How to Maximize Your SEO Campaigns

Thank you Matt.

We didn’t make any coffee, but.

I was. As a major. I was studying things like deconstruction. Taking apart things people wrote in to see what made them important. When I went to law school, we were when we studied using the Socratic method, which meant answering questions in front of a professor, usually involving homework from judges and their opinions. And it would also often involve breaking the judicial opinion down to nine parts questions by the action part. So there’s a lot of analysis of those opinions. And literary criticism was a lot of analysis of those. And that sort of flowed over into what I was doing with websites. And that’s the insight stuff.

I helped some friends. I was a service mechanic at a car dealership. He hated his job.

I was reading a book on how to run an incorporation business in Delaware, and it required that you had an open office and that you received mail. Those are the technical responsibilities of being a registered agent. You charge people, say, 100 bucks a month or whatever, to be the registration. You accepted the mail. You told them that they were being sued.

You said that’s a good idea, Bill, but. No Web site hand that. Oh, the store I bought it to amount to weeks book.

In two weeks. I’ve got a website. Wow. You helped. Just get another friend. Had a law office. Yeah. And my first friend moved into the empty room in the law office. Ran the appropriations business from there.

Ended up being a successful business. And I did that.

You did that for eight years.

First thing I did, one of the first thing I did wasn’t SEO.

Looking at the customer base.

Taking QuickBooks Mail merge.

That’s some customers. So we can build people.

Sales is priority number one.

The site looked like a Christmas card.

It’s simpler, but. I still started doing the technical stuff, looking at the patents and writing about them.

It was up to our WordPress site.

It took me 2 hours to build parts. 

We didn’t touch those and lost four.

There are like three areas of specialization in law that the Supreme Court allows lawyers to say they’re involved in the example of the patent law. Those are two of them.

But. But. Patent law means normally that you have some type of scientific background. And in this case, it was. We were trying to sell a business based on legal merits to people wanting to have the business in Delaware.

After three years of law school, that was really easy to list. Why you might want to choose Delaware.

The law doesn’t have a judicial framework for business.

For chancery. So these equity judges don’t use jury trials.

Opinions. So if somebody has a shareholder dispute, the chance to hear the case or write down an opinion. And this is an opinion that everybody throughout the U.S. can rely on.

It has story decisis. Is it n the plant-based on that judicial opinion?

So law school throughout the country. People studied Delaware law.

There are little tiny things like the disappearance of spinning globes on websites.

You. Yeah. Yeah, I know. You know, I used.

They’re the little things involved in SEO that you may not notice right off. Yeah. Yeah. There are information retrieval methods behind them. Like an inverted index.

An inverted index means that the words you query your words, probably show up in the document you’re creating.

So if you have a query that’s five words long or five words, show up on that document. Yeah, the words. And it’s part of an inverted index that the search engines track, in all those words, what they might be.

Now, something like PageRank. They’re also adding the strength of links to the page. Okay, that’s a little bit tricky. I know some of the toolmaker’s struggle with this.

Correlation studies, but. A link from a page like the home page of the New York Times is worth a thousand times a link from the home page of a local newspaper.

It just carries that much more weight. Yeah. So if you can find the weight of the links to a page with the number of words on the page from the first index, you get an idea of how much the page ranks.

Net your idea behind Netscape.

Okay. So you come up with the phrase like a featured snippet. Okay. You expect to see the words featured snippet on a web page. Absolutely. Bada of the burning index. Okay, so it’s a list of words that might appear in a document. Okay, we’ll put these huge index word traps. All of these words that might appear on documents looked up.

Absolutely.

As The New York Times is a popular website, it gets traffic and hits lots of links, pointed to it by many other papers.

When it links to somebody, people say, Oh, this is from New York Times.

So for.

We see the evolution. We saw in 2012 that Google came out with what they call a knowledge graph to know. They started identifying entities instead of keywords and said, okay, there are certain people, places, and things. People want to find out more information. They want to find facts related to those things. And. We see it all the time. We get millions of searches a day.

We have an idea of what’s happening here. We get all these query logs where we know what types of things people are searching for, what types of questions, and how they phrase things.

