Y Combinator Startup Podcast - #95 - Eli Schwartz

Episode Date: September 27, 2018

Eli Schwartz is the Director of SEO and Growth at SurveyMonkey. He came in to answer commonly asked questions about SEO.Eli blogs about SEO and growth at elischwartz.coYou can find him on Twitter at 5le.The YC podcast is hosted by Craig Cