What is GEO and how does it differ from SEO?

GEO (Generative Engine Optimization) refers to optimizing content so it gets cited by AI-powered search engines like ChatGPT, Google AI Overviews, or Perplexity. Traditional SEO focuses on ranking in classic Google results. The two are not mutually exclusive but require different strategies.

How do I check if my Google Ads ROAS is real?

Compare the conversion values reported in Google Ads with actual revenue from your CRM or accounting system. If there are large discrepancies, check your Customer Lifecycle Goals settings and whether conversion values are being automatically modified by the platform.

Is AI-generated video worth investing in now?

For hooks and short social media clips, yes. For premium brand content, not yet. The maximum 720p resolution works for social platforms, but quality varies and the risk of visual sameness is high without a solid creative brief.

GEO, AI Search and ROAS: What Matters in 2026

Four developments from the past week that deserve attention. Not because any single one is dramatic, but because together they paint a clear picture: digital marketing is moving faster than the tools we use to measure it.

GEO now outpaces SEO in freelance demand. AI search data sources have trust tiers. Google Ads ROAS can be artificially inflated without you noticing. And Google Gemini Omni creates video clips with your face, no camera or editor required.

What do they share? All four ask marketers to check what they measure, not just how much.

GEO demand has outpaced traditional SEO

A Fiverr survey published this week reveals a shift many anticipated but few had quantified: 67% of search freelancers report increased Generative Engine Optimization (GEO) requests over the past year. By comparison, only 56% saw growth in traditional Google ranking tasks.

But the numbers tell only part of the story. 61% of those same specialists still consider classic SEO the strongest long-term investment. The market wants GEO, but practitioners still bet on Google. That gap matters.

What this means for businesses: GEO demand comes from clients who have heard about AI search and want to be present. It is reactive demand, not strategic. The difference is significant: those who can turn reactive demand into a coherent plan win projects. Those who simply add "GEO" to their service deck spend money on execution without direction.

Another detail worth noting: 61% of freelancers say clients expect lower prices because of AI, with typical projects running between $150 and $300. Yet 83% of specialists report being hired more often for strategic advice than for execution. Value is migrating from "do my SEO" to "tell me what to do." Much like the broader debate around AI visibility and publisher compensation, what matters is not whether you are visible but in what context.

(Source: Search Engine Journal)

AI search has a data source hierarchy

A recent article on Search Engine Journal classifies AI search data sources into four tiers. Tier 1: confirmed and active, meaning Google Search and Google Maps through real-time grounding technology. Tier 4: based on strong evidence but not officially confirmed. In between, a range of sources with varying degrees of certainty.

The practical takeaway: product feeds (Google Merchant Center, OpenAI commerce specifications) directly power AI shopping recommendations. If you sell products online and your feed is not current, you do not exist in purchase conversations generated by ChatGPT or Google AI Overviews. This is not a long-term visibility issue. It is an immediate presence problem.

Reddit illustrates the volatility. Google pays $60 million annually for access to Reddit content, but the contract may not be renewed. A data source many SEO professionals considered stable is actually fragile. For local businesses, direct partnerships between AI platforms and Yelp (including real-time bookings) point the direction: local data matters more than ever. The equivalent for businesses outside the US: treat your Google Business Profile with the same seriousness as a landing page.

The recommendation: do not try to optimize for every source at once. Start with what is confirmed (Google Search, Maps, product feeds), monitor what is probable, and ignore what is speculative. As we discussed in the context of how advertising works inside AI platforms, where you show up matters as much as how often.

(Source: Search Engine Journal)

Google Ads ROAS looks good. Maybe too good.

Google Ads launched Customer Lifecycle Goals, a set of features that separate new customers from returning ones at the bidding and targeting level. The concept is sound: treat a first-time buyer differently from a repeat customer. The implementation raises serious questions.

A detailed article on Search Engine Journal identifies the core problem: some settings artificially inflate conversion values. The reported ROAS looks better than actual business performance. The author calls it "ROAS inflation," and it is hard to detect without checking numbers outside the platform.

Common errors: misconfigured customer lists, confusion between targeting and optimization in Performance Max (which only optimizes, it does not target), and duplicate campaigns that distort analysis. The simple rule proposed: if your customer list represents less than 1% of your total target population, you do not need Customer Lifecycle Goals. Basic audience exclusions will do.

What matters for marketing managers and business owners: ask where the ROAS numbers in your reports actually come from. A 5:1 ROAS that includes automatically adjusted conversion values is not the same as a 5:1 ROAS calculated on real revenue from your CRM or accounting system. The difference can be significant, and budget decisions based on inflated numbers are expensive decisions.

(Source: Search Engine Journal)

Gemini makes video with your face. Production gets democratized.

Google Gemini Omni now allows users to create a digital avatar from a 5-minute facial capture. The avatar can be placed in any scenario described in text: presentations, social media hooks, video inserts. Clips are capped at 10 seconds but can be concatenated for longer content.

The immediate practical application: 2-3 second hooks for short-form video. Instead of filming 15 intro variations to test what works on each platform, you generate them instantly. The prompting formula combines four elements: subject, action, environment, and camera direction, all in natural language.

Beyond generation, Gemini can edit existing footage without a human editor: change the environment, add motion graphics, remove objects from the frame. The 720p resolution works for social media but not for a premium website or brand production. That is a real limitation, not a niche one.

The risk? It is the same one we flagged with AI-generated text content: if everyone uses the same tools with the same prompts, the result is visual uniformity. The production barrier drops, but differentiation grows in importance. The tool is not the strategy. The brief is.

For marketing teams with limited production budgets, Gemini Omni is a real option for social media content. But it does not replace a content strategy. It only accelerates the execution of the strategy you already have.

(Source: Social Media Examiner)