What Happened in SEO in 2024? Year-End Review

In-depth retrospective covering Google core updates, helpful content changes, and generative search rollouts in 2024.

Speakers: Samet Özsüleyman, Can Mutioğlu

17 Dec 2024

Video Transcript

⏱ 42 min read• 7,486 words

Read the full speech or switch languages

— Hey everyone, welcome to Zeo Talks. Today Can and I are covering 2024 SEO trends. We have our topics ready. Can, I’ll hand it over to you first. Let’s talk about Google updates. We actually hosted a meetup on this together. I’m curious about your final thoughts. We’ve seen so many core updates lately. Plus, there were several spam updates. Google pushed these updates so fast. It’s been tough to analyze the impact of each one. Many SEO experts are reporting the same struggle.

What’s your take on this? Especially regarding the core and spam updates?

— Hey Samet. Google really went all out with updates late this year. Like you said, tracking which update caused what was a nightmare. It was especially hard for large websites. Sites that dropped during the November-December updates are struggling. These updates hit many sectors hard this year. They impacted authoritative insurance sites. News sites are also constantly being hit by these updates. Updates after the second half of the year really shook news sites. We talked about this personally. Local and smaller news sites gained value. They started ranking better on Google. Big news sites lost authority. They lost traffic along with that authority.

This update was tough for news sites. I haven't had time to analyze the latest one yet. But I see a slight recovery for news sites. People are saying a lot about the December update. Some call it a continuation of November. Looking at the results, I disagree. It was an interesting process. Google now rolls out updates without notice. Before, they used to signal everything beforehand. Every SEO tool showed update signals. We tracked them closely on Semrush. Even without an announcement, the update shouted. "I'm coming." It was a year full of Google updates.

The summer update didn't seem to have much impact. If I recall correctly, most sites saw nothing. Updates vary by sector, of course. Maybe some small sites felt it and we missed it. From what I read, there wasn't a massive update process there. There are always annual spam updates, though. Sites hit by spam updates are easier to fix. I mean, compared to those hit by core updates. You can take quick action on sites hit by spam updates. You can handle backlink or off-page fixes fast. But sites hit by core updates struggle to recover. We saw this clearly in the news sector this year.

— Exactly. I’ve noticed this with my own test sites too. Looking at them, I see similar patterns. They often mirror the growth or trends from the November 2023 update. A stable site with no new articles or SEO work suddenly spikes. It shows the core update fixed something. Maybe some sites don't need constant updates. Take a constitutional law article, for example. It stays the same for years. Maybe the update addressed that. We already covered spam and these updates earlier. We should talk about the new 2024 spam policies. Like expired domain abuse. Think of a charity site from the 90s that expired.

Bad actors exploit these expired domains for junk content. Or site reputation abuse. That was definitely the biggest focus area. Google warned sites back in May. They set a deadline for site reputation abuse. They threatened manual or algorithmic penalties after that date. It’s about your site's main context. Say you’re pumping out coupon codes. Just having coupon content doesn't make you a bad site. Google cares if you’re leveraging site reputation to rank higher. That’s the real issue. Defining spam is subjective, Can. My definition differs from yours.

Google’s view on spam is different too. Be careful before labeling a page as spam. Don't just think about backlink spam. Content-based reputation abuse is real. We’ve had a wild update cycle. Every analysis we ran overlapped with another update. These updates were exhausting to track. December update, then a spam update. It feels like they’re just checking off a year-end to-do list.

— They're wrapping up the year.

— Exactly. I’ve been thinking the same thing. I hope the new year is better. I hope good work gets rewarded. I hope fewer people get hit by these updates.

— I hope so.

— My next point is about Google Search bugs. We sometimes catch these a few days late. At Zeo, we manage many accounts. We have a big team. That helps us spot them faster. We check in with each other constantly. But sometimes, even we miss them. Or we can't fix certain Search Console bugs. What’s your take? I feel like we saw way too many this year.

— It wasn't just ranking bugs. We saw plenty of others. Bugs in settings and request tools. Don't even get me started on position bugs. We see these constantly. Since we manage so many accounts, we know they're bugs before Google even says anything. For example, on the 25th, positions across all accounts jump from 8 to 25. That’s nearly impossible. You’d have to de-index the site or hit spam filters overnight. That’s highly unlikely. We’ve seen many bugs. Even if we know it’s a bug, it ruins our end-of-month reporting. This is huge for large accounts. It skews year-over-year and month-over-month comparisons.

