Google Webmaster Trends Analysts Q&A Interview

Welcome to a special edition of Poly Mondays. Today we have some guests that probably you don't know nothing about because you have been living the last 30 years with Dr. Manhattan in Mars or something. But just in case, just in case, and you are back to Earth, can you please introduce yourself? You are Googlers. Everybody knows that you work here. But what do you actually do on a day-to-day basis? I'm a software engineer.

I'm John.

Nice to meet you, John.

I work at Google. I try to connect the search engineering side with people who make websites so that we can kind of understand a little bit better where people's problems are, where people are struggling, and try to find ways to make that easier.

I'm Lizzy. I write the documentation. So I work with feature teams at Google who want to design something for search and then help them communicate it to everybody else. Perfect.

Okay, thank you. You're the—

yeah, I'm Daniel.

I work very closely with the Search Console team. I'm based in Tel Aviv where the Search Console engineering team is, and I work with them with the product and also like doing advocacy for search.

So if anybody has any idea of a feature or a request for the Google Search Console, is it you, right? Exactly. Good to know.

Good to have that.

You want them to use the Submit Feedback in Search Console, George.

Exactly.

Thank you, Martin.

You're welcome. Or the forum.

We'll have a link to the forum in the description and you can ask any questions. Twitter, forum, and feedback. Thank you.

Perfect.

Martin? So hi, my name is Martin Splitt. I am also a Webmaster Trends Analyst and I am working closely with the rendering team. So I look into JavaScript problems, I am in the mailing list for JavaScript Sites. I answer questions on Twitter. You can also ask us questions throughout the forums. And I show up in John's office hours every now and then when there's technical questions. I think on Friday we have more technical questions, so I'll probably be there as well. And I, yeah, I work with the rendering team. And I am trying to bridge the gap between SEOs and developers. Because I think if we all work together, we reach better results. And I work a lot with Lizzy so that we can have documentation for both SEOs and developers to tackle the more technical terms. By the way, this documentation would be crap without Lizzie.

I see your work there. That's nice.

Hi, I'm George.

Thank you, Barry.

I'm also a WDT. I think I'm the second oldest after John. I typically work on specing with specifications with Lizzie, like the Robots.txt standard, for example.

Open sourcing that one as well?

Yeah, blah. Then I work quite a bit in infrastructure, search infrastructure, but there I do coding mostly, like I'm changing things, and I'm also doing lots of debugging, ranking debugging, when Something looks fishy, like someone reports that, I don't know—

Poltergeist type of cases.

Yeah.

Interesting.

So entering more into topic, I wanted to ask you, and please feel free to share your thoughts, which are those top trends that you think SEOs are overlooking right now, that people are not talking enough right now and are important?

I don't know, robots, text. What do you think?

I was thinking about images rather.

Images, okay.

Media in general. Like this year and also past year, I was doing quite a few keynotes and panels on media in general. And that's not because I'm bored and I hate my life, but well, that's partially true, but also because We noticed that there are very basic things that SEOs don't do. And if you think about it, like, what, like, in your SEO process, where is like putting alt attributes on images?

Yeah.

I looked on many SEOs, like, this is your cheat sheet, like, what should you do? And the vast majority didn't have like, like, look at alt attributes. And we also see this in indexing, of course. Like, when we start talking about just do that one thing, like put alt attributes and just put some text that's related to the image in the alt attribute, we actually see in indexing that the numbers are going up. Like, at Webmaster Conference, we used to have these talks quite often, and every single time we see like an uptick in image indexation.

So to give a little bit of context of why might be that, right, also from the SEO side of things, it's not that we don't know how important it is to have an alt description, even for accessibility matters, right, beyond SEO influence. But sometimes it's because there are other very critical aspects to be optimized first that need to be taken care of. So with this, you will say that we can expect that image will earn more or will start being used more within search, or why do you think?