They’ve asked for certain types of things. We can answer these questions. We’re just matching the keywords in their queries with words on documents. We’re answering questions about identities and people ask.

I think it’s been going on for a long time. People are not necessarily paying attention.

We see it in places like Yahoo! Where people are asking questions about celebrities. We see it in Google, where people are asking questions about businesses.

The same reason why at one point, Tom was considering practising law.

Rich. It’s. It’s. Questions are solved by answers with their information filled.

They’re there. To use the old statement from the Wizard of Oz. Ignore the man behind the curtains, you mean? Debbie Harry sure does know a lot.

There are a lot of them. I’ve seen a lot lately that it’s just an old technology called Lanes Matic indexing, which is 88.

Semantic. And it is panning by a Bell Labs for you to drive past it. On time ago, the labs in New Jersey were developed for small static databases.

There’s a patent related to it. An example of the database made out of eight books.

The patent says every time you want to update this index. You need to start from the beginning. Every time you add something new to the index. You need to start over again and re-index.

The Web changes too quickly, people. Ed commented people to lead start people add new pages. Yeah. Create stuff to put on. It’s not small. It’s not static.

So. So chances are Google’s released paper saying we don’t use make indexing. Yeah, we use something that’s called probabilistic weighting, semantic indexing for advertisement

Semantic indexing, which means it’s just a statistical approach.

It doesn’t just encompass its more static database. It’s better than that.

The net all search for Google through signing Google.

The idea behind it tends to be to exclude other people from using your intellectual property.

So one of the issues with that is sometimes Google will cancel things and you’ll see something similar to that from Bing, and then something very similar come out from Yahoo! 

That is exactly the same. The words are different. They run a little bit differently, but they do very similar things. Now, Google probably has a head start on the other two because they’ve developed more stuff in the past few years. Google’s taken away some of the search engine talents from the other two sources.

Sometimes you can read Pat and say, you know, this is in play, this is the news, and something you and say, I know who the heck these guys are. I’ll go look up the authors. And Lincoln said, Oh, you were the head of core web search in Google for 14 years. You must know.

I’m trying to do the research behind this, but this guy was with some machine learning initiative for years.

There’s one guy who was involved in what was called the annotation framework.

Google and went to work for Foursquare. But not.

But before he was in charge of 13 other search engineers of Google who developed patents, then developed into the knowledge graph. 

It’s following the history and reading about some of these people means that you have to go beyond the patent walls.

That’s a point behind the pattern. Yes.

You want the.

You’re working on a lot of stuff that’s somewhat related.

All your work for ten years. Can’t use it.

It’s like the LSI is based on similar variable testing.

There are some limitations.

It’s not quite as complex as some modern stuff that’s doing similar things like word vectors.

Which are being used, in some of the language models.

So. So King is to queen as man is to woman.

That’s their head in certain directions.

You hear the meanings of those words and words around them to keep that in mind, how it fits. There was a linguist, a computer linguist named John Firth. You know the phrase you shall know a word by the company keeps.

In just the way words are connected to each other. When you hear knowledge craft.

When he refers to the word knowledge in a patent, they’ll define it for you sometime.

So they’ll tell you. Knowledge means relationship.

So. So now scrap is a relationship graph. It’s a relationship graph.

All these entities are related to each other.

There are different ways to implement the knowledge graph. So the one one that I would think about using is I would look for what is known as frequently totaling words. Okay. Right. For the same thing. So you have a page that’s about the president of the United States is going to complete meaningful phrases on it, such as Oval Office or Secretary of State or so on that. If you take them, look at them all together and say, Oh, that’s probably a page about the president of the United States.

That’s the appropriate semantic topic model.

So there’s one if we go all the way back one of the original ones from Google is actually Stanford’s patent that was assigned to Stanford. Okay. Once Page developed it, it was a page, right?

Each requisite a way of. Better sorting pages. They were hiring and worked a certain way. It followed something that page called the Random Search for the model. And one of his employees, Jeff team, used a lot of stuff dealing with how Google calls pages, updated it, and turned it into something called the reasonable serve model a reasonable, reasonable search for model. The reasonable model was finding a link on a page and asking yourself, what is the probability that someone might click on this link to go to another page? The higher the probability, the more patrons it’ll pass.