It’s a nightmare for quarterly or annual reports. These bugs just keep happening. Sometimes we have to manually cache the data from the last query for certain accounts. Search Console is our most trusted tool. We rely on it for every analysis. These bugs lead to bad data. Managing many accounts, we catch it and warn clients. But for a single site owner, it’s a massive problem. If there’s a bug, just send a notification in Search Console. Just tell us, "We had a bug on the 27th, FYI." It should be clear. Users shouldn't have to hunt for answers on Twitter or SEO news sites.

I don't think these bugs will ever be fixed. A 100% bug-free system is impossible. It’s not really about fixing them. But bugs reflected in Search Console should at least inform the user. Notifications like "70% of your data is shown, connect to BigQuery" are effective. They show Google is aware of the bugs. People want to save or see all their data. Keeping users informed about this is crucial.

— I totally agree. The real issue is explaining a bug on the 25th to a client. Anyone doing this work will back me up on that. A day or two of lost data or a drop in average position can ruin your month. Clients just look at those positions or their specific KPIs. Explaining that is a huge headache. I hope for fewer bugs next year. They cause serious reporting issues. Of course, bugs happen in Search Console and other areas too. We see bugs in index coverage or crawl stats. I totally agree with you. I want to pivot to Google Search Console. There have been many updates in 2024. I want to highlight those.

These updates are huge for webmasters and SEO pros. The latest update shows Search Console datasını from the last 24 hours. This is great data, especially for big clients. Say you get zero traffic in the morning. Maybe you see a flow between 8 and 12, then it drops. Seeing this hourly in Search Console helps us build better strategies. I really like this new Search Console update. We’ve also seen many documentation updates. Google is doing a great job here. When they update a doc, they note the date.

They usually explain why. They say they updated it for more clarity. We’ve seen many other Search Console updates too. What’s your take on all this?

— That 24-hour hourly view was a must. We used to track data 3 to 5 days behind. That slowed down our actions. This move was exactly what I was waiting for. Maybe not 24 hours, but seeing the day in real-time is huge. 24 hours is actually faster than I expected. It’s a big deal. Most SEO actions aren't urgent, though. They don't need instant fixes. Unless there's a bug. Like accidentally adding noindex or messing up canonical tags. Usually, there's no rush. But tracking it within 24 hours is great. In some sectors, we track rankings in real-time.

High-volume keywords can shift hourly.

— Seeing that is a great feature, right? We had a client whose queries changed hourly. They were obsessed with that data. Having Search Console datasını show that is a game changer.

— Exactly. We used to rely on rank trackers. We’d send requests and wait for the data. Or we’d use SERP API'ları via Google Sheets. Different locations always caused issues. It was a local query nightmare. This update makes everything so much easier. Beyond just spotting bugs, we can track rankings and impressions. We can even analyze hourly trends to gauge user behavior. It’s pretty solid. It’s exactly what we expected. A welcome update. Search Console also had some interface tweaks. They make small changes all year. I’m still waiting for them to improve the crawl stats report in Search Console.

— Right? It could be much better.

— I might be one of the few waiting for this. They need to make that report more effective. It’s the most valuable report. You can use it like a log analysis tool. It shows how fast requests are returned. We talk about this constantly. You can see how request volume scales with response time. It shows which page types get the most hits. It tracks if new pages are discovered. Google should update this to 24-hour data. They already have the data. I’m waiting for that upgrade. Setting aside the bugs, it’s still a top-tier SEO tool.

It’s Google’s own tool, so it’s in a league of its own. It still provides the cleanest data. We’ve also been loving JetOctopus lately. It’s a great addition to our SEO tool'ları stack. I know, I'm jumping around a bit.

— I'll get to that. It's already in my notes.

— Setting bugs aside, 2024 was great for Search Console updates. Especially lately.

— Exactly. Also, Google Trends had a major update. The UI changed. We got more data. This is huge. Publishers can now use Google Trends datasını much more effectively. The 24-hour data is a game changer. There is so much data now. You can export it. Tons of details were added. You know the insights screen in Search Console? The BigQuery entegrasyonu is even more critical. Like you said, we can even see 50/50 split projects. Search Console tells you that. It says you'd see 50% more with BigQuery. My 2025 advice for brands? If you manage a large account, set up that BigQuery entegrasyonu. You definitely need it.

Try to get it done. It’s great for analyzing Search Console datasını. Also, the Page Experience report was removed. That’s worth mentioning too. Regarding the 2024 updates in Search Console.

— It wasn't a very actionable report anyway.