So, I think both. One of the main issues is— well, not main issues. One of the things that we see is that sites that have images don't get the expected traffic from Google Images, traffic that we would expect it to get. And when we start looking at what's wrong, what's happening, do we have a ranking issue or an indexing issue? Well, we do have an indexing issue, but not caused by us, but caused by the site itself. And usually the issue is with alt attributes, and then there are other issues as well, like lazy loading that messes up indexing images. But that's a different—

Lazy loading is incorrectly implemented, that is. If you do it right or use the new native lazy loading in Chrome, you don't have that problem.

Right.

I also see the images as one of the things that pops up all the time with mobile-first indexing. So that's like we're shifting all of our indexing to the mobile version, and a lot of people still focus mostly on the desktop version. And things like alt attributes, even the positioning of the images on a page, things like the headings on the page, titles, these are all things that people often forget on mobile because they're just not as visible. Like you can't Like, when you're on your phone, you don't look at the source and it's like, is there an alt attribute here or not? But we need that for indexing. So, that's something where some of the sites that haven't been shifted over to mobile-first indexing, we do hear that they have issues with the images.

I have to say that a few people who did rely, especially publishers, right, rely on image traffic. The way that image results were— the display was changed. They have to stop relying on that traffic too because at the end of the day, they they didn't get that much traffic after the way the images displayed were changed. And so—

But I think that's something that depends quite a bit on what you're trying to achieve with Image Search. And I think that's something that a lot of SEOs don't really have figured out that much yet, in that a lot of times people just see images and indexing like you see it as a count, kind of like you have so many pages indexed. But what you really need to think about is how people might use visual search to come to your pages. And that's kind of the point where you have to consider, well, what kind of images do I have on my page? How can I be at the right place at the right time? Or do people just not search visually for my content? Like, if you're writing SEO blogs, I don't think anyone's going to image search and searching for SEO graphs or something.

Basically, I think this change, or these changes, will help websites get more qualified traffic, more targeted traffic, traffic that actually wants to be on that site. Because if I have a picture of apple pie, for example, then there might be several reasons why I would search for a, for a picture of an apple pie. Sometimes it might be just ideas like how it should look like, But other times it's like the recipe, right? And in that case, for example, it's quite important to see like how it will look like because perhaps I want to recreate something that my father used to bake. And based on the image, I can identify the type of pie. And also the badges that we nowadays do, or some images get for products and recipes and some other things of that nature. That's also what helps the publisher get more qualified traffic because people know that they can buy something on that page or they can find a recipe or whatever else the other pages are.

Okay, whenever an image will better satisfy the user need, of course, and make sense to have an image format for that information.

If you think about it, that's our goal. Yeah, like we want people to find the right information at the If they are looking for a recipe of an apple pie, then we don't want to show them something that's unrelated, like just apples, for example.

This, I have to say, if you are using any usual platform CMS, should be pretty straightforward to set, even programmatically. It's an initial step. Then another thing is quality, right? Because I understand that, for example, to get an image to be shown to be taken into consideration. For example, for Google Discover, the quality should be a certain minimum, right? That is a little bit trickier, right? Because there's always the back and forth of the image cannot be too big, but it also needs to be certain quality, right? So—

It depends.

Indeed, it depends. A lot of priorities here. Realistically, I have to say, for me, it always made sense that because a lot was said about voice, right? Voice search, huge topic in the last 3 years or so. It's almost 2020 and I don't think voice is even a word anyway, but I was going to ask you on the other hand, which were those trends that you would think that SEOs are focusing on too much and they shouldn't? So for example, I was mentioning voice before, that for me is too much of a— you don't think it is? Do you really think it is as important?

So I do think that people think too much about it because voice search essentially is like— there's not that much to optimize there, right? Basically. And same goes for BERT and RankBrain. RankBrain and other— they are on the same page in the terms of like SEO is overthinking them because in each case you would just create content normally. I was saying this many times and others were saying this many times as well, just create content and make it sound natural. That's what you need. Create a content, create a piece of content and write it in a way that it sounds human.

I mean, that's just like a writing tip in general.

But that's what you—

listen to the writer.

