Responsive search ads (RSAs) are Google's standard search ad format in 2026, allowing advertisers to provide up to 15 headlines and 4 descriptions that Google's AI assembles into the best-performing combination for each auction. Writing high-performing RSA headlines, optimizing ad strength, and knowing when to let automation handle the rest is the difference between ads that convert and ads that drain budget.
This guide covers everything you need to know about responsive search ad optimization in 2026: headline strategy, description best practices, what ad strength actually means for performance, and how autonomous management removes the manual grind of ongoing RSA testing.
Google fully sunset expanded text ads (ETAs) for new creation back in 2022, and by 2026, RSAs are the sole search ad format available for creation across all standard Search campaigns. If you are running Google Ads, you are running RSAs. There is no alternative format to fall back on.
This means your ability to write strong headlines, structure descriptions correctly, and manage the interplay between Google's ad assembly AI and your own strategic intent is not optional. It is foundational to every dollar of Search spend you run.
Google's responsive search ad system works by combining your provided headlines and descriptions into different permutations at auction time. You supply up to 15 headlines and 4 descriptions. Google selects which combination to show based on the search query, user context, device, and historical performance signals.
The key detail most advertisers miss: Google does not test every possible combination equally. Its machine learning model quickly converges on a subset of combinations it predicts will perform best, which means low-quality headlines may never get meaningful impressions. Writing 15 headlines is not a checkbox exercise. Every headline needs to be strong enough to earn its place in the rotation.
With ETAs, you controlled exactly which three headlines appeared and in which order. With RSAs, you cede that control to Google's AI. The tradeoff is reach and relevance: Google can adapt your ad to match different queries and contexts. The cost is that weak copy gets amplified if you are not deliberate about what you provide.
The practical shift for advertisers in 2026 is this: your job is no longer writing a single perfect ad. Your job is creating a library of high-quality components that work in any combination. That requires a fundamentally different approach to ad copywriting.
Every RSA headline is capped at 30 characters. This is a hard limit, and it forces precision. The best-performing headlines typically follow one of these patterns:
Direct keyword match. Include the core search term as closely as possible. "Affordable CRM Software" beats "We Have Great Software."
Specific value proposition. Lead with a number or concrete benefit. "Save 40% On Shipping Costs" is stronger than "Lower Your Costs."
Social proof or trust signals. "Trusted By 10,000+ Teams" or "4.9 Stars On G2" fits within the character limit and adds credibility.
Clear CTA. "Get A Free Quote Today" or "Start Your Trial Now" works as a headline, not just a description element.
The mistake most advertisers make is writing headlines that are too similar to each other. Google's system penalizes redundancy by limiting how many near-duplicate headlines it shows, which means five headlines all saying variations of "Best PPC Agency" waste your slots.
Keyword insertion using the {KeyWord:Default} function still works in RSAs and remains a strong tactic for improving expected click-through rate. However, in 2026, the smarter play is combining dynamic keyword insertion in one or two headline slots with static headlines that carry your unique value proposition.
A practical framework: use keyword insertion in headlines 1 and 2 (which appear most prominently), then fill the remaining slots with benefit-driven, differentiated copy. This ensures relevance to the query while giving Google enough variety to optimize combinations.
One important note: ad strength and Quality Score are related but distinct. Keyword relevance in headlines directly influences expected CTR, which is a core Quality Score component. Writing headlines that match search intent is not just about ad strength. It directly affects what you pay per click.
Pinning forces a specific headline to always appear in a designated position (headline 1, 2, or 3). Google recommends against excessive pinning because it reduces the number of combinations available for testing. But the recommendation to "never pin" is overly simplistic.
Pin when: You have a legal disclaimer that must appear. Your brand name needs to show in headline 1 for trademark or consistency reasons. A specific CTA must always be visible.
Don't pin when: You are trying to maximize Google's ability to find the best-performing combination. Your headlines are already diverse enough that any combination reads well.
The middle ground: Pin two or three headlines to position 1 (not just one). This ensures one of your preferred headlines always appears first while still giving Google flexibility. The same logic applies to position 2 and 3.
Pinning a single headline to a single position reduces ad strength and limits optimization potential. Pinning multiple headlines to the same position gives you control without sacrificing too much flexibility.
Google gives you 15 headline slots. Use all of them, but use them strategically. Here is a practical framework for filling all 15:
Slots 1 through 3: Primary keyword variations. These should closely match your top search terms.
Slots 4 through 6: Value propositions. What makes your offer different? Price, speed, quality, guarantee.
Slots 7 through 9: Social proof and trust signals. Customer counts, ratings, awards, years in business.
Slots 10 through 12: CTAs and urgency. "Get Started Free," "Limited Time Offer," "Book A Demo Today."
Slots 13 through 15: Differentiators and objection handlers. "No Long-Term Contracts," "Cancel Anytime," "Free Shipping."