— Exactly. It was a headache for us too. The data often lagged behind. Clients would check Core Web Vitals, then Page Experience. They’d ask why their scores were low. It was tough to explain. Google sent a clear message by removing it. Everyone got the point. They already stripped out mobile usability and HTTP URLs anyway. It lost its relevance over time. This leads me to the leaked Google documents. That was a huge 2024 topic. We discussed the link data internally quite a bit.

It even had specific details. It checks the last 16 months of a page. We immediately thought about the 17th month. If we keep a page updated, we show Google we're performing well. Also, if we want Google to wipe old content. Maybe we need 15 or 16 updates. What are your thoughts on these leaked Google documents?

— You mean the Yandex leak.

— Yes, that too.

— Then the Google leak followed. It was totally unexpected from Google. Nobody thought a Google leak would ever happen. These were just rumors and known developments before.

— Sorry to interrupt, but they dodged it well. They stayed quiet and calm. It just faded away. It’s only a topic now because we’re talking about it. Most people didn't even care.

— It comes down to this. These aren't major ranking factors. We’re fine. We’re already improving everything. There’s no immediate crisis. The leaks just confirm what we’ve been discussing. These are things we’re already acting on. Keeping content fresh is key. Content scores matter. Accuracy is everything. EAT is obviously a huge part of this. I’m talking about content now. Link scores are also in the mix. The leaks just validate our current strategy. We’ve read a lot about this. We’ve discussed it internally. Honestly, I expected something more sensational. I wanted a "secret" nobody knew about. We’ve already done so much work here.

Remember when we used AI ile to map out our strategy? We checked for any gaps. We’ve done a lot of work. The leaks mostly just reinforced what we already knew. They confirmed our strategies were solid. It was a weird move from Google. For our clients, the leaks actually helped justify our actions. It made our strategy click for them.

— Exactly. I think about it from my perspective too. You see certain metrics. Maybe we fall short in our reporting. Like explaining the math behind a metric. The leaks really clarify that. See how critical that math is? We can even build our own formulas for reporting. You can see it right there. When improving a page. Define a metric like "can renk." Then run your own statistical calculations on it. It’s all visible there. It clears up so many lingering questions. But you know, details are scarce. We often just guess what metrics mean by their names.

There are metrics for new sites. There are details on how page context is parsed. We’ve seen them and taken action. If you haven’t looked yet, you really should. This was one of the biggest events of 2024.

— Exactly. Let me add a quick note on updating content from the last 16 months. It’s crystal clear in news sites. They publish so much content. It shows why staying current is vital. Even evergreen content needs updates to stay relevant. Update your pricing and dates at year-end. Take action on this. If you’re an authority site with strong competitors, act early. It puts you one step ahead of the game.

— Exactly. Let’s talk SearchGPT. You’re doing a ton of research on this. You’re building on the AI side too. I’ll leave the floor to you. We can actually see these AI results in our logs. In 2025, we’ll likely see way more AI bots in our log analysis. What’s your take on SearchGPT? Especially with all the 2024 developments.

— SearchGPT was expected by year-end. There were plenty of benchmarks. Everyone on Twitter is debating usage rates. They say it’ll match Google by 2026. SearchGPT works differently than our standard browser logic. It’s disorienting at first. We’re used to typing, browsing, and clicking results. SearchGPT does something else. It takes your request, pulls answers from the web. And synthesizes them with its own common knowledge. It’s basically a web-based Q&A or referral system. What’s the real difference from Google? Let’s compare the moves. You still send a request. If it’s informational, it gives you the answer directly. No browsing or searching required.

Everyone treats this as informational at first. It’s used for informational purposes. AI's text-based nature fits this perfectly. It’s great for Q&A or gathering info. SearchGPT is a unique OpenAI tool. It’s a departure from the norm. SearchGPT doesn't run on Google. It’s a search engine built on Bing. It’s more of a next-gen search engine. It runs entirely on Bing infrastructure. It uses Bing data exclusively. You can see SearchGPT traffic in Bing Webmaster Tools. They’ve broken it out there. Like you said, it’s showing up in reports. It will show up even more in 2025. Don't just focus on SearchGPT.

We started with SearchGPT. But there’s Perplexity and Coma AI too. Many tools now pull live web data. Their usage rates have spiked recently. They’re still used less than traditional browsers. But we expect massive growth ahead. Especially as it shifts from informational to transactional. Take this example. You search for a specific red shoe. You describe the exact model and cut you want. You find the product just by describing it. This will speed up sales. E-commerce will see a huge boost. Users will reach their target products much faster.

