Editorial Team Generative search is completely changing how people find things, so advertisers have to rethink their paid media strategy just to stay visible. It’s not about simple keyword matching anymore. The whole game now is figuring out what a user actually wants when they ask a long, conversational question, and advertisers need to position their campaigns to grab attention inside this new search experience.
Key Takeaways
- You’ve got to shift budget to platforms with conversational AI and better semantic targeting if you want to show up in generative results.
- Your creative needs a total overhaul. It has to give direct answers and full info, getting ahead of the complex questions users are asking.
- Forget just tracking clicks. For gen search campaigns, you need to focus on engagement with the AI snippets and conversions coming from those detailed answer boxes.
- A winning strategy here uses your own first-party data to create hyper-personalized ads that go way beyond old-school broad audience segments.
- You have to constantly A/B test your ad copy and landing pages, treating them like prompts, to keep them relevant and capture that generative search traffic.
Campaign Teardown: “Future-Proofing Footwear” with Generative Search Paid Media
I was just running a campaign for StrideFlex, a direct-to-consumer athletic shoe brand, for their new line of sustainable running shoes. Our whole goal was getting visibility inside generative search environments because we knew people weren’t just typing “running shoes” anymore. They’re asking the AI stuff like, “What are the best running shoes for trail running with recycled materials?” or “Compare eco-friendly running shoes for marathon training.” We had to build a strategy that could handle that kind of detailed question.
Strategy: Anticipating Conversational Queries
Our whole strategy was built on guessing what complex, conversational questions people would ask and then answering them directly in our ads. That meant we had to go deeper than standard keyword research and get into natural language processing (NLP) patterns, focusing on long-tail, question-based keywords that we plugged into our paid search campaigns on Google Ads and Microsoft Advertising, especially the new AI-powered campaign types.
We set a budget of $150,000 over a 12-week duration. Our main KPIs weren’t just clicks. We were laser-focused on impression share within generative AI answer boxes and the click-through rate (CTR) on those big, expanded responses. We were aiming for a 15% bump in qualified leads and a 10% better return on ad spend (ROAS) than what we’d seen on past launches that used old-school paid search.
Creative Approach: Beyond the Headline
Getting the creative right was the toughest part. Catchy headlines and a basic description just don’t cut it. For generative search, your ad copy has to be super informative and directly answer a potential question right there in the ad. We ended up building out a ton of structured snippets and ad extensions that were designed to feed information straight to the AI models. So instead of “Buy StrideFlex Running Shoes,” the copy looked more like, “StrideFlex Trail Runners: Sustainable design with responsive cushioning for off-road performance. Made from 70% recycled content. Explore sizes and colorways.”
We also built out specific landing pages for each shoe model, structuring them with clear FAQs, comparison charts, and detailed product specs that an AI could easily scan using clear headings (we didn’t mess with full schema in the ad copy, but the landing page structure was key for the AI’s interpretation). The whole point was to hand the AI everything it needed to build a complete answer, with our ad sitting there as the source or the direct solution.
Targeting: Intent-Based and Contextual
Our targeting went way beyond simple demographics or interests. We focused almost entirely on intent signals we pulled from those conversational search patterns. In Google Ads, we built custom segments around people searching for things like “best eco-friendly running shoes,” “sustainable athletic footwear reviews,” and “vegan trail shoes.” We also used contextual targeting to show up next to articles and reviews about sustainability in sports. A huge chunk of our budget, around 40%, went straight to these more specific, intent-driven targeting methods.
Our negative keyword list was also bigger than usual because we were trying to filter out any vague queries that didn’t signal a real intent to buy our specific shoes. This really helped keep our Google Ads Quality Score high in the context of these new generative searches.
What Worked: Direct Answers and Enhanced Visibility
The results were pretty instant. Our ads started popping up in the “answer box” or “snapshot” in generative search results way more than we expected. Our average CTR on these generative placements hit 8.2%, which blew away the 3.5% we’d see on normal text ads in similar campaigns. It just showed that when you give users a direct, complete answer in the ad, they’re much more likely to click.
We also tracked a 17% lift in branded search queries for “StrideFlex” and specific shoe models during the campaign, which tells me that seeing the brand in the AI answers was building real recognition. The cost per lead (CPL) for conversions we could attribute back to these generative placements was only $28.50, a solid 15% cheaper than our typical paid search CPL of $33.50. It proved it’s just more efficient to meet people right when they’re asking detailed questions.