This framework ensures that no matter which combination Google assembles, the resulting ad contains a keyword match, a reason to believe, and a reason to act. That is the structural foundation of a high-performing RSA.
When you work with groas, your dedicated account manager builds this kind of structured headline library for every ad group, informed by competitive analysis and performance data. The AI agents then continuously monitor which headline combinations are winning and adjust creative direction accordingly, something that would take a human team hours each week to do manually.
RSA descriptions have a 90-character limit per slot, and you get 4 slots. The critical principle: each description must read well alongside any headline combination. This means descriptions should not repeat headline content verbatim.
Think of descriptions as the "supporting evidence" layer. Headlines grab attention and state the proposition. Descriptions expand on why the proposition is credible and what the user should do next.
The most effective RSA descriptions in 2026 follow a consistent structure:
Description 1: Expand on the primary value proposition. Include a specific benefit or outcome.
Description 2: Add a trust signal or social proof element.
Description 3: Present the offer clearly, including any pricing, discounts, or free trial details.
Description 4: Close with a strong CTA and an objection handler.
For example, if you are advertising a B2B SaaS product, your descriptions might be: "Reduce onboarding time by 60% with automated workflows. Trusted by 2,000+ companies." and "Start free for 14 days. No credit card required. See results in your first week."
With 90 characters, you have more room than in headlines but not unlimited space. Prioritize specificity over cleverness. "Save 15 hours per week on reporting" beats "Work smarter, not harder" every time. Google's AI responds to concrete language because users respond to it, and CTR data reinforces that signal.
Ad strength is Google's assessment of your RSA's component diversity and relevance. It ranges from "Poor" to "Excellent." Here is what many advertisers get wrong: ad strength is not a performance metric. It is a readiness score.
An "Excellent" ad strength rating means Google believes your RSA has enough diverse, relevant components to test effectively. It does not mean your ad will convert better than one rated "Good." Many accounts have "Good" rated RSAs that outperform "Excellent" ones because the copy itself is sharper and more aligned with user intent.
Instead of chasing ad strength, focus on these metrics:
Click-through rate (CTR). The most direct measure of whether your headlines and descriptions resonate.
Conversion rate. Strong RSAs drive not just clicks but qualified clicks. If CTR is high but conversion rate is low, your ad is attracting the wrong audience.
Cost per conversion. The ultimate measure. An RSA rated "Good" that delivers conversions at half the cost of an "Excellent" rated RSA is the better ad.
Impression share. If your ad strength is limiting your impression share (which it can, as Google uses it as a minor factor in ad rank), that is worth addressing. Otherwise, focus on outcomes.
Google's ad strength recommendations are directional, not prescriptive. When Google tells you to "add more unique headlines," it is generally good advice because more variety gives the system more to work with. But when Google recommends headlines that dilute your messaging or add generic phrases like "Learn More," you should push back.
The best approach: get to "Good" or "Excellent" ad strength by following the 15-slot framework above, then optimize purely based on performance data. Do not add weak headlines just to appease the ad strength meter.
A thorough audit of your account will often reveal RSAs running with only 5 or 6 headlines, which limits Google's ability to optimize and can suppress impression share. Filling all 15 slots with quality headlines is one of the highest-leverage fixes you can make.
Testing RSAs is fundamentally different from testing ETAs. With ETAs, you ran two or three ad variations side by side and picked the winner. With RSAs, the testing happens within the ad itself, across headline and description combinations.
To run a meaningful RSA test, you need to change the component library meaningfully. Swap out 5 or more headlines, change the angle of your descriptions, or test a completely different value proposition. Small tweaks to individual headlines are unlikely to produce statistically significant results because Google's internal optimization absorbs minor variations.
Give each RSA iteration at least 2 to 4 weeks and a few hundred clicks before drawing conclusions. Shorter test windows produce unreliable data.
Google assigns performance labels to individual headlines and descriptions: "Best," "Good," "Low," and "Learning." These labels are based on relative performance within the RSA.
The practical rule: replace "Low" performing headlines with new variations and monitor "Learning" headlines for at least two weeks before judging. Keep "Best" headlines in rotation and write new headlines that explore adjacent angles.
Asset performance labels are one of the most underused feedback mechanisms in Google Ads. They tell you exactly which components are dragging down your RSA and which are carrying the weight.
Stale ad copy leads to creative fatigue, even in Search. A reasonable refresh cadence for most accounts is every 4 to 8 weeks, replacing underperforming headlines and descriptions while keeping top performers.
The challenge is that most teams do not have time to refresh copy across dozens or hundreds of ad groups on a consistent schedule. This is where the gap between what should happen and what actually happens in most accounts is widest.
The biggest bottleneck in RSA optimization is not strategy. It is execution. Most advertisers know they should be refreshing headlines, pruning low performers, and testing new angles. They just do not have the bandwidth to do it consistently across every ad group.