Yeah, exactly right.

Basically, that's what you need. And we had that in the SEO guide as well for since 2018. 6, 7, or whatever. Same goes for titles, same goes for like, how many characters do you need in a title?

More than—

it's like, write it. Yeah, write the title and make it relevant to the page.

Well, no, but it is true, 100%. Although I have to say too, from an SEO perspective, you really want to make sure that you make the most out of the available characters that will be shown to the user in search results, right? And you don't want them—

And you know how many of those are? Because on the phone it might be different from—

Yeah, they change. It depends. It depends again. They change based on the device.

For instance. So if you have a relevant title, we might just use it. If we think that we have a better suggestion, we might rewrite it to the query. But fundamentally, make a good title. for a human being and the rest will follow.

I think the tricky part is, on the one hand, SEOs research a bunch of keywords and they try to target those keywords and then they try to just artificially include all of those keywords in the content, in the titles, in everything, which often leads to that situation where it's like, well, I have 10 keywords and I have 1 title tag.

I will I have to say that I think that most of us who work, I mean, at least with real businesses, right? I mean, real companies that make money, that have employees, etc. It's not that we just select random titles that we might think that could be connected and end up including there as much as possible, but actually those that will connect and satisfy the user search need and will address the user question at the end of the day. I think that most of us already follow that type of thinking.

Yeah, but I mean—

And not only purely because of SEO, it's because of user satisfaction, click-through rate, and conversions from those users at the end.

Yeah, but I mean, that's essentially what leads to voice search. If the content is readable, if it's understandable, if it matches what people are looking for—

It should be already okay.

That's kind of what you need.

Important clarification for everything that we said. before and probably what we are going to say. Like, not everything applies on every single SEO. Like, some SEOs already do pretty much everything as they should, and they are not focusing on keywords. They are writing naturally. They are doing image alt tags, alt attributes. And they are doing well already. What we are hinting at is that there is still a large section of SEOs who are not focusing on image alt attributes or focusing too much on— but it doesn't apply on everyone.

There are exceptions.

Communities grow. This is large, but there's plenty of people who are doing well, plenty of people who are doing it not so well. And within the group of people doing not so well, there's many different ways of doing not so well. Unfortunately, we have a tendency to focus on those who are not doing so well because it's our job to help them do better.

Yeah, of course, you need to target those who need the more help, where there's more room to grow or to fix or improve at the end of the day. Makes sense. So, any other area that you think that people focus on too much and should just go?

Performance scores and tooling. Like, performance, website performance is such a It's a process rather than like a state or one thing you do. And we have different tools for slightly different purposes and audiences. And they have to— like, performance is such a huge topic that we have to somehow break it down so that you know, like, top-level, 10,000-foot view of what is your site doing in terms of performance. And then like you can drill in. But the problem is when you have this 10,000-foot view and you have one number, let's say like Lighthouse score. Then some people are like, I have a Lighthouse score of 90 and I'll spend billions of hours now to just like get from 90 to 100. That's not necessarily what you want to do. That's where you overfocus because you're not looking into what is the actual problem and how are you actually doing. Because if I see like my score is 90 and then I see every metric that I care for, and that depends on what my website does and what my business wants.

Mm-hmm.

then I might be fine and there's nothing to do here. However, if I have a score of 90 and the metrics that I care for are bad, then that means that I still need to have to do some work. Which does not mean that I can just ignore a Lighthouse score of 5.

You can't.

But it's a bit of an indicator of how are we doing roughly. And then you look at the— you understand and look at the different metrics. And Lighthouse is a tool geared towards developers. So you don't have to understand every single metric, but it's something that you can give your developers and you have to trust them to make the right decisions based on the data. WebPageTest is a mix of developers and SEOs, I would say. PageSpeed Insights is less technical, but still a little technical. And then we have like TestMySite and a few other tools that are even less technical than that.

Decision-making.

Exactly. So like, you have to choose the tool. And I'm being asked like, therefore, so Martin, which is the one tool that we should use?

