Ranking first used to mean you won. You’d check your position, see "#1," and know your SEO work paid off.

That’s over.

Today, ranking first might put you fourth on the page. Or fifth. It depends on how many ads run above you, how long the AI overview stretches, and whether there’s a local pack pushing you down.

"I think it’s not anymore like important just that you are first but where exactly on that page you are," says Milos Djurdjevic, VP of Engineering and Product at SerpApi. His team has been building rank tracking tools for years, and they’ve watched this shift happen in real time.

The question isn’t "where do I rank?" anymore. It’s "where am I actually visible on the screen?"

The Three Metrics That Define Visibility Now

Traditional rank position still matters. When SerpApi loses a top position, they see it in their traffic numbers. But it’s not the complete picture.

You need to track three things:

Traditional rank position: Your numerical placement in organic results

Pixel position: How many pixels from the top of the page your result appears

AI citations: Whether and where you’re referenced in AI-generated answers

Each metric tells you something different. Traditional rank tells you how search engines value your content. Pixel position tells you if users will actually see it. AI citations tell you if you’re making it into how people actually search now.

Why Pixel Position Matters More Than Rank

Milos uses a restaurant analogy: "A restaurant can rearrange its layout to make shopping more efficient or to encourage customers to browse longer."

Search engines do the same thing, but they change constantly. They test features, add AI overviews, insert more ads, and shift elements based on the query.

You can rank first and still be below the fold. It happens all the time. A site holds the #1 organic position, but users scroll past three ads, an AI overview, and a local pack to reach it. That’s not a #1 position.

Pixel position tracking solves this. Instead of just knowing your rank number, you know exactly how many pixels down the page your result appears. You can measure this in real time, for any location, and compare it across queries.

The data confirms what you’d expect: pixel position correlates strongly with click-through rates. A result at pixel 800 performs differently than one at pixel 1600, even if both are technically ranked #1.

This matters for locations and languages too. The same query in different cities produces different layouts. Add multiple languages, and you’re tracking dozens of variations for a single keyword. Real-time pixel position data lets you see exactly what users see in each scenario.

The Technical Reality of Real-Time Tracking

All position tracking needs to happen in real time because search results change constantly.

"These search engines, they change layout sometimes, you know, on weekly basis," Milos explains. They introduce features, pull them back, test variations, and iterate based on user behavior. An AI overview might appear for a query on Monday, disappear on Tuesday, and return in a different format on Wednesday.

This creates a challenge for anyone trying to maintain accurate data. You can’t scrape once a day and call it current. You can’t cache results and assume they’ll stay valid. Every query needs fresh data because the landscape shifts continuously.

The scale compounds the difficulty. Tracking one keyword for one location is straightforward. Tracking thousands of keywords across multiple locations, languages, and devices requires infrastructure that can handle constant, high-volume requests while adapting to layout changes in near real time.

AI Citations: The New Ranking Frontier

More users are moving from traditional search to chat interfaces. Instead of clicking through results, they’re getting answers directly. This shift makes AI citations critical.

Being cited in an AI-generated answer is the new visibility metric. It’s not enough to rank well in traditional search if users never see the search results page. You need to appear in the AI’s response.

SerpApi is seeing more requests for tracking citations in various AI engines. Users want to know: Am I being mentioned? Where in the response? How often? For which queries?

The challenge is that each AI system works differently. ChatGPT structures its responses one way, Perplexity does it another, and search-integrated AI overviews follow their own patterns. Tracking citation position across all these systems requires monitoring multiple interfaces simultaneously.

This is where rank tracking splits into two worlds. Traditional search ranking still matters for users who browse results. AI citation tracking matters for users who want direct answers. Most businesses need to track both because their audience uses both approaches.

What Search Engines Are Optimizing For

Search engines operate on feedback loops. They test changes, measure user behavior, adapt based on the data, and repeat.

"It just feels that Google operates that way," Milos says. "It has been that way like for years, even before the AI boom. Constant changes in layouts, constant changes in design and UI."

