GOOGLE ADS SITELINK EXTENSION BEST PRACTICES

The google ads sitelink extension best practices that matter most are not the ones Google publishes in its help center, those are structural minimums, not performance strategies. This article is grounded in what I observe consistently managing real accounts: which choices produce measurably better impression share, which copy patterns lift click-through rate, and which operational decisions around asset hierarchy save time without sacrificing control. Understanding how sitelink extensions work in the auction is the prerequisite, once you know how Google selects and renders them, the best practices follow from the mechanics.

The framework I use has four layers: link text quality, description line quality, asset hierarchy decisions, and pool size management. Each layer is independent but compounds with the others. A perfect link text paired with empty description lines leaves expanded format eligibility on the table. A dense, well-written description attached to a generic link text wastes the copy budget. The google ads sitelink extension best practices I lay out here address each layer separately and then show how they interact in practice. One thing worth stating upfront: sitelinks do not have their own search volume or independent bidding, they appear on top google search results when Ad Rank is strong enough, which means every improvement to asset quality translates directly into more frequent and more prominent display, not into a separate traffic stream to optimize.

Link text: the most underestimated variable

Every Google Ads practitioner knows that link text is capped at 25 characters. What fewer internalize is that the 25-character ceiling is not a target, it is a hard wall, and the best link text usually lands well below it. The google ads sitelink extension best practices around link text are built on a single discipline: name the destination with the minimum words that are still specific enough to carry intent. The same economy that applies to ad copy in responsive search ads applies here, every word must earn its place, because there is no room for filler. When you follow sitelinks best practices consistently from the moment you create sitelink assets, link text becomes a precise entry-point label rather than a generic navigation item.

Generic labels are the most common mistake

In accounts I audit, the most common sitelink failure is not character overflow, it is semantic emptiness. Labels like "Products," "Services," "Learn More," and "About Us" appear in a substantial share of accounts, including ones with otherwise sophisticated campaign structures. These labels do not help users understand where the click will take them, and they do not help Google's expected impact model assess the asset's relevance to a specific query. The result is lower sitelink impression share than the account's Ad Rank would otherwise justify.

The replacement discipline is direct: take the generic label and replace it with the most specific thing a user arriving at that page would actually want. "Services" becomes "Enterprise API Plans." "About Us" becomes "Our Founding Team." "Products" becomes "Compare All Tiers." Each alternative is still under 25 characters, still usable as a compact row link, and carries an intent signal the generic version cannot. According to Google's official sitelink assets documentation, link text must be unique across sitelinks in the same asset set, which forces differentiation and, as a useful side effect, forces specificity.

The rendering argument for shorter text

Shorter link text is not only better for relevance, it is mechanically better for display frequency. A compact sitelink row on desktop fits four to six links side by side. Link text averaging 12–15 characters allows all four to show without truncation. Text averaging 22 characters on one or two sitelinks can force Google to drop a link from the row to preserve readability. Across a set of eight sitelinks, the ones with the most compact and specific labels tend to appear more often than longer alternatives. This is one of the most reliable patterns I observe, and it is consistent with the google ads sitelink extensions best practices 2026 guidance to prioritize specificity and brevity together rather than treating them as a trade-off. For the exact numbers behind the character cap, the dedicated page on google ads sitelink extension character limit covers every field in detail, including how double-width languages are handled differently.

Matching link text to query intent

The highest-impact application of this principle is intent matching at the ad-group level. When a campaign has ad groups targeting distinct intents, say, one for pricing queries and one for comparison queries, the sitelinks attached at the ad-group level should reflect those intents explicitly. The pricing ad group gets sitelinks like "Monthly vs Annual Plans" and "Request a Custom Quote", that last one doubling as a call to action that guides the user toward a concrete next step. The comparison ad group gets "Side-by-Side Feature Table" and "Why We Beat Competitors." Neither set would be appropriate at campaign level. This is the foundation of the asset hierarchy discussion below, but the copy discipline starts here, with the link text.

Description lines: unlocking the expanded format

The second layer of google ads sitelink extension best practices is description line copy. Each sitelink has two optional description lines, each capped at 35 characters. They are optional in the sense that Google will not reject a sitelink without them, but from a performance standpoint, leaving them empty is one of the most costly omissions in a Search campaign.

Why both lines must be filled

Filling only one description line does not unlock expanded sitelink format. Both must be present. The expanded format, where each sitelink shows its two description lines beneath the link text, stacking vertically, only renders at positions 1 or 2 and requires sufficient Ad Rank. But the eligibility gate is the presence of both description lines in the asset itself. Advertisers who fill one line and leave the second empty are permanently locked out of the expanded format, regardless of their bid or Quality Score. I treat this as a non-negotiable in every account I manage: if a sitelink is worth creating, both description lines are worth writing. Adding descriptions to every sitelink in the pool is the single highest-return action most accounts can take immediately, it costs nothing in media spend and directly unlocks a more visible format. This requirement is confirmed in Google's About assets documentation, which consolidates format eligibility requirements across all asset types.

What to write in 35 characters per line

Thirty-five characters is enough for one short declarative claim. The discipline is to treat description line 1 as the primary benefit and description line 2 as a secondary differentiator or risk-reducer. For a sitelink labeled "Enterprise API Plans": line 1 might read "From €499/month, SLA-backed" and line 2 "Dedicated onboarding included." Both lines add information the link text cannot carry, and neither restates what is already visible in the label. Restating the link text in the description is the second most common description mistake after leaving both lines empty, it passes the format gate but contributes nothing to expected impact. Notice also that the description should be written independently of any specific search term, it is attached to the asset, not to the query, so it needs to remain relevant across the full range of searches that can trigger the parent ad.