Campaign Performance Snapshot (12 Weeks)
| Metric | Traditional Paid Search (Pre-Campaign) | Generative Search Focused Campaign |
|---|---|---|
| Budget Allocation | N/A (Baseline) | $150,000 |
| Impressions | ~5.2M | 7.8M |
| Click-Through Rate (CTR) | 3.5% | 8.2% (Generative Placements) |
| Conversions | ~1,500 | 2,200 |
| Cost Per Lead (CPL) | $33.50 | $28.50 |
| Return on Ad Spend (ROAS) | 3.2x | 3.8x |
What Didn’t Work: Over-Optimizing for Single Keywords
At the start, we screwed up by trying to stuff too many old-school, short-tail keywords into our generative ad copy, and it just diluted the message and tanked our relevance scores. We learned that the AI models want ads that sound like a natural response to a complex question, not something packed with high-volume terms. Our first few attempts to just “generative-fy” old ad groups by adding more keywords was a total failure. It’s a completely different animal, and you need a new playbook for creating ads.
Attribution was another headache. A user might see our ad in an AI answer, but then search for the brand directly later on. User journeys have a lot of touchpoints. To get a real sense of generative search’s impact, we had to build out a more advanced multi-touch attribution model and stop relying on last-click.
Optimization Steps Taken: Iterative Refinement
So, we made a few key changes as the campaign went on:
- Prompt Engineering for Ad Copy: We basically started treating ad copy creation like we were engineering prompts for an AI. We A/B tested different sentence structures and how much detail to include. For example, we tested “StrideFlex: The eco-friendly trail shoe” against “Looking for sustainable trail running shoes? StrideFlex offers superior grip and recycled materials.” That second one, the question-and-answer format, performed way better in generative results.
- Landing Page Refinements: We looked at the data from users coming from gen search and tweaked our landing pages. We added more detailed comparison tables showing material breakdowns and sustainability certs, directly answering the deeper questions people were asking. We also made sure the main reasons to buy were right there above the fold, answering the “why us?” question immediately.
- Budget Reallocation to AI-Driven Campaigns: We kept shifting more of the budget (another 15%) into Google’s Performance Max campaigns and Microsoft’s Smart Campaigns. These campaign types are built to use AI for targeting and creative, so it let the platform algorithms do the heavy lifting of finding and capitalizing on new generative opportunities for us.
- First-Party Data Integration: We plugged our customer data platform (CDP) into the ad platforms so we could build super-specific audiences. For instance, we could target people who had bought sustainable stuff from us before or read our environmental blog posts. This allowed us to get even more personal with the ad copy that was showing up in the generative answers.
By the end of the 12 weeks, the campaign beat our initial ROAS goal, hitting 3.8x. The cost per conversion ended up at $68.18, which was a nice improvement over our baseline. And our impressions shot up by over 50% compared to previous similar campaigns, which proved we were getting much better visibility in these new search interfaces.
Getting this stuff right is a present-day reality, not something you can put off. The advertisers who are already building detailed, answer-first content and using smart targeting are going to get a huge lead. If you ignore it, you’re just handing your market share to competitors who get that search is now a conversation, not just a query box. For more on how AI is changing PPC, check out these insights on how AI in PPC reveals ROAS gains.
What is generative search optimization?
It means tweaking your content and paid media so they show up and look good inside the AI-generated answers and summaries on a search results page. You’re creating complete, answer-focused content that an AI can easily grab, synthesize, and present to a user.
How does paid media contribute to visibility in generative search?
Paid media lets you strategically insert your brand, products, or content right into or next to the AI’s answer. You could show up as a cited source, get your product recommended in an AI summary, or just have a traditional-looking ad that’s super relevant to the user’s conversational query.
What kind of ad creative works best for generative search?
The best ad creative is really informative and feels like a complete answer. It should provide direct solutions or detailed info that gets ahead of what the user is asking. Things like structured snippets, rich ad extensions, and landing pages packed with FAQs and specs work really well.
Are traditional keywords still relevant for generative search paid campaigns?
Yes, but how you use them has changed. While short-tail keywords still define a broad topic, winning in generative search is all about understanding the user’s intent behind their long, conversational questions. You should be focusing on those question-based keywords and natural language patterns to build your targeting and ad copy.
How do you measure success for paid media in generative search?
You have to look past the old metrics. For generative search, your main KPIs should be things like your impression share inside the AI answer boxes, how many people are engaging with those expanded AI responses, and the CTR on your ads when they appear as a direct answer. Cost per lead (CPL) and return on ad spend (ROAS) are still important, but you need a multi-touch attribution model that can actually account for these new generative interactions.
