Sitelinks That Actually Get Clicked: Patterns From Real SERPs
The mechanism behind which sitelinks earn clicks, the exact conditions where they pay, and when adding more hurts you.
The short version
- Sitelinks earn incremental clicks only when each one serves a distinct sub-intent your headline cannot cover.
- Provide six to eight sitelinks and let Google pick the best four per query; do not hand-pick exactly four.
- On navigational brand queries, watch for sitelinks cannibalizing main-headline clicks and dragging down conversion rate.
- Preview mobile separately: fewer sitelinks show and they can bury your main CTA.
- Every sitelink destination needs a next action; 'About Us' and 'Blog' spend budget to produce bounces.
The sitelinks that get clicked are the ones that offer a destination the headline does not already cover: a specific intent branch (pricing, a product category, a location) that a subset of searchers wants instead of your main landing page. If your sitelinks just restate the ad or point to the same funnel, they add visual bulk but earn no incremental clicks, and in some layouts they can cannibalize your main headline click.
What sitelinks actually do in the auction
Sitelinks are not a ranking lever in the way people assume. Google does not reward you for simply having them. What they do is change the physical footprint your ad occupies and the number of distinct intents your ad can satisfy in one impression. When your ad shows four sitelinks on desktop, you are no longer offering one answer to the query. You are offering five.
The decision Google makes per impression is whether showing sitelinks (and how many) improves the expected experience versus the alternative uses of that space. This is why sitelinks show inconsistently. The same ad can show four sitelinks for a broad brand query and zero for a narrow transactional one, because on the narrow query the main headline already resolves the intent and extra links are noise.
The practical consequence: sitelinks earn clicks when they let a searcher self-sort into a sub-intent your headline cannot address in 30 characters. A searcher types your brand name, but half of them want pricing and the other half want support. Your headline picks one. Sitelinks let the other half land where they wanted without a bounce.
The pattern that works: orthogonal destinations
Look at the sitelinks that survive on competitive SERPs and you see a consistent shape. They are orthogonal to each other and to the main ad. Each one points to a genuinely different page serving a genuinely different intent. 'Pricing,' 'Free Trial,' 'Integrations,' 'Case Studies' on a SaaS ad. 'Men's,' 'Women's,' 'Sale,' 'Store Locator' on retail. No overlap, no redundancy.
The failure case is sitelinks that are synonyms. 'Our Solutions,' 'What We Offer,' 'Services,' 'Products' are four links to roughly the same place in the searcher's mind. They fill the space but give nobody a reason to click a specific one over the headline. Google's own systems often suppress these because they predict low incremental value, so you lose the real estate anyway.
A useful test before you save a sitelink set: could you swap any two descriptions and have the ad still make sense? If yes, they are too similar. Each sitelink should answer a different question a searcher might have at the moment of the query.
Where sitelinks pay, and the numbers to use
Sitelinks pay most on high-intent-ambiguity queries: brand terms, broad category terms, and competitor terms where the searcher's specific goal is not encoded in the words they typed. On a tight long-tail query like 'buy 10mm hex socket set,' the searcher told you exactly what they want. A sitelink to your blog is a distraction.
Decision rule for auditing an existing set: pull sitelink-level performance from Campaigns, then Assets, and segment by sitelink. If a sitelink has served a meaningful number of impressions (wait for roughly a few thousand before trusting it, because sitelink impressions accrue slowly and small samples lie) and its click share is a tiny fraction of the others, it is dead weight. Replace it, do not just delete it, because fewer than two eligible sitelinks often means none show at all.
Give Google raw material to work with. Provide six to eight sitelinks at the ad group or campaign level even though only up to four show on desktop and fewer on mobile. The system picks the best combination per query. Providing exactly four removes that optimization and forces whatever you wrote to show even when a different mix would perform better.
Where sitelinks backfire
The first trap is cannibalization on navigational brand queries. If someone searches your exact brand name intending to reach your homepage, and your top sitelink is 'Login,' a chunk of your main-headline clicks shift to the login link. That can look like healthy sitelink CTR while your overall conversion rate drops, because login traffic does not convert the way new-visitor traffic does. The sitelink stole a click, it did not create one.
The second trap is mobile. Mobile SERPs show fewer sitelinks, often as a horizontal carousel, and the vertical space is scarce. Aggressive sitelinks on mobile can push your actual landing page CTA below the fold of the ad block. Check your mobile preview separately. What reads as a clean four-link grid on desktop can be a cluttered scroll on a phone.
The third trap is sending every sitelink to a soft page. 'About Us,' 'Blog,' 'Careers' feel like filler for a reason: they rarely convert. A sitelink that earns a click but lands on a dead-end page spends budget to produce a bounce. Every sitelink destination should have a next action, not just information.
How to build a set this afternoon
Start by listing the three to five distinct reasons a person who sees this ad might not want your main landing page. Pricing hesitation, wanting proof, wanting a specific product line, wanting to compare. Each of those becomes a sitelink pointing to a page that resolves it.
Write descriptions that preview the destination, not the brand. 'Plans from $X/mo' beats 'View Our Pricing' because it answers the question in the link itself and filters out tire-kickers before the click, which protects your conversion rate. Use the two description lines when the extra context earns the click; skip them when the title is self-explanatory.
Then leave it alone long enough to read. Sitelink data accrues slowly because they only show on a subset of impressions. Resist swapping sitelinks weekly. Give a new set a few weeks or a few thousand sitelink impressions, whichever comes first, before you judge any single link. What you cannot see is the counterfactual: Google will not tell you how the impression would have performed with a different sitelink mix, so some of your read is inference, not measurement.
Questions people also ask
How many sitelinks should I add?
Add six to eight at the campaign or ad group level even though only up to four show on desktop and fewer on mobile. More options let Google's per-impression optimization choose the best mix for each query. Providing fewer than two eligible sitelinks risks showing none at all.
Why do my sitelinks sometimes not show?
Sitelinks show when the system predicts they improve the result for that specific query and when your ad has enough Ad Rank for the expanded format. On narrow, high-specificity queries the headline already resolves intent, so Google often suppresses sitelinks as redundant. This is expected behavior, not an error.
Where do I see which individual sitelinks are performing?
Go to Campaigns, then Assets, and view at the sitelink level. Segment by individual asset to compare impressions and clicks. Wait for a few thousand impressions on a sitelink before trusting its read, since sitelink impressions accrue slowly and small samples are misleading.
Can sitelinks actually hurt performance?
Yes. On navigational brand searches a prominent 'Login' or 'Support' sitelink can pull clicks away from your main headline, shifting traffic toward lower-converting pages. High sitelink CTR with falling overall conversion rate is the signature of this. Sitelinks pointing to dead-end pages like blog or careers also spend budget without a next action.
How often should I rotate sitelinks?
Not often. Because sitelinks only show on a fraction of impressions, data accrues slowly and frequent swaps reset your read before you have a reliable signal. Give a new set a few weeks or a few thousand sitelink impressions before judging any single link, and replace underperformers rather than deleting them to keep enough eligible.