The google ads sitelink extensions best practices 2026 context makes this more important, not less: as Google auto-generates descriptions for sitelinks in dynamic mode when fields are absent, the bar for manually written descriptions is that they must be better than what the algorithm would produce on its own. Vague or restated copy rarely clears that bar. Write descriptions that contain information the algorithm cannot infer from the landing page alone, a price point, a guarantee, a specific feature count, a concrete differentiator. That is where manually written descriptions earn their place over dynamic fallbacks, and it is a large part of why sitelinks perform better when copy is written with intent, not generated as a placeholder.

The description-to-link-text relationship

One copy pattern I return to consistently: treat the link text and both description lines as a three-part unit. The link text names the destination; description line 1 answers "what will I find there?"; description line 2 answers "why should I prefer this over the obvious alternative?" A sitelink that satisfies all three questions in 25 + 35 + 35 characters is doing maximal work within the constraint. That three-part discipline is one of the google ads sitelink extension best practices I apply in every account at initial setup and at every copy refresh cycle.

Asset hierarchy: where most accounts leave performance on the table

The google ads sitelink extension best practices around asset hierarchy are the most operational of the four layers, and the most commonly ignored. Google resolves asset conflicts by specificity: ad-group-level assets override campaign-level assets for impressions in that ad group. Most accounts I audit use only campaign-level sitelinks, which means every ad group in the campaign shows the same set of sitelinks regardless of what the user searched for. Before thinking about copy, the first structural question is always the same: should these assets live at the campaign or ad group level? And when there are assets set at multiple levels across multiple campaigns, it is worth reviewing them at account level periodically to ensure the hierarchy is still coherent and that no outdated assets are suppressing more relevant ones.

When campaign-level is sufficient

If all ad groups in a campaign target the same intent, a single-product campaign where every keyword is a variant of the product name, campaign-level sitelinks are appropriate and operationally simpler. The ad groups are not meaningfully different in terms of what the user wants, so differentiated copy at the ad-group level would not outperform the campaign-level set. In these cases, the google ads sitelink extension best practices focus shifts entirely to copy quality: make the campaign-level sitelinks as specific and well-described as possible, since there is no lower level to override them. A useful signal that campaign level is sufficient: your ad groups share the same primary landing page or the same conversion action.

When ad-group level is necessary

As soon as intent varies between ad groups, pricing intent vs. feature intent vs. competitor comparison intent, ad-group-level sitelinks become the highest-leverage intervention available in the account. A user searching "google ads sitelink pricing" who sees a sitelink labeled "Pricing Plans" has a far more relevant experience than the same user seeing a generic "About Our Agency" sitelink inherited from campaign level. Expected impact, the auction factor that rewards assets for predicted click contribution, is sensitive to this relevance match. Intent-matched sitelinks at the ad-group level consistently produce higher impression share for those specific sitelinks than campaign-level fallbacks would. The policy framework governing what sitelinks can and cannot say is covered in the dedicated guide on policy requirements for sitelinks, worth reviewing before scaling ad-group-level assets, since policy violations are enforced at the asset level and can silently suppress individual sitelinks without surfacing a clear warning in the interface.

Pool size and rotation: how many sitelinks you actually need

The fourth layer of google ads sitelink extension best practices is pool management. Google selects which sitelinks to display in each auction from the available pool, optimizing for the combination most likely to drive clicks for that impression. A pool that is too small removes the algorithm's ability to differentiate between queries. A pool that is large but inconsistent in quality dilutes the average asset quality and can suppress the best performers. Both extremes hurt performance in different ways.

The minimum viable pool

Four sitelinks is the structural minimum for a four-link compact row. In practice, four sitelinks gives the algorithm no flexibility, the same four will always show, in whatever order Google chooses, with no rotation for learning. My working minimum is six sitelinks per asset set. When you add sitelinks beyond the structural minimum, you give Google the material to learn which combination performs best for which query signals, and that learning compounds over time. Six allows rotation in different combinations of four and begins that learning process. Eight to ten is the range where I see the most consistent impression share gains, because the algorithm has enough material to differentiate meaningfully between high-commercial-intent queries and broader informational queries within the same campaign. Beyond ten, pool quality becomes harder to maintain without introducing generic or low-intent assets that drag down overall expected impact scores.

Refreshing the pool without resetting performance data

One operational nuance worth knowing: editing the link text or destination URL of an existing sitelink resets its performance history in the account. Adding a new sitelink to the pool does not affect the history of existing assets. When a sitelink is underperforming and needs a copy revision, the cleaner approach is to create a new sitelink with the revised copy and pause the old one, rather than editing in place. This preserves the performance data on the original asset for reference and lets the new version accumulate its own history without inheriting a potentially distorted baseline. The google ads sitelink extensions best practices 2026 environment makes this more relevant because dynamic sitelinks are now competing for impression share alongside manual ones, keeping manual assets sharp and clearly differentiated is how you retain meaningful control over what users actually see.

Testing within the pool

Sitelink A/B testing in Google Ads is not native the way RSA headline testing is, you cannot pin sitelinks to positions and measure them in isolation. The practical alternative is to run two versions of the same conceptual sitelink simultaneously (e.g., "Request a Demo" and "Book a Free Call") with similar destinations, watch click share over 30–60 days, and pause the underperformer. This is a rough signal rather than a controlled experiment, because Google's selection algorithm already optimizes for expected performance, what you observe is the algorithm's verdict, not a clean head-to-head test. That limitation is worth knowing before drawing strong conclusions from a single round of testing. For deeper dynamic personalization beyond static copy, the guide on keyword insertion nelle sitelink extensions explains when dynamic keyword insertion adds value to the pool and when it creates more noise than signal.