YT YouTube2Text

YouTube2Text → Transcripts → You're Doing SEO Wrong (This Hacker Proved It)

You're Doing SEO Wrong (This Hacker Proved It) — Transcript

by Stephen Baker · 2,838 words · 426 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00In 2024, Mark Williams Cook, a security
0:02researcher and SEO expert, found a crack
0:05in Google's [music] system. He found
0:07that when he sent a request to one of
0:08Google's communication channels in just
0:10the right way, Google accidentally sent
0:13him back a bunch of data and information
0:15that should not have been available to
0:17the public. Now, Google's algorithm
0:19doesn't live in one place. You can think
0:20of it like a postal service. They've got
0:22a bunch of different departments that
0:23are constantly communicating with each
0:24other about what content should rank and
0:27what content shouldn't rank. But, Mark
0:29found one of these channels, and he
0:31discovered something wild that when he
0:32poked at it in just the right way, they
0:35accidentally showed him everything in
0:37plain text. So, naturally, Mark alerted
0:40Google about this hole in their security
0:41system, and they actually thanked him
0:43for it. They also sent him $13,000 as a
0:46reward. But, that's not the end of the
0:48story because before Mark told Google
0:50about this, he got access to massive
0:53amounts of data. 2 TB worth of data, to
0:56be exact. To put that into perspective,
0:58that's 90 million Google searches worth
1:02of information. So, this let him see the
1:04rules that Google has to decide which
1:07websites rank number one and which ones
1:09disappear to page 10. Now, what he found
1:12shocked him, and it's going to shock
1:13you, too, because it's the complete
1:14opposite of what most people believe
1:17about SEO. Right now, you're probably
1:18focusing on things that don't matter.
1:20Your SEO agency is focusing on tactics
1:23that are outdated. Or, business owners,
1:26you're paying thousands of dollars for
1:28an agency that isn't doing the right
1:30work when it comes to ranking your
1:32website number one on Google. I'm going
1:34to show you the four biggest myths about
1:36SEO. And with Mark's discovery, the
1:39truth behind how Google actually works.
1:42By the way, Mark's full lecture on what
1:44he found when he hacked Google is on
1:46YouTube right now. You can watch the
1:47full thing. I'm going to link it in the
1:49description, and I would highly
1:50recommend you watch it because I found
1:52it extremely fascinating, and he goes
1:54into way more detail than I'm going into
1:56in this video. The four myths that will
1:58be debunking in this video are as
1:59follows. One, you need a high domain
2:02authority to rank number one on Google.
2:05Two, you need to hit a specific keyword
2:07count on each of your pages. Three, you
2:10need high-quality backlinks to rank. And
2:12four, you could just make a bunch of
2:14content and a bunch of pages on Google
2:16and you'll eventually be successful.
2:18These myths are so common in the SEO
2:20space today and I still see tons of
2:22comments on my channel and across the
2:24internet about them. But to understand
2:26where they come from and why people
2:28believe these things about SEO, we first
2:32have to look at the history of Google
2:34and the origins of how its algorithm
2:36ranked websites and how it's evolved
2:38today. In 1998, Google was founded by
2:41two Stanford students, Larry Page and
2:43Sergey Brin.
2:45Their idea was to create a search engine
2:47that ranked websites based off of
2:49citations. They noticed that other
2:51search engines like Yahoo, for example,
2:53didn't have a proper structure or a
2:55proper algorithm to rank websites
2:57properly and so often times when someone
3:00used these search engines, the results
3:02were pretty sloppy. What they originally
3:04set out to do was rank websites by
3:07citations just like research papers. It
3:10was a clean and elegant algorithm that
3:12they had and so websites ranked higher
3:14based on how many backlinks or citations
3:18they had. This is where the myth that
3:20you need to have backlinks originated
3:22from. Over the next few years though,
3:24people found this out and they tried to
3:26gamify their rankings by getting more
3:28backlinks, buying them, link building.