That includes both external links from other sites and internal ones from your own site.

And it could.

Play it’s partly because it’s one of the features, one factors that determines how pages rank in search results.

Google’s been saying in patents that some links are losing value.

It’s from real time results. So in tweets, earthquake. Chances are there’s an earthquake.

And nobody’s had time to link to that tweet that says earthquake.

Our news story, something that’s like a brand new news story. Empire State Building catches on fire.

If it catches fire, it’s it’s burning regardless of whether or not somebody links to it or not.

I was in Virginia earthquake hit 60 miles south of me canned beans rolls off one of my shelves hits kitchen floor. I knew it sounded like somebody ran across a tin roof and sound like something. Right. No legs.

It was definitely an earthquake.

Stupid enough to drive 60 miles south and drive through the town. Drive past a sign that says nuclear power facility.

Ask myself, why am I here?

It’s not a cheat. Yeah. Say you start with a new client. You ask to talk to people who work for the client. You talk to the salespeople. You have to talk to the customer service people. Okay. So contact with customers on a daily basis?

So you can ask what questions are people ask and what are they? We we get a lot of the agency I work with. It has been a number of people at this point in time.

It does. What? What are known as creative working campaigns or creative content campaigns.

Ask questions like that. We worked with the limo dealership and in the limo driver. Drivers told us where people were getting engaged in being here. So we created an interactive map. People can go to click on get photographs. So those place where people get engaged. This was all proprietary information from a limo driver. It’s absolutely. There is stuff people are furious about. We get lots of visits. We get lots of people who linked to the site who talked about the site. We’ve had campaigns like this, not necessarily that one where people on radio stations or TV stations talked about them.

No, it. It did. We worked with the jeweller.

And one of my coworkers visits Reddit a lot.

He met somebody on Reddit user hydraulic press website and YouTube with 500,000 subscribers.

He talked the guy into taking one of these diamonds from jewelers and crushing it. With his hydraulic press.

Right. Wow. So because 10 million views in a matter of months.

It had a link to the jeweler.

It’s like they crushed it. Wow. I’m sure he asks. Question How well does a crushed garment burned using stone torch? And these videos that.

It was a 1.3-carat diamond. So it’s worth more than a thousand bucks.

I would think so, too.

Writer Data Information. People don’t know this. They won’t find the answer.

Yeah. So one of the first tools that I gave somebody was my first in-house IFCO job.

Guy who incorporated businesses. You know, Pat, I said anybody asked you any good questions? Write them out. And then written answers for our website.

It was a Delaware intercourse nature. It did really well in terms of incorporating businesses.

Yeah. My friend. Who is his search mechanic job?

You’re mentioning like you like incorporating people.

He sold it to his partner who had signed somebody else.

He didn’t have to be a mechanic anymore.

You do have tools like training frog that you call website.

And they’ll allow you to develop. Visual graphics about how you sort of connect.

You create flowcharts.

Show how pages are connected and identify which pages you want the most links to.

Go back and create points.

It’ll show you the visual aspect of it.

It proves some things are true in private, which lets you do the work. Yeah. I want to show you our actual short quotes.

Internal linking does have lots of value.

There are pluses and minuses with plug-ins.

Is relying on them too much. They can end up slowing down your site.

Too Bad.

If somebody enjoys going to your site and has a good experience there.

You know, before other people, you say, this is with you more likely than not. Yeah.

And conveyor belts in factories sort of things that.

So we’re trying to set it up so that it is emulating Google boss.

Before screen from there was an older tool I used to use called genie which was it helps you identify broken links and pages to. Oh yeah it, it will identify architectural problems of the website. Sometimes you have problems. Such as spider traps.

Just disappear as it gets.

Duplicated even though they shouldn’t be.

And then it’s worth trying.

It’s close to £200.

My birthday happened one year that just came out. I said, Yeah, that’s my birthday present.

Sort of tethering? Couple of things there, but there is a. Links are definitely part of the value that Google looks at when it shortage.

However. Chances are Google’s diminished the amount of weight that carries.

So it doesn’t necessarily have the same amount of weight that it did ten years ago.

Now we’re saying the real-time and news results. Not every backlink.