And I'm like, well, The absolute truth metric, that score that I should trust. Exactly.

And I'm like, what are you trying to accomplish? What are you trying to measure? And most of the time people don't have an answer for that. So they are over-focusing on these numbers and also sometimes even smaller sub-aspects of it. I recently got into a tweet storm of people asking me, what does this mean? Because there was a slide in Mountain View saying that at some point when you excessively consume CPU with JavaScript, We're going to cut you off. And then they're like, what's the moment? What's the timeframe? What's the score? How do I measure this? And I'm like, look, if you look into your analytics or whatever you have and you see that people are annoyed by your website and bouncing off after a couple of seconds because it hasn't finished loading, that's what you need to look into. So like it becomes a user problem and a customer problem by that long before it becomes an SEO problem. So when you overfocus on that, you are missing out on the bigger problem.

I love that you mentioned that because at the end of the day, SEO should be like part of an outcome.

Yes. Usability, accessibility.

It's a means to an end.

Yes, indeed. User experience, excellent. And context matters. Metrics are just references and what is important is to take context into consideration. That's great, thank you. So, the next question, what's the most overlooked useful feature of the Search Console? That gives me— people are not making the most out of.

Rendered page.

Which one?

Homepage.

Help center. Oh, thank you.

I should ask the experts.

The help center being people asking or?

No, it's okay.

Okay.

We will get your answer.

Well, actually one of a really small feature that I think it's quite nice is on the performance report that you can actually in all the tables if you click one specific row, it adds a filter to all performance reports. So I think that's a pretty neat feature. I was just talking about it the other day for rich results. So when measuring optimizing performance with rich results, if you're looking at one specific type of rich results, so for example products, you can just go to the Search Appearance tab and Click on that row, and then it will add to all the performance reports. This feature that you can see all the queries that got that these specific rich results got, and all the pages and countries and all the. And I have a video on the Search Console training series that you can watch on YouTube on the Google Webmaster channel.

That's nice.

That's nice.

YouTube.com slash Google Webmasters. Subscribe.

It's like how you get to.

Exactly.

Martin, you had a— Yes, so in the testing tools, if you're testing stuff, every SEO I know looks at the screenshot and gets hung up with the screenshot. Look at the rendered HTML. That's what you want to look at.

Okay, that realistically is because visually we want to see if something is loading and rendered, and we only get the viewport, first viewport thing.

Like, scroll, I want to see No, you want to look at the renovated email.

Okay, trust me. Okay, that's a good tip right there. Perfect. Here, what is the most common question that you get all the time from the community?

How can I make my website rank? Exactly, that question is number one probably. Yeah, like what is the most frequent question you get is a question that we get asked quite a lot. Really?

Oh, really? I'm sorry, I thought that I was being unique here. Sorry, that's the first question. What is the question that nobody asks you to, and you wish people will ask?

How are you?

Let's remember that they are people. They are people.

You're gonna get so many hugs.

This is not like being a model.

I'm just kidding.

Everyone actually asks that. Everyone's nice. Yeah.

Everybody's nice.

To me.

Yes, Barry. Um, I don't know. How do you make better content for your users is a question that I have not really been asked much, and that's like the core question, maybe.

But what would you say? Like, someone depends.

It's true.

Maybe Lizzy has an answer to how to write better.

No.

No, no.

Yes, yes, it depends. So many factors. And like, what, what is the thing? That just sounds like a nicer way of saying like, how do I rank number 1?

Fair enough. But I mean, in the end, write for humans, have descriptive text, have short sentences, explain concepts well.

I think I have a Python script for that.

Use active voice.

Yeah, that's very hard nowadays, right, to automatize Yeah, it's like, what did you say?

How many words do I have to put in a sentence?

5.

5. Okay.

Okay, well, 5? Oh God, no, people will—

please. Oh no!

Of course not. It depends. But like, if you— there's a bunch of documentation out there how to write better. Um, and yeah, it's like, use active voice, have reasonably long sentences. describe concepts when you introduce them, all that kind of stuff. A lot of that I learned from—

There's a classic book as well.