I mentioned this in a meeting recently. I don't think OpenAI will invest heavily in SearchGPT. Sure, they have plans and structures in place. But I believe OpenAI will focus on new models and SuperAGI. SearchGPT is active now. Users are testing it. Our industry uses it a lot. We want to see how it pulls search results. Is it based on Bing rankings? We use SearchGPT and Perplexity often due to client questions. Honestly, I don't see OpenAI making massive investments there. They’ve already launched it.

Updates will come as usage grows. But I think their big money goes into AGI. Still, using Generative AI in search is an undeniable reality. Brands want to take action. They want to show up there. First, every site must open itself up to AI bots.

— Yeah, some people are still blocking them.

— Yeah, I see it.

— It was a craze. At first, everyone blocked it in robots.txt. They pulled some stats, but you need to open it up.

— Yeah, that was the initial reaction. Brands feared it would crawl their site. Sure, it can crawl. Let it crawl. If it's a brand decision, we can't argue. But from an SEO perspective, it doesn't hurt your site. If you're getting too many requests and it's crashing your servers, that's different. You have to take action there. If these AI bots are causing server issues, you have to handle that. But generally, they need to see your site. Not just for today, but for the future. I think about it like Google. Does OpenAI crawling a site extensively make it more authoritative in the future?

That's how OpenAI works at its core. More data means more authority. We might see that happen soon. If you have the content and the resources, I say open it up. We want users to see us. We want traffic from Search. We want traffic from SearchGPT too. That’s the main goal. Basically, your optimizations for Google for Google için also work for Bing. Bing has some small tricks. Google has its own differences. But we optimize for all search engines. We don't just build for Google.

Google is dominant. That’s why we share Google insights. Their documentation is better. But search engines work similarly. Our optimizations help Bing too. They help SearchGPT as well. If you want to optimize further, I think you’ll agree. Make your informational content more answer-focused. That’s key. I mean answer-focused. We used to do this for snippets. For "2025 ticket price," we’d state the price clearly. We wanted to fit the featured snippet structure. We wanted to give users direct answers. Optimizing for featured snippets is effective for AI. It provides clear answers. Users go to SearchGPT or Perplexity. They just type "2025 ticket price."

It shows how a query works. Or something similar. It sends the query to Search. It pulls info from vectors or Search. Then it brings it to us. If you have the data and AI bots crawl you, you'll likely show up. The new year will bring changes to Search, GPT, and AI.

— I should add something. You mentioned Perplexity. Some people are actually making money from it. Check out those videos. It will still be huge in 2025. Let's connect this to Google AI overview results. It started in the US. Then it hit other English-speaking countries. Now it's spreading to places like Brazil. Will Turkey see this in 2025? We saw AI overview in 2024. I think there's a quality threshold for these queries. Google might publish it in Turkish once that threshold is met. We see many studies on this abroad. You just mentioned them. That is what we do, Search Engine Optimization.

We do this for everything. Maybe check where AI overview triggers in the UK. See which sites rank. If it hits Turkey or other markets, how can I optimize for those queries? Analyze that data. Most SEO tool'ları provide it. Don't overcomplicate it. These optimizations are key. They’ll help you prepare for AI overview integration in 2025.

— I agree. You follow this closely. You’re more informed. I just see the Twitter chatter. AI overview results have bugs. That’s totally normal.

— The design keeps changing, too.

— Exactly. They keep tweaking the design. Bugs are inevitable. I expect plenty of bugs in the Turkish version. They’ll stabilize it by 2025. I think it launches in Turkey by Q1. They often test in Turkey mid-cycle. They spot the bugs and iterate. We’ll definitely see more of this in 2025.

— Let's see what's coming. I use a VPN for Discord. Sometimes it stays set to the US. I get confused. I see Turkish results and wonder where I am. Is it in Turkey yet? We'll see how it impacts things. My next note is about Reddit, Can. You know Google has a deal with them. They're all over the search results now. I wanted to touch on that. Reddit was huge for us in 2024. They started showing up for so many queries. I think you'll agree with me here. We see Reddit more often now. Not just for Turkish queries, but everywhere. Reddit is definitely on the rise. We had to mention it for 2024.

