Latest / The SEO Podcast with Fexingo: Search Engine Optimization, Google Rankings, and Organic Traffic / How Search Engines Score Page Speed in 2026
Transcript
- Lucas: So we need to talk about something that feels like it should be easy but keeps tripping up even the biggest brands: page speed. And I don't mean the old-school notion of just getting the HTML down the wire quickly. I mean the actual feeling of using the site. Luna: Right, because everyone says their site is fast until you try to click a button on mobile data and nothing happens for three seconds. Lucas: Exactly. It's that disconnect between the visual render and the interactive reality. For years, Google used to focus heavily on LCP, which is Largest Contentful Paint. Basically, when does the main hero image or headline appear? But back in early twenty twenty five, they started shifting the weight toward INP, Interaction to Next Paint. Luna: Which measures how long it takes for the browser to actually respond after a user clicks or taps? Lucas: Precisely. And now, with the September twenty twenty six updates rolling out across the board, that metric has become even more punitive. If your site looks great but feels sluggish, Google is essentially saying, 'We know you're hiding something behind that pretty animation.' It’s not just about aesthetics anymore; it’s about trust in the interface. Luna: That makes sense. Users bounce immediately if they can't interact. So how much does this actually move the needle on rankings compared to backlinks or content quality? Lucas: It’s a gatekeeper metric. Think of it like this: high-quality content gets you into the tournament, but slow page speed knocks you out in the first round. I looked at a recent dataset from a major e-commerce player last week. They had incredible product descriptions and strong domain authority, but their checkout flow was built on a legacy JavaScript framework that added over four hundred kilobytes of unused code. Luna: Four hundred kilobytes of junk? That sounds like a nightmare for mobile users. Lucas: It was a disaster. Their INP scores were sitting around eight hundred milliseconds on average. Google’s threshold for good performance is under two hundred milliseconds. So they were quadruple the limit. What happened next was fascinating. Their organic traffic didn't drop overnight. It decayed slowly over six weeks as competitors with cleaner code started stealing their featured snippets and top-three spots. Luna: So it wasn't a penalty per se, just a gradual loss of visibility because the user experience was inferior? Lucas: Spot on. It’s a subtle erosion. And here is the kicker: the brand didn't change their content strategy at all. They kept posting the same blogs and updating the same product pages. But because their technical foundation was rotting, Google’s crawlers started spending less time on their site. Crawl budget became a real issue. Luna: Crawl budget. That’s usually reserved for massive sites with millions of URLs. How does that affect a smaller business? Lucas: Even mid-sized businesses with fifty thousand pages are feeling this. If your site is slow, Googlebot gives up after crawling a few hundred pages because the server response times are too high. You end up with fresh content that never gets indexed. It’s a silent killer. I’ve seen marketing teams pour money into promoting articles that literally don’t exist in Google’s index yet because the site couldn’t handle the request volume. Luna: That’s terrifying. So what’s the fix? Do we just strip all the JavaScript and go back to plain HTML? Lucas: No, you don’t want to go full retro. The modern web requires interactivity. But you do need to audit your bundle size. The most common culprit these days is third-party scripts. You know, the chat widgets, the analytics trackers, the cookie consent managers. Every single one of those adds a layer of processing delay. Luna: Yeah, I noticed my favorite news site has like ten different pop-ups before I can even read the headline. Is that what you mean? Lucas: Exactly. Those pop-ups often block the main thread. While the browser is trying to show you the cookie banner, it can’t process your click on the article link. In twenty twenty six, Google’s algorithms are specifically looking for those 'main thread blockers.' If your script is holding up the UI for more than fifty milliseconds, you’re losing points. Luna: Fifty milliseconds. That’s practically instant for a human, but huge for an algorithm. Lucas: For the algorithm, yes. And here is the thing most agencies miss. They optimize for desktop. But ninety percent of search traffic is mobile. Mobile CPUs are weaker. Mobile networks are slower. A site that flies on Wi-Fi with a MacBook Pro might crawl to a halt on a phone with LTE. You have to test on real devices, not just simulators. Luna: Simulators lie. They always make things look faster than they really are. Lucas: They do. So I recommend running Lighthouse audits on actual phones. Not just the default scores, but the waterfalls. Look at where the time is going. Is it waiting for DNS? Is it waiting for the server? Or is it waiting for JavaScript to parse? Usually, it’s the parsing. You’ll see thousands of lines of code executing just to render a static paragraph. Luna: Why do developers write so much unnecessary code? Isn't there a standard for clean coding? Lucas: There is, but corporate inertia is strong. Marketing wants the flashy animated banner. Design wants the custom font that loads from three different CDNs. Engineering is told to ship features, not optimize performance. It becomes a tug of war. And SEO suffers in the middle. The best advice I can give is to make performance a requirement in the design phase, not an afterthought. Luna: Integrate it early. Otherwise, you’re just paying tech debt interest forever. Lucas: Precisely. Tech debt accumulates interest. And right now, the interest rate is your organic traffic. I saw a case study where a travel agency reduced their JavaScript payload by sixty percent. They switched from a heavy React app to a lighter Vue implementation for their search results page. Within two months, their conversion rate went up by fifteen percent. Not because the products changed, but because the friction disappeared. Luna: Sixty percent reduction? That must have taken some serious refactoring work. Lucas: It did. But the ROI was undeniable. Fifteen percent higher conversions from a technical fix is the holy grail. Most companies are chasing viral marketing trends while ignoring the fact that their website is fundamentally broken. It’s like putting a Ferrari engine in a shopping cart. Luna: A very expensive shopping cart. Okay, speaking of keeping things moving smoothly... Lucas: If these conversations about technical SEO and site architecture have helped you spot a bottleneck in your own workflow, remember that keeping this network ad-free relies on listener support. Buy me a coffee dot com slash fexingo helps us keep digging into these data-driven deep dives. Luna: Yeah, if you find value in cutting through the noise, that little gesture goes a long way toward sustaining the research we put into every episode. Lucas: Absolutely. It just helps us stay independent and focused on what actually works for your business. Luna: Back to the site speed issue though — do you think Google will eventually measure cumulative layout shift differently too? Lucas: I think CLS, Cumulative Layout Shift, is already quite mature. The issue is that developers are gaming it by adding empty height dividers to prevent text from jumping. Google knows this trick now. The latest updates penalize intentional layout stabilization if it’s done solely to manipulate the score rather than improve usability. Luna: So you can’t cheat the system anymore. You actually have to build better interfaces. Lucas: You have to. The era of quick hacks is over. If your site is fast, responsive, and honest, Google rewards you. If you’re trying to game the metrics, you’ll get caught. It’s a simpler rule set than people think. Build for the human, not the bot. The bot will follow. Luna: And if the human has a smooth experience, the revenue follows naturally. Lucas: Exactly. So check your INP scores this week. If they’re above two hundred milliseconds, you have homework. And if you fix them, you might just wake up to more traffic than you expected.