3:31This became its own industry and people
3:33essentially tried to gamify their
3:35rankings. Then 2005 rolls around and at
3:38this point, SEO is established as its
3:41own dominant industry where people are
3:43making real money with this. One of
3:45these SEO companies called Moz, which is
3:47still around today, created a metric
3:49called domain authority. Domain
3:52authority essentially allowed people to
3:54compare the relative authority of one
3:56website to another. However, this was
3:59never a factor that Google had created
4:02to rank websites. Users of Moz saw that
4:05if a website has a higher domain
4:07authority score, they typically rank
4:09higher and get more traffic. And so,
4:12they took that as a causal factor in
4:14rankings rather than just a correlation,
4:16which is exactly what it was. Now, at
4:18the end of the day, it's hard to blame
4:20people for thinking that. But, Google
4:21has never used domain authority as a
4:24metric in their algorithm for ranking
4:26websites. But, this is where the myth of
4:29needing to have a high authority score
4:31came from. From 2005 to roughly 2011,
4:35Google continued to make more and more
4:37updates to refine its algorithm so that
4:40factors like backlinks weren't the only
4:42things that mattered anymore. This is
4:44where keywords came in. Now, people
4:46figured out that if you have more
4:48keywords in your content and you have
4:50longer content on each of your pages,
4:52they have a better chance at ranking
4:54high on Google. So, this is where myths
4:56number two and number four were formed.
4:59Just add more keywords to your page and
5:01publish more content more frequently.
5:04Fast forward to 2022. This is when
5:06ChatGPT came out and it's the AI content
5:09explosion. People figured out that if
5:11the formula to rank on Google was get
5:14more backlinks, write more content more
5:16frequently, put more keywords in the
5:17page, etc., you could just automate most
5:20of that with ChatGPT. As you would
5:22expect, this was not a good thing for
5:24Google. Their original vision of having
5:26a search engine that had true
5:29high-quality search results for its
5:31users was essentially being taken down
5:33by AI. Now, even worse than this, a big
5:36group of people who were in the SEO
5:38space using these AI tactics were
5:41getting great results. So, these myths
5:43were solidified in their their that if
5:45you do these things, you will rank on
5:47Google. But here's the thing. Since that
5:50time, Google has made so many updates to
5:53their algorithm quietly often in the
5:55background and people haven't noticed.
5:58And those same people who are still
6:00doing SEO the old way and trying to
6:02optimize for those old metrics have seen
6:04a decline in their rankings and their
6:07progress on Google. So the question is,
6:09what is the Google algorithm
6:11prioritizing to rank websites today?
6:13Well, let's jump back to what Mark
6:15actually found. So when someone looks up
6:17into Google something like painters near
6:19me, the first thing Google is doing is
6:22filtering which websites are even worth
6:25considering. And this is where most
6:26people go wrong. Mark discovered this
6:29when he accessed Google's algorithm.
6:31>> Google had what they called a site
6:33quality score for every single website
6:36on a subdomain level. It gave them a
6:38score of zero
6:40uh to one. And
6:43we got this data for 800,000 domains.
6:46>> So every single website, including
6:48yours, gets a score not determined by
6:51backlinks or content optimization, but
6:54something else entirely. Here's how Mark
6:56explains what determines that score.
6:58>> Site quality score is calculated based
7:00on how many times people specifically
7:03search for your website along with other
7:06search terms, how often they select your
7:09website even when you're not number one,
7:11and how often your website name or brand
7:14name, if you want to call it that,
7:15appears in anchor text around the web.
7:18>> So Google is asking three questions. Do
7:20people really search for this business
7:22by name? Do people click on this website
7:26even when it's not ranking in the first
7:28position? And do other websites mention
7:31this site by name? These are brand
7:33signals, not the typical SEO signals
7:35that most people think about. Now,
7:37here's where it gets interesting because
7:39when Mark compared websites that ranked,
7:41he found something that contradicts what
7:43everyone believes.
7:44>> So, this is one of our clients rehabs.
7:47Um it's ranking top for a YMYL type
7:50query around addiction. It's outranking
7:52uh UK, which has been around way longer
7:55than them. It's got like 15 times as
7:56many links. It's outranking NHS. Again,
7:59on all those other metrics, from a
8:01classical SEO point of view, they should
8:03be losing, but they're not.