— I love Reddit. I use it all the time. It feels like Google is playing favorites. Like they're saying, "You guys keep going, we'll handle it." Reddit is a massive site. It was already ranking. You could always find Reddit for complex solutions. It wasn't huge in Turkish, but it was everywhere in English. Maybe it just stands out more now. But the stats definitely back it up. Reddit's traffic is climbing. I don't know. Sure, it's a forum with real solutions. It's helpful for users. But was this really necessary? I'm not sure if we needed to know about this partnership.

It’s reached that point. Reddit is basically Google’s teacher’s pet now.

— Google basically told Reddit, "I've got the formula, don't worry." They boosted them. Next up is structured data and schema improvements. I had to cover this. We’ve already done some great work here. Think product variant improvements. PDP pages and profile pages. Google is pushing author pages and 3D models. They’re also adding more support for energy labels on appliances. I read that support will grow even more in the coming months. Carousels. We use them on many projects. Finally, returns and Merchant Center integrations are huge. These are the most critical structured data topics for me. They’ve been technically efficient this year. Many people see the value, but they struggle with implementation.

Technical gaps or sprint priorities often get in the way. But heading into 2025, if you run an e-commerce site, use product variant schema. Use it for sizes, models, and variations. It’s a must-have recommendation. If these aren't in your 2025 roadmap, add them. Google prioritized these heavily in 2024. They displayed them directly in Search. Sites that implemented them definitely saw the results.

— Agreed. You really need to use product variants. Many e-commerce sites struggle with this. They have variants that get flagged as duplicate content. Using the variant schema is a must here.

— Exactly. Also, check the documentation updates. Google is detailing more schema types now. They want those added too. You have to pay attention. There were huge updates in 2024. I need to repeat this. The Rich Results Test isn't everything. It won't flag every error. It doesn't mean your data is perfect. Always check the specific guidelines. Mark up your content accordingly. In 2025, manual actions aren't just for spam. You can get hit for bad structured data. It can be a partial or full site penalty. You have to be careful.

Schema will stay huge in 2025. My next note is on web stories and EMP. I'm passing the ball to you. You have a rare web stories client in Turkey. At least, that's what I think. What’s your take on EMP? Support for EMP has really dropped lately. It might be gone in a few years. It’s worth remembering why it launched. I’ll hand it over to you. EMP was for accelerated mobile pages. Now, responsive mobile sites are often faster than EMP. It’s often better to ditch EMP and switch to a responsive mobile version.

What are your thoughts on web stories and EMP?

— I agree with you on EMP. I think support will end soon. We still get good traffic from EMP, but for different reasons. Honestly, relying on EMP for traffic is a problem. We’re working on that now. You’re getting requests from too many versions. Publishers want to shift that traffic to mobile for ad revenue. Shifting that traffic is crucial. I remember when we first talked about web stories. For some reason, nobody wants to use web stories.

— I think I wrote a blog post when it first launched. I should check the SEO blog. It’s wild that nobody uses it.

— Turkey has a really advanced publishing scene. Especially regarding SEO.

— Yeah, there are some great teams.

— We’ve worked with some excellent teams. You remember them too. But nobody really adopted web stories. They didn't want to. We gave them the recommendations. They just didn't want to take action. I see them used abroad all the time. With AI overview and everything, their importance has dropped. They’re fading away like EMP. But I never saw them used before that. Given how much we use social media. They should have been a priority for the visuals. And it’s a tough one.

— Let's say it didn't work. It just didn't catch on.

— It was weird. It didn't impress anyone. But we got great results with web stories. At the end of the day, Samet, you know how it is. You're the only one doing it. Users aren't used to it yet. You end up feeling like you're just shouting into the void. But the setup isn't that hard. You build a system. The web stories part generates automatically with EMP logic. At least that's how our system works. I don't know, it didn't take off. We're still doing web stories anyway.

I'm sad it didn't work. It's a feature I really like.

— Let's see what happens. Maybe Google says they're killing it or changing it. That's unlikely though. We'll see. I definitely wanted to cover this. Next up: legal issues and search changes. I wanted to dive into that. We're already discussing this. Here's what we saw. In some EU countries, legal rulings came out. Google hotel results are being limited or removed. Legal issues definitely impacted our work in 2024. Keep that in mind. For example, in certain countries, the hotel pack disappears. You just get organic results. You see PPC ads first, then organic results follow.

You had to watch legal outcomes in 2024. It might be even worse in 2025. Then there are search shifts. We see hundreds of search tests. If a search change hit your brand, you had to track it in 2024. We saw a lot of this in the USA. Products on sale or coupon codes appeared. Some search appearances have no technical data. There’s zero documentation on how they actually trigger. It doesn't come from Merchant Center. It’s not in structured data. We had to deep dive into these issues. We just left a year with too many search appearance details. Average positions still matter, though.

