.
Key takeaways
- What the shift actually removed: Queries with a single factual answer — specifications, prices, simple comparisons, "what is" definitions — now resolve on the results page.
- What still converts: First-hand testing with evidence: photographs, measurements, failure notes, long-term follow-ups.
- Traffic sources that are not search: Email is the most durable.
- Disclosure and duty of care: Disclose affiliate relationships clearly and near the link, not in a footer nobody reads.
What the shift actually removed
Queries with a single factual answer — specifications, prices, simple comparisons, "what is" definitions — now resolve on the results page. The click never happens, so the affiliate link is never seen.
This hit the middle of the funnel hardest: the roundup posts and spec tables that existed only to intercept a search on the way to a purchase.
It barely touched queries that require trust, testing or a recommendation from someone the reader recognises. Those still produce clicks because the reader wants a judgement, not a fact.
What still converts
First-hand testing with evidence: photographs, measurements, failure notes, long-term follow-ups. Nothing about this is cheap to fake, which is exactly why it retains value.
Narrow use cases: not "best laptops" but "laptops that survive site work", not "best CRM" but "CRM for a two-person letting agency". Specific buyers convert at multiples of general audiences.
Recommendations delivered inside a relationship — a newsletter, a community, a channel — where the reader already trusts the source and the link is incidental to the advice.
Traffic sources that are not search
Email is the most durable. You own the list and no ranking change removes it overnight.
Communities and forums work when you are a genuine participant and fail immediately when you are not. Moderators recognise extraction quickly.
Video and short-form still send buying traffic, particularly for physical products where seeing the thing matters. The production overhead is real but so is the conversion rate.
Disclosure and duty of care
Disclose affiliate relationships clearly and near the link, not in a footer nobody reads. This is a legal requirement in most jurisdictions and a trust requirement everywhere.
Recommend against products when they deserve it. A review section that never says no is read as advertising, because it is.
Keep commercial pressure out of the ranking. Once payouts decide the order, the reader eventually notices and the asset is finished.
Realistic economics
Conversion rates on genuinely relevant traffic run in the low single digits, and commissions vary from a few percent on hardware to substantial recurring shares on software.
That means income scales with buyer intent, not pageviews. A thousand readers who are actively purchasing beats fifty thousand who are browsing.
Recurring commissions on subscription products are worth disproportionate effort, because they turn one good recommendation into two years of income.
Building something that survives the next shift
Assume the traffic source you rely on will change. Diversify to at least two, and make one of them owned.
Invest in the parts that cannot be generated: your own testing, your own data, your own reputation for saying inconvenient things.
Keep pages updated. Outdated pricing and dead links are the most common reason otherwise decent affiliate pages stop converting, and the easiest thing to fix.
What changed and what did not
Answer panels absorb the informational queries that used to send traffic to definition and comparison posts. Clicks now concentrate on queries where the reader wants to see something for themselves: prices, hands-on results, specific configurations.
What did not change is that people still buy after reading someone they trust. The channel shifted; the mechanism did not.
Plan for fewer, more valuable visits. Traffic-volume strategies are the ones that broke.
Content types still earning
First-hand testing with photographs, measurements and things that went wrong. Comparisons of specific configurations rather than category overviews. Long-term follow-ups after six months of use, which almost nobody publishes.
Also: content for narrow situations. 'Best laptop' is answered by the panel; 'best laptop for field recording in cold weather' is not, and the reader is much closer to buying.
Each of these is expensive to produce, which is exactly why it survives.
Diversify the destination, not just the source
Email remains the highest-converting channel and the only one you own. Capture from day one, even at low traffic, and write to the list as though it were the product.
Add a second discovery channel where your audience already gathers — a forum, a video platform, a community — rather than treating search as the only front door.
Track revenue per visitor rather than sessions. In the current environment, that number can rise while traffic falls, and optimising the wrong one leads to a lot of wasted publishing.
Frequently asked questions
Is it worth starting an affiliate site now?
As a pure search play, no. As a narrow, genuinely tested resource with an email list attached, yes.
How long before an affiliate site earns?
Twelve to eighteen months is normal, and that assumes real testing and a distribution channel beyond search.
Do AI-written reviews rank?
Occasionally and briefly. They cannot supply first-hand evidence, which is the property that both readers and ranking systems increasingly require.
Which programmes pay best?
Recurring software commissions, followed by high-ticket B2B. Consumer marketplaces pay the least per sale and convert most easily.
Is affiliate content dead?
No, but thin roundups are. Original testing and narrow, specific advice still convert well and now face less competition.
Should I disclose AI assistance?
Disclose material connections as the law requires and be accurate about whether you tested a product. Claiming hands-on testing you did not do is the fatal error.