There's lots of books about this.

What is the one resource that you will recommend to SEOs to read about good writing?

Well, the one that I use is the Google Developer Style Guide, and they've made that publicly available like 6 months ago or so.

I will look for it and I will put it below.

It's useful to me because I write developer documentation, so it would It's helpful, uh, but being consistent and finding a style guide that you like, uh, and not changing.

Yeah, so either hire a copywriter or, uh, learn to write well.

Nice. Oh, that, that I didn't— I never expected that answer, so that is actually very nice.

You're welcome.

What is the most common reason, cause that you see of Drop rankings. Whenever someone comes to ask you and you actually take a look at the website, what is the most common one that you see?

I don't discuss ranking.

No?

Or drop traffic in general.

I don't know what Gary— Gary will probably say something smart. But the one—

No pressure.

The one that I most frequently see is when a website is kind of mediocre, which is something that— it's sometimes really hard because I realize a lot of these websites are labor of love and people have been working them for years. But when you look at them as an outsider, you're like, well, it's kind of like all of the others of the same kind. It wouldn't be something that I'd be able to take to the search quality team and say, hey, we should be ranking this website for this query instead of these other ones. Because it's like, well, it's all kind of the same. It's not really outstanding.

Yeah. So I can imagine that is more tricky, right? It's easier when you're in conversation where you're saying something and you're like, I don't want to Yeah, I can understand. Also, if it is something much more straightforward, like all pages canonicalized to the homepage or robots.txt blocking the mobile and tablet.

Unfortunately, those are not very common, I think. It happens even with the best SEOs sometimes that they come to us and they are like, you dropped our site from search! And it's like, well, yeah, you have noindex on every single page. And they are one of the leading SEOs. But it happens, like shit happens. But it's not common. It happens, but fortunately it's not common.

It's more common to see like a website that doesn't necessarily provide like a good experience.

So actually I wanted to say the same thing that John said. So in essence, it was smart. There are things that you can tell people. And like when I'm on stage and someone asks me about a specific website, site, then I will not be able to help them for various reasons. One is that I'm not able to look at the site or compare. But many times when people catch me after the talks and we sit down and we talk about the site, then I can give them concrete tips about what they could do. And it's not ranking tips, it's more about comparing websites and let's say that there was a core ranking algorithm update and some site went down and some other site obviously went up. And when we look at that, when we compare the 2 sites, the one that went up is significantly better. But not only content-wise. There are many little aspects that you can optimize and take away. And if your content is actually better, but overall you are not giving the right hints to people, and I said people and not ranking signals, that you are an authority on a specific topic, Then how would people know that?

Mm-hmm.

Like, if I have applepie.com, just people would not necessarily trust me just because I have applepie.com when it comes to apple pie recipes.

Mm-hmm.

But if I say that this Michelin-star chef recommended this recipe or wrote this recipe, Then that already has a different ring to it. If I can also link out, for example, to specific resources that people can use to dive deeper in what kind of baking soda to use, for example, and comparison on baking sodas, even better. There are these little things that help humans understand the content and the authority, or recognize the authority of an author or a site. I said author, I shouldn't have said that.

No, but you know, interestingly enough, I like that you say an author because actually the principles that you are discussing are the ones that you also take into consideration when you are writing a book.

Yeah.

When you're writing information in general, right? Whatever the output or the format, at the end of the day, you want it to be relevant, useful, meaningful, as much authority as you can. You invite people to contribute and you refer to other useful resources. Makes sense. Yeah.

I realized something. Did you ask about like what we don't like?

Because I want to say what's wrong somehow.

SEO articles about SEO experiments that were run on 1 or 2 sites For 8 keywords.

Okay.

Sounds like a business survey.

You're showing very specific things right now.

That's why I don't know. For the period of 1 week.

Yeah, just to make sure, this is something that you dislike, not like.