Some SEO tool'ları show pixel rank or base rank. Your base rank can spike when image results drop. You really had to pay attention to those nuances. We saw so many search shifts in 2024. You must document them. If they hit your industry, you need a plan to rank. What are your thoughts on this?

— Yeah, search changed a lot. Google ditched infinite scroll for pagination. It’s a weird move. Users were so used to infinite scroll. They went back to clicks. 2024 was packed with new features and snippets. Things were constantly added or removed. It changes based on the query. I track these changes closely. I manage a sports news site. There’s always something new. One day you get great traffic. The next day, clicks drop. Your position stays the same. Then you see a new feature.

It starts eating your static traffic. You mentioned legal stuff. That’s true. There were huge legal developments this year. Updates brought legal changes too. This isn't just about the TR market. It’s global. Google was busy there too. You have to track search changes. Forget average position. Track pixel position instead. What are your pixel changes? That’s what we track at WOP. With AI and AI overview, search will change even more. I’m curious about what’s next. Search for a team's fixture and it pops up directly. No clicks needed. Can we grab this with AI? Will these features stay? Or will they become integrated?

I'm curious. Where are we headed? Will we vanish as pixels? AI overview will change everything globally. Big shifts are coming. We need to study the US market. I haven't checked the post-AI search changes yet. We need an analysis for TR. Some features might disappear. We could see traffic return.

— Exactly. Take the Turkish 2nd League. Search for a team and the logo is missing. You get the fixture and standings, but no logo. Maybe AI overview or new search features will fix this. Google might just lack a high-quality source. I remember this being an issue years ago. AI overview might solve these minor glitches. Let's see what happens. We'll find out together. Remember, in 2024, we used to have 10 blue links. That’s not a thing anymore. For brand queries, page one often shows only 5 results. Sometimes even just 3 organic results. Let’s see what’s coming in search for 2025. You must be ready for these changes.

Next up is Core Web Vitals. There have been tons of updates. Some are quick wins. We’re even struggling to keep up with some. Remember the Interaction to Next Paint change? Google added AVIF support. Plus, constant updates to the Google Chrome developer console. Lighthouse was already integrated into Chrome. They created more documentation on using the developer console for SEO. The SEO community was kept in the loop. 2024 was a very productive year for this. You know. We discussed how to use the Google Chrome developer console efficiently for Core Web Vitals. Those were my main takeaways for Core Web Vitals in 2024.

— Like you said, a new metric hit Core Web Vitals. Everyone was confused about it. It’s different from standard Core Web Vitals. It’s not just about page load speed. It measures the data or time from user actions. It’s fundamentally a different metric. We had to figure out our next steps. When it first launched, most sites had issues. We got tons of warnings. We struggled to fix them. It’s an interesting metric. Like I said, it’s an action-based metric. Google doesn't really track form actions.

— The name says it all. It’s about interaction. You have to look at it that way. People just say fix your JavaScript. I want to ask, did you even read the definition? You aren't offering specific advice. You just blame third-party scripts. Sure, they matter. But read the definition first. Optimize based on that. That’s the logical way to give advice.

— Exactly. Also, there were tons of rumors about Core Web Vitals this year. John Mueller had a lot to say about it.

— He hasn't been off Twitter for months, by the way.

— Suddenly, people claimed Core Web Vitals don't matter. But they impact page experience and user experience. How can they not be important? It’s been a huge debate. I’m not sure, but these metrics still matter.

— They definitely matter. Absolutely.

— They impact user experience no matter what. Cumulative shifts, load times, and meaningful paints. You need to optimize these. It’s crucial to see a clean layout when you land on a page. We talked about DevTools in our last training. It’s a killer tool for site speed. It gives you so many metrics. You can test everything in real-time. You can spot layout issues in performance. You must use it for Core Web Vitals. Sure, PageSpeed Insights or Lighthouse show the metrics. They show Largest Contentful Paint or cumulative shifts. But they don't always let you debug live. PageSpeed doesn't quite provide that level of actionable insight.

That’s why these things feel vague to us. But DevTools makes them easy to solve. You can clearly see CLS, LCP, TTFB, and load times. You should definitely use it for performance. Like you said, there are plenty of resources on this. Maybe we can do a DevTools session on Core Web Vitals, Samet.

— Samet, did a Core Web Vitals SEO Talks episode happen already? I think it did.