People from Google have produced documents where they’re talking about high quality sites and oh yeah, one questions that they have associated with that is, would somebody come to your website and use it? Would they feel so? And that’s that’s if they would feel safe using pressure on your site. It’s a higher quality site.

Okay so often. We look for skills, things that we can employ. He would search for an employer in their workers, like a college. It shows that they’ve accomplished something.

There’s some type of history when I’m recommending to people that they study certain things to become CEOs. One of them is building a hobby site. Maybe I’m necessarily mission-critical, but you can learn from you can put your search console on it. Now you have an idea of how works within the framework of the web of things.

You can start with a hobby.

Something more serious.

That you can build a hobby site is a plus.

There are lots of moving parts. It doesn’t.

Obscene sites sexually dropped by one person.

I work with the company that has around 70 right now.

It’s why I’m recommending people don’t have it strongly. Usually, it’s the people who are interested enough that they’ll do everything.

Usually, somebody needs to come to you with the idea that. I said I would be helpful to them.

And then when I do one, I’ll try to cover everything I can cross every T and dot every. I want to see what works well. What doesn’t work well?

What I might recommend changing.

I usually use Green Frog first and create what I refer to as a content inventory.

It tells me what exists on that site, which phrases people have chosen, and which they’re using in titles, descriptions, and headings.

Multiple directions. For a little while, FuelWatch was grabbing. Phrases from reviews express sentiment and information from those reviews and knowledge graphs. All these statements about sentiment, dealing with products or services from inside.

And that would give people who visited a Web page that had knowledge about for certain business, information about that business and how it works. The idea was to create the general impression that sometimes people will prefer to as a brand.

This is part of transformation. Again, it’s like when Google is updating stuff like medical sites, they’re using a slightly different classification approach.

So. There’s a lease one patent from Google that is referred to as website representation vectors. Which uses neural networks to classify the language on a web page to give people an idea of who the speakers are. If it’s a medical website created by doctors, was it created by students at medical school, or do they have an idea of who the answers are, who the people decide which questions they should probably be answering.

Can get good sense of which search results should come. From which pages? From which sites?

And to some degree is speculation. We don’t know. But it seems like that’s probably what’s going on.

There are some medical sites that are written by people who are. More popular scientist.

Like. Like. Show off the benefit of walking through the beach barefoot. Okay. Cool response.

Maybe it isn’t the best way to cure it then. Things like. Like heart disease.

No websites that do things like that.

I can hit one quick.

Okay. So we had a we had a site that was having problems appearing in top stories in Google searches.

I had looked at the website and noticed that in the old tags, all the entities being referred to in those pictures were being mentioned twice. So. Oh, that cured stuffing.

Get rid of one mentions. Name these for each of those. They got rid of them and started appearing in those top stories again.

They were stuffing the keywords. Right.

Like looking at a picture of Matt Damon. You say this. Matt Damon. Matt Damon.

Text. No reason to name twice.

Every day was Christmas. Yeah.

But you name it, he wants and that should be enough.

I have a law degree now.

It’s diminishing.

So when Google identifies analyzes videos and they’ve had a video of a movie with the car scene, and it’s like Fast and Furious.

They know there’s a car chase it nowhere. And on the video page does it say there’s a car chase? Google knows there’s a car chase movie clip about the Fast and the Furious. 

They they go through the audio of the video. They analyze it. They find the car chase. Someone searches for it. They get to see the car chase.

From the image analysis. Right.

Like I was saying, Google is getting smarter using neural networks and things like that to classify pages. They have a better idea why not just flow to with the description of words Google might use on a YouTube page to describe a video? Yeah, actually, they’re watching the video. It’s like Google’s machine learning. Talked about in the past about watching videos to identify which cats were and they were the ones that really identify which videos had cats with cats were, which sounds they make

Or machine learning.

I’m not sure I can share them, but. Yeah

Fair enough. Well, I’ve seen certain things, certain technologies that are described in patents, but nowhere else worth.

And when they. When they do one wonders, natural. Whether that shows that supports that provides evidence. So this was a good idea, this one.

So there, people.

For people who say to complain that there’s no actual eat score.