This is exactly the problem groas solves. As a full-service Google Ads management service, groas combines AI agents that run 24/7 with a dedicated human account manager who oversees creative strategy. The AI agents continuously monitor asset performance labels across every RSA in your account, identify underperforming headlines and descriptions, and flag opportunities for creative iteration. Your dedicated account manager reviews these signals, makes strategic decisions about messaging direction, and ensures every change aligns with your business goals.
Unlike an agency where junior account managers cycle through your account, groas gives you a single point of accountability paired with AI that never stops monitoring. The result is RSA performance that compounds over time because optimization never pauses.
Most teams treat RSA optimization as a periodic task: check the account once a week, swap a few headlines, move on. groas treats it as a continuous process.
Here is the difference in practice. A typical freelancer might review your RSA performance labels once a week and make changes when they remember. An agency might batch creative updates monthly, if the junior account manager assigned to your account has bandwidth. groas AI agents evaluate headline and description performance daily, surface recommendations to your dedicated account manager, and implement changes as soon as the data supports action.
The dedicated human account manager brings something no AI can replicate on its own: strategic judgment about your brand voice, competitive positioning, and business context. They ensure that new headline variations are not just statistically optimized but strategically sound. This AI plus human combination is what makes groas fundamentally different from both traditional agencies and self-serve optimization tools.
When you compare this to managing RSAs yourself or relying on a stretched-thin in-house team, the math is clear. groas costs a fraction of a single in-house salary, delivers senior-level strategic oversight, and provides 24/7 AI execution that keeps every RSA in your account performing at its peak.
Responsive search ad optimization in 2026 comes down to two things: providing Google's AI with the best possible components, and ensuring those components are continuously tested and improved.
The first part, writing strong headlines and descriptions, requires strategic thinking and an understanding of your audience. The second part, ongoing testing and iteration, requires consistent execution at scale.
If you are spending meaningful budget on Google Ads and your RSAs are not being actively managed on a daily basis, you are leaving performance on the table. groas eliminates the gap between what should be done and what actually gets done, combining always-on AI agents with a dedicated human account manager who owns your strategy. See how groas compares to your current agency, freelancer, or in-house approach, and decide for yourself whether your current setup is actually keeping up.
You should fill all 15 headline slots that Google provides. Each headline should be unique and serve a distinct purpose: keyword relevance, value proposition, social proof, CTA, or objection handling. Leaving slots empty limits Google's ability to find winning combinations and can suppress your impression share. Use the 15-slot framework outlined above to ensure every headline earns its place.
Ad strength is a readiness score, not a performance metric. It measures how diverse and relevant your RSA components are, but an "Excellent" rating does not guarantee better conversions or lower costs. Many accounts see stronger results from RSAs rated "Good" because the copy is more tightly aligned with user intent. Focus on click-through rate, conversion rate, and cost per conversion rather than chasing the ad strength meter.
Pinning is useful when you need to guarantee that a specific message (like a brand name, legal disclaimer, or primary CTA) always appears in a certain position. The best practice is to pin two or three headlines to the same position rather than pinning a single headline, which gives you control without eliminating Google's optimization flexibility. Avoid pinning all positions, as this essentially recreates the rigidity of expanded text ads.
A reasonable refresh cadence is every 4 to 8 weeks. Replace headlines and descriptions labeled "Low" by Google's asset performance ratings, keep those labeled "Best," and give "Learning" assets at least two weeks before making a judgment. Consistent refreshing prevents creative fatigue and keeps your ads competitive.
Yes, the {KeyWord:Default} function still works in RSA headlines and remains an effective tactic for improving expected click-through rate. The recommended approach is to use keyword insertion in one or two headline slots (ideally positions 1 and 2) while filling the remaining slots with benefit-driven, differentiated copy that communicates your unique value.
Meaningful RSA tests require changing the component library significantly, not just tweaking a single headline. Swap out at least 5 headlines or shift the core messaging angle in your descriptions. Allow 2 to 4 weeks and several hundred clicks before drawing conclusions. Testing within the RSA (by refreshing components) is more effective than running multiple RSAs in the same ad group, which splits data.
groas is a full-service Google Ads management service that pairs AI agents running 24/7 with a dedicated human account manager. Unlike a freelancer who checks your account a few times a week or an agency that batches creative updates monthly, groas AI agents evaluate headline and description performance daily, flag underperformers, and implement changes as soon as the data supports action. Your dedicated account manager ensures every creative decision aligns with your brand voice and business strategy. This continuous optimization cycle means your RSAs improve every day, not just when someone remembers to log in.
Managing RSA copy testing across multiple ad groups is time-intensive and easy to neglect. Rather than hiring a dedicated resource or adding it to an already-stretched team's workload, groas handles it as part of its full-service Google Ads management. You get continuous creative iteration powered by AI agents plus strategic oversight from a dedicated human account manager, at a fraction of the cost of a single in-house hire. This ensures your RSAs are always optimized without adding headcount or management overhead.