AI has accelerated this cycle. What used to change monthly now changes weekly or daily. New features appear, perform well or poorly, and either stick around or disappear. The whole system feels algorithmic rather than human-directed, like a massive A/B testing engine running continuously.

This makes prediction difficult. You can see current patterns, but projecting what comes next is speculation.

The practical impact: if you’re tracking rankings or building tools that depend on search result structure, you’re constantly adapting. Layout changes break parsers. New features require new data models. What worked last month might not work today.

The SEO Migration to AI

SEO is moving from search engines to chatbots. That’s the direction the data shows. More queries go to AI interfaces, fewer to traditional search pages.

"I feel that the SEO will indeed move to the AI, to the chatbots, away from search engines," Milos says. "That’s how it looks right now."

This changes what "ranking" means. In traditional search, ranking is a position in a list. In AI responses, "ranking" might mean being the primary source cited, or being mentioned at all, or being linked in the references. The concept needs redefinition.

For businesses, this means tracking visibility across both paradigms. You still need traditional rank data because search isn’t going away immediately. But you also need AI citation data because that’s where the trajectory points.

The challenge for rank tracking tools is supporting both modes without overwhelming users with data. Users need to know: "Where am I visible?" The answer now requires multiple data points across different systems.

What This Means for Your Strategy

If you’re still optimizing solely for traditional rank positions, you’re measuring an incomplete picture.

Track three things:

Traditional rank position: Still matters for understanding search engine valuation and for queries that haven’t shifted to AI answers yet 1.

Pixel position: Shows actual visibility accounting for all the elements pushing you down the page 1.

AI citations: Indicates visibility in the growing segment of users who skip search results entirely

Each metric serves a different strategic purpose. Traditional rank guides your SEO fundamentals. Pixel position tells you if your visibility matches your rank. AI citations show if you’re prepared for where search is heading.

The Infrastructure Challenge

Building rank tracking infrastructure at scale isn’t viable for most organizations anymore. It might have been possible a few years ago with simpler search layouts and slower change cycles. Today, the technical and operational overhead makes it impractical.

The infrastructure needs to handle:

• Real-time data collection across hundreds of search variations
• Constant adaptation to layout changes
• Parsing complex, dynamic content including AI-generated elements
• Scaling to thousands of queries per client
• Maintaining accuracy across different locations, languages, and devices

Each of these challenges alone is solvable. Combined at scale, they require dedicated teams, significant engineering resources, and continuous maintenance. For most businesses, it’s more cost-effective to use specialized tools like a SERP tracking API than to build and maintain the infrastructure internally.

Where We’re Headed

The rank tracking industry won’t stabilize in 2026.

"I don’t see like that normalcy happening too soon," Milos says. "Like I think that this will continue in next year. I feel that all like all the prominent companies into AI are still exploring things, still kind of searching and in a way fighting between each other to or who is going to be on top."

AI companies are still exploring how to present information, search engines are still testing layouts, and users are still adapting their behavior. This transition period will continue.

What’s clear: the definition of "ranking well" is expanding. It now includes traditional position, actual visibility (pixel position), and presence in AI-generated content. Tracking one metric without the others leaves you blind to parts of your actual visibility.

The shift from search engines to chatbots will continue. This doesn’t mean traditional search disappears, but the weight shifts. More queries get answered by AI, fewer click through to websites. Businesses need visibility in both channels.

For anyone serious about search visibility, the strategy is straightforward: track all three metrics, understand what each tells you, and optimize for both traditional search and AI citations. The tools and techniques that worked five years ago aren’t sufficient anymore. Visibility requires a more complete view.

The technical challenges of tracking these metrics at scale aren’t getting easier. They’re getting harder as search engines and AI systems evolve faster. This makes specialized tracking tools more valuable, not less.

If you’re still making decisions based solely on traditional rank positions, you’re working with incomplete data. The question isn’t whether to adapt to these new metrics, it’s how quickly you can implement tracking that covers the full picture of your visibility.