But this happens. Like, it happened years ago, it happened, it happens all the freaking time. And it's so frustrating because like, you don't even have a proper methodology. Like, like you, like nothing is good about your study, but you published it because like it's sensationalist or worldwide.

Or arcane test sites where like, yeah, I mean, you have a pagination with 20 pages and all of them is lorem ipsum dolorum sit amet. And Google didn't index it. Google only indexed the first 2 pages. And I'm like, yeah, we did.

Sorry, sorry that we tried.

Okay, makes sense. Of course, do not Making conclusions out of tests that are not necessarily following a fully reliable type of process or methodology.

And it's definitely possible to run proper studies, like good studies, because people like, I don't know, like Eric Enger, for example, they were running studies, like really, really good ones, and very often, interestingly, their results were the same as Hmm.

Okay.

But you actually have to invest in it, and, and you actually have to have a proper methodology and know how to actually run studies, because it's very much statistics. And when it comes to statistics, then typically big numbers win.

And now to wrap up, what are your predictions for 2020?

We will have blog posts about predictions for 2021.

That can be for sure.

It will happen.

And yeah, that's— I think predicting the future is very risky, so we should not do that.

Well, we can say what general things we would want to— general things we want to focus on. Like, I really want to get to areas of the world where I know there is a need for more information. Because we see this all the time that certain areas of the world are not— or asking pretty basic questions, for example, which generally means that like they are novice SEOs, for example, and it's easy to fix it. Yeah.

Yeah, that's why we're doing webmaster conferences pretty much all around the world. We're trying to bring that into more areas as well.

That's nice, that's really nice.

We're gonna definitely look more into ecommerce as well. And I want to learn more about what we can do to support people building JavaScript websites. That's an ongoing, ongoing thing.

It's annoying, huh?

No. That's definitely something that we're gonna work on. Yeah, what else?

Search Console.

Search Console.

Okay, so all the feature requests?

All of them.

Crawling, API?

Everything.

Google Data Studio integration?

Continuing to build documentation. That's also something that we haven't—

videos.

Yes, videos.

Yeah, you have invested lots of videos. It's funny that you didn't mention anything regarding speaking about trends and importance. You spoke about images and you didn't say anything about media.

Media.

Media, okay.

We just focused on images because that's a quick win and pretty much everyone has images. Not everyone has the capacity to produce video, but generally what you said applies to all media.

Okay, that's—

Another thing that is on our mind is how can we communicate better when we have issues? Like when we have indexing, crawling, serving issues, how to communicate better in a more structured, more proactive way. That's something that preoccupies me particularly a lot because we have to connect to different parts of infrastructure to make that happen.

Yeah.

And another topic that I just started digging into is how to Or helping also those who have special needs when it comes to consuming content.

Accessibility.

Yeah, because I was in Tokyo and I listened to a very inspiring talk about being a software engineer when you can't hear or talk. And how, how that person was overcoming their challenges and how we can help them more with random things that we never even thought about. So yeah, I thought that was inspiring and something that we should look more into.

Yeah, you could also even consider Googlebot to basically be a user with accessibility needs in a certain way. I mean, yeah, yeah, a user with special access needs. And that's something that Ilya Grigorik, who is in our larger group, phrased that way. So I took that phrase from him, and I find that quite matching, quite fitting to say, like, Googlebot might not be able to see images. If we make it easier for Googlebot to understand what the image is about, I'll see it. Attribute, it's easier for us to index this image, and it's better for users who can't see the images.

Win-win.

If you have easy-to-read text that has a reading level of, I don't know, 6th grade or something like that, that's better than like super academic text that no one understands what's going on in it. It's like, think of your users, make it easier for humans to consume your content, and usually you make it easier for bots to consume your content.

That's that's awesome, and a nice way to finish also today's episode. And thank you very much for for all your answers. I really appreciate you giving like this your time and quite insightful answers too. Thank you, and looking forward for today's event too. Awesome.

Thank you very much.

Thank you.

Thank you.

Chronic Mondays.