— We can do a DevTools session for Core Web Vitals.

— Exactly. I won't drag this out, but there's a lot to cover. In 2024, I noticed many sites using caching. They cache certain resources. But only for one or four hours. You’re a static site. Your content doesn't change that fast. I suggest webmasters and SEOs check this. Look at your caching details. Use the Chrome developer console to verify them. Is the cache duration too long or too short? Checking that is much more useful. Also, look at data from PageSpeed Insights or GTmetrix. They give you tips like shrinking the DOM.

If you use the Chrome developer console well, you can see everything. I showed this before. You can spot which HTML elements bloat the DOM. See which ones are problematic or have too many child nodes. This is crucial when talking to dev teams. You can't just say "shrink the DOM." They won't know where to start. Use the developer console to give them a clear brief. I can say this for 2024. It will hold true in 2025 too. I have notes on translated results, Can. You see them in search, right? If quality Turkish sources are lacking, Google shows translated results.

This existed before. But in 2024, they added languages like Arabic, Gujarati, Korean, Persian, Thai, Turkish, and Urdu. That’s one of the things that caught my eye regarding these translated results.

— I think this is a smart move. Seeing this increase makes sense. Google also plans to reduce spam and clickbait. They want to prioritize serving high-quality, accurate content. If there's accurate Arabic content, I want to see the translation. Would I click it? Is the site trustworthy? That's debatable. If an unknown site shows up in Search Console, are we really clicking it? Looking at user behavior, it's unlikely. But if we assume the strategy is providing correct info, it makes sense. You deal with this more often. Especially on the Arabic side.

I run into this constantly in Arabic. It's the most common area for me. You probably have more experience here. I think it will spike in 2025.

— Definitely. It's important, Can. As a user, if I see a translated result, I keep going. I feel like it provides good data. I don't bounce just because it's translated. That's my personal take. I've seen wild cases in Arabic. Egypt, Saudi Arabia, and the UAE speak the same language. Yet, you see very different results. Since it's not the Latin alphabet, it's hard to grasp the intent. You have to be careful here. I think these results will spread to more languages in 2025.

We see better optimization in translated results. Search Console has a report for this. Feel free to check it out. I'm reaching the end now. I saved SEO tool'ları for last. Specifically the indexing API. I had to include it, Can. It comes up so often. Developers deep into SEO always mention using the indexing API. They claim faster indexing or share unique cases. That is the point of the indexing API. I'll promote myself here, but I wrote a Zeo blog post on this. It covers how to use it and the guidelines.

It launched for job boards. Job listings expire quickly. It was meant to get them indexed fast. If I recall, Google updated this in October. Google updated the indexing API documentation back in October. It is very clear there. If you use the indexing API, it undergoes strict spam checks. Don't think you're clever by using the indexing API to spam. If you think you can game the system, think again. You might want to step back. Google updated the indexing API documentation in 2024. They’re getting much stricter about it now.

— SEO pros always want to exploit every new feature to the max.

— Plus, there’s a quota limit there.

— Exactly, there’s a quota. If it works for you, use it. If you run a job site, you should definitely set it up. It’s a no-brainer. But should a news site run both a news sitemap and the indexing API? Are we trying to break Google’s results? Where does it end?

— We don't know what Google is doing on the backend. They might be doing it. We don't get notifications for that. You have to be careful to avoid spam triggers.

— I was just about to say something about that, Samet. I expect transparency from Google. They should say you dropped because of the indexing API. Stop using the indexing API. Learn to index your pages properly. I want that transparency. It should show up in Search Console just like a manual action.

— They actually used to do that, Can. Back when manual actions were common, they’d slap a spam penalty on your site. They’d tell you your site was hit. You’d submit a reconsideration request. You’d explain you fixed the issues. I think it was a reviewer note. The person writing it put in real effort. They actually reviewed the site. It was valuable feedback. They’d say why they couldn't approve the request. It would be nice, but it’s unlikely. Even if they did, it would be vague. They’d just say something like "quality issues." It’s always a very high-level response.

— Quality issues. I think people look at the indexing API the wrong way. You should definitely use it in the right sectors. It’s like not using a sitemap for news sites. You have the option to use it for news. You should take it. Use it. I think you should use the indexing API here. I don't expect them to loosen the indexing API policies. I doubt they'll make those policies more transparent either. We’re a community that fills requests even from Search Console. Publishers operate the same way. I don't think this will get out of hand. Still, it’s a great development.