8:06>> So, a newer website with 15 fewer
8:08backlinks outranked the NHS. This proves
8:12that having high domain authority and
8:14tons of backlinks does not guarantee
8:17your rankings. What matters is whether
8:19Google recognizes your business as a
8:22legitimate brand or not. Now, there's a
8:24second filtering stage that Google is
8:25going to go through when it's ranking
8:27websites. But, the crazy thing is that
8:29if your website does not even pass that
8:32first filtering stage, if your quality
8:34score is too low, you won't even get put
8:37into the second stage.
8:38>> And what we found is if you have a site
8:41quality less than 0.4, you are not
8:43eligible to get things like featured
8:46snippets or PAAs. It doesn't matter what
8:48guides you follow on optimizing for
8:50these things, you won't get them.
8:52And that's something that became
8:53apparent in this algorithm is there's
8:55lots of like heat races you need to
8:57qualify to get to the next kind of stage
9:00in ranking.
9:01So, it doesn't matter what you're doing
9:02over here if you're not even in that
9:04race cuz you haven't filled this
9:06prerequisite first.
9:07>> So, in other words, Google is first
9:09asking, is this a real brand? And if you
9:12don't pass that test, then your website
9:14is not even showing up on that search,
9:16and no amount of optimization is going
9:18to help you. But, if the answer is yes,
9:20and you pass that filtering stage of
9:22being recognized by Google as a real
9:24brand, then and only then will Google
9:27look at your content. And this is where
9:29a lot of people waste time. People will
9:31spend hours writing blog content,
9:33hitting specific keyword goals, and
9:35making landing pages. They think, "Hey,
9:36if I hit this specific 2,000 word count
9:39on my page or on my blog, then I'll
9:41rank." But, Mark discovered Google does
9:44not work this way anymore.
9:45>> Google tells you to make helpful
9:47content. And I'm certainly not saying
9:49this is a bad thing. You should
9:50absolutely do this, but the Google
9:53algorithm has a target. And if you think
9:56that target is helpful content, you'll
9:58probably find out like many people, you
10:00will spend ages making helpful content
10:02and find out you don't actually rank.
10:04This is because there are other forces
10:06pulling on this,
10:08which in Google's case is making ungodly
10:11amounts of money.
10:12>> People try to optimize their content and
10:14they still don't rank. Why? Because
10:17Google isn't measuring word count
10:18anymore. Mark was pretty blunt about
10:20this.
10:21>> It doesn't make a lot of sense for
10:22Google to rank a site on something that
10:25you have full control over like that as
10:27a metric.
10:28>> If Google ranked based on word count,
10:30someone could just write 10,000 words of
10:32absolute garbage and Google would still
10:34rank them. But, that's not what Google's
10:36trying to do. It's trying to show people
10:39the right information and helpful
10:40content. Google doesn't care if your
10:43blog or your website page is 2,000 words
10:46or 200 words. It matters if the content
10:49is correct. So, myth number two doesn't
10:51exist either. Reaching a specific word
10:54count is not going to help you. Google
10:56is asking, "Does this content answer
10:59what the person is searching for?" Not,
11:01"Is this content 2,000 words or more?"
11:03What about the other myth though? Of
11:05just publishing hundreds or thousands of
11:07pages to rank. Well, here's something
11:09that happened in real time that proves
11:11this doesn't work.
11:12>> I gave you this example of multiple
11:14sites going from zero to over a million
11:16visitors in 16 weeks
11:19um on brand new domains just using uh
11:22zero volume keywords and basically
11:24spamming out AI [music] content. And
11:26that was because of recent Google
11:28updates had left this as like a big open
11:30goal for them. We had a whole second
11:32renaissance in niche blogging, I e.
11:35people just setting up blogs and they
11:36found anything ranked really easily.
11:39Um and Google was just kind of like,
11:41"Whoops, probably shouldn't be doing
11:42that."
11:43>> Like I mentioned back in 2022, when
11:45ChatGPT came out and just changed
11:47everything, people would publish
11:49hundreds or thousands of pages, but
11:51here's what Mark saw happen. Sites were
11:53publishing thousands of pages of AI
11:55content and getting millions of views,
11:58but Google shut this down immediately.