It is as a concept is something that’s related to search results and having doing well within means that you probably. Meaning those little algorithms that result in showing off expertize, authority and trust and not getting some flatline eats for that tells you all you’ve met. Eat today. Congratulations.

The thing is.

Those. Those algorithms exist.

When? When your website shows that there are experts behind this.

From the experts behind it.

So when I was doing it solo CEO for comments one thing I would give people was a guide. The Stanford Web portability guidelines, are one page and it’s ten items. And it’s things that you can add to your website that show that there’s an actual eat or a website. So, so showed that there’s a real organization behind your website. You can put information about your business on your website, and show you’re expertized and the people who run the website. So you show. Degrees. And you show. You talk about things that you’ve done. Experiences show on your show.

Don’t just tell people what you show me. States, show me that you’ve done this before. Show website exists that does these sorts of things. So you’re showing off eat, but you’re not yet showing that you’ve met these algorithms. Rules come up in older patterns that show they know how to do these things, that show that you know how to do these things.

Credibility guidelines.

I’ve been doing this to clients since it came out around 2004, so I’ve been with them for years, long before, long before there was me, long before there was a very rigorous fashion rule. The funny thing is, they’re both from Stanford.

And Lawrence Ross Page, and Sergey Brin were both students.

And they were working on a project, working with the professor who wanted to battle Japan for a year, but they got a trailer to work at. Yeah. So they’re doing some electronic image stuff, and they’ve started doing the work behind the trailer and found more pieces. They said We’re not going to do this work on the web. So they saw working on the Web instead. The guys, supervisors in Japan watching samurai games and maybe doing some work, get in touch with somebody once, once this as the project’s going to go off real good.

Yeah. So they developed stuff like PageRank. Filed a provisional plan for PageRank that described how it worked.

Sure. Web pages. I know what design thinks. That is the sabbatical, professor. What they’re working on after the fact.

 It’s assigned to Stanford as an exclusive license to use. It’s expired since then.

It was dropped in 1998. I paid last 20 years. So it expired in 2000.

So the recommended approach to that is what we call key performance indicators. 

What is the best way of showing off a successful campaign when it’s okay?

So that’s what we do, what we’re working on. One of those things like how many I draw prisoners it takes to crush a diamond.

I have some idea of how much is good traffic to the website. 

For a business that would use cars and dealers, you put the type of car you’re looking for in the past, the dealer of one of their dealerships. It would tell you when they had the car.

So, we developed a database of the most popular cars in each state.

It did well. It had like. Three hundred thousand views in a couple of months.

And that was one of our keeps.

It varies. It involves the station and so on. But yeah.

I. The whole black and white have been just about all western.

It’s not necessarily something that. I think that too much effort was put into these programs at a scale well.

On the watch works but. You can often do a lot of stuff that involves a lot of effective work.

This involves questioning, asking people, salespeople, etc. One client we worked with was a. Chip manufacturer.

Different chips for different customers based on what they’re trying to create. Hmm. We had to have some idea of what their customer base is. Yeah. We helped them create case studies because they wanted to resell different ideas behind different chips and different manufacturers.

So. So we would drop case studies so that people would come along and say, this is our case study. This song applies to us. It’s the one nearest to us.

And that was a strategy in that their customers were self-identifying.

To know what their potential client can do for them.

You work with somebody who has a good department.

Helpful health benefits program. I’ve probably lost £30 in the past four months or so from all the stuff like that from me. Well.

I have a stationary bike. I will ride 40 minutes a day.

It trails around me. Uh, I’m in Southern California. It’s always sunny here and there. Always take walks is nearby. Looks good. That’s with a five-foot-wide three-mile trail. And being able to drive out there and walk around is nice.

And mention beach to and walk along beaches real nice.

Vincent De.

Fair enough. Let’s see. I like your coffee.

I like the areas of the past to now.

I go to the beach. I’ll go to Coastal one on one along the coast in California here because there are a lot of old diners and shops and things like that.

It’s fun.

Become lazier, like my dad.

Now you talk about learning styles.

True. I think it varies depending on what I’m learning.

Lots of things we like to do and not necessarily the best.

And that I make myself a cup of coffee.

You too. Thanks. Bye.