— Exactly. It’s worth repeating. One site might succeed using it. But a bad example isn't a good blueprint. That site could crash in a month. Play it safe. It’s always better. The guidelines are clear. The rules are set. Maybe nothing happens today or next week. But we don't know about three months from now. Act accordingly. One of my final notes, Can. I wanted to add SEO tool'ları. We had the chance to test many a different SEO tool'u in 2024. The creators of these SEO tool'ları put in a lot of work. They’re launching new ones. Even Ahrefs integrated AI overview results into their platform fast.

I remember something. Before I worked at Zeo. Google announced new link types like sponsored and user generated content. Ahrefs added them to their interface in just a day or two. SEO tool'ları performed well in 2024. They’re great for tracking trends. They really support SEO pros. Lots of new stuff popped up. Semrush even acquired Search Engine Land. We saw big shifts in various SEO tool'ları. You mentioned it earlier. JetOctopus is a solid tool. It’s a great alternative to Ahrefs and Semrush. Give it a try if you’re curious.

Those were the first ones that came to mind.

— We measured this in 2024. Who integrated AI overview? If I recall, SEO Monitor did. It’s part of their core philosophy. So, the integration was natural.

— Yeah, it’s already integrated.

— We always talk about the big SEO tool'ları. I really respect SEO Monitor. It’s incredibly stable. It just works without issues. You know how Ahrefs is. Sometimes Ahrefs updates its keyword data. Everything changes in an instant.

— Yeah, it messes up the analysis. A keyword database update hit during that update. It’s visible. The database updated the same day. You can't tell if it’s the update or the database. I’d be more relaxed if it didn't tell me that.

— We lived through exactly that. We even talked about it. Ahrefs shows a drop, Semrush shows a rise. Search Console datası is three days behind. We can't track it. We face these issues, unfortunately. But Ahrefs has really stepped up lately. But I think our SEO tool'ları were solid in 2024. One new tool I’d mention is JetOctopus. It honestly blows my mind. It’s just...

— Not a sponsor, by the way. We’re just sharing what we use.

— Not an ad. It’s a great tool. We’re used to SEO tool'ları like Ahrefs and Semrush. This one felt different. It offers deep analysis. The UI is beautiful. It gives you everything. The workflows are clear and actionable. These are huge upgrades for Screaming Frog.

— I was going to say that. Screaming Frog is the gold standard. Shoutout to them.

— It was a great year for Screaming Frog. Their AI integration and OpenAI features are effective. The new UI updates are great too. They just need to move to the cloud. They need to process data in the cloud. We don't want this just for local performance. Cloud access lets us automate everything. We’re stuck pulling data locally right now. It creates a bottleneck for our workflows. I think a cloud move is coming soon. Sometimes we pick cloud-based tools over Screaming Frog. We might choose Lumar instead. It’s easier to just let it crawl and check back later.

Screaming Frog is a given. Nothing left to say there.

— One wish for Screaming Frog though. I know it’s a stretch. The spelling error report. It has Urdu but no Turkish. Adding Turkish would be huge. You can’t just add it yourself. It pulls from a language platform. They say you need to donate. But even then, there’s no guarantee. I hope it happens. We could fix those issues too. It works great for other languages. We just don't have that resource for Turkish yet. Nothing else comes to mind, Can. Beyond what you covered.

— We really covered a lot of ground.

— Yeah, it went long. Hope we didn't bore the audience. It’s been an hour and seven minutes. Hope this was a useful stream. For 2025, I wish everyone tons of clicks. High organic clicks and better average positions. May your pixel rankings improve. May organic revenue rise even against foreign currency. And may your conversion rates hit new highs. I hope we get great results.

— That’s a great wish. Truly the best wish for organic traffic. I hope 2025 is a great year for SEO. Let it be a year of growth, not bugs. A year of development. I expect 2025 to be packed with updates. Core updates, spam updates, and helpful content are all integrated. I expect plenty of updates. Especially if AI overview hits Turkey. 2025 will be a core year for us. I hope it’s a great one. I think 2024 was a solid year. I’m talking strictly from an SEO perspective. Overall, it was a good year. I hope 2025 is the same.

— I hope it goes well. Thanks to everyone for listening and watching. Doing a classic YouTuber sign-off here. If you liked the video, hit like and subscribe. Have a great day, everyone. Goodbye. Take care.

— Thanks everyone. Happy New Year.

This transcript was generated using AI speech recognition and verified against subtitle style rules.

Yiğit Konur

Content author

Yiğit Konur

both

Contact us

How we handle your data

Open as a full page