12:00Publishing massive amounts of content
12:02won't work anymore. Google first filters
12:05based on whether it recognizes this
12:06website as a real brand or not. So, myth
12:09number four debunked as well. So, if
12:11domain authority doesn't matter,
12:12backlinks don't matter, word count,
12:14keywords, publishing a bunch of pages,
12:17what actually works in 2026 for ranking?
12:20>> We've even got historic comments from
12:21like Eric Schmidt, the ex-CEO of Google,
12:23saying brands are the solution, not the
12:25problem. Brands are how you sort out the
12:27cesspool.
12:28>> Google's philosophy, straight from the
12:29founder, is that brands solve the
12:31problem of content spam. A well-trusted
12:34brand is more likely to give good
12:36content [music]
12:37than a new website with highly optimized
12:39content. This is the core of Google's
12:41new algorithm.
12:42>> This is Tom Capper. He did a brilliant
12:44study for Moz where he found the sites
12:47most impacted by the helpful content
12:49update were ones with high
12:52uh domain authority metrics. That's
12:54pretty much like how many links you have
12:56and low brand authority metrics. So, if
12:59it looks like you've been doing lots of
13:00SEO, lots of SEO success, but I've never
13:03heard of you, that's kind of fishy. And
13:06there was a really obvious correlation
13:07to those sites getting penalized.
13:10>> So, websites that look like they'd been
13:11manipulating SEO tactics, but had no
13:13real brand authority, got absolutely
13:16crushed. This is Google's message. We do
13:18not care about your SEO tactics. We care
13:21if you're a legitimate brand that people
13:23trust. So, if you're a a business or a
13:26business of any kind of that matter, you
13:28need to focus on building these trust
13:29signals before you even try to optimize
13:32for SEO content. And there are hundreds
13:34of ways to do this. Get your business
13:36profile up and make sure that it's
13:38active. Get your name listed in
13:39directories and citations. Get backlinks
13:42to your website. And no, not just spammy
13:44links that you paid 20 bucks for. Go to
13:46real websites that also have a real
13:48brand authority in the real world and
13:51not just on Google and get them to link
13:53back to your website. We've also talked
13:55a lot about AI content in this video and
13:57this is not to say that it's bad to
13:59write content with AI. AI can be super
14:02helpful and I use it, my team uses it
14:04when we're writing content for our
14:05clients as well. But it also has to pass
14:08the filtering stage of a human. We have
14:11to determine if this blog or this
14:13landing page that's written by Claude or
14:15ChatGPT is truly helpful to someone. AI
14:18alone is not going to rank you on
14:20Google. You need to give it clear
14:23guidelines for the content and
14:24information that you're looking for and
14:26proofread everything before publishing
14:28it on your website. What it seems to me
14:31is that Google in 2026 is trying to
14:33replicate real-world information and
14:35authority in the form of search engine
14:38rankings. So, for example, if I really
14:40want to go for Italian food this Friday
14:42night, I'm going to maybe ask a couple
14:43of my friends where they've gone to in
14:46the past and they're going to give me
14:47information that they know is credible.
14:50I know it's going to be credible as well
14:52because they've been to these
14:53restaurants know that they have good
14:55service, great food, etc. These trust
14:57signals that we experience with our
14:59friends and family in conversation all
15:01the time is what Google is trying to
15:03replicate and is what you should be
15:05optimizing for if you want to rank on
15:07the first page of Google as well. With
15:09all that being said, I hope you guys
15:11enjoyed this video and got some useful
15:12information about it, especially for
15:14your SEO tactics. If you want to see
15:16more informative and educational content
15:18around SEO and around Google, check out
15:20the rest of the channel, but either way,
15:22thanks for watching and I'll catch you
15:23in the next one.

About this transcript

This page contains the full transcript of You're Doing SEO Wrong (This Hacker Proved It) by Stephen Baker, generated from the public captions YouTube serves with the video. The transcript has 2,838 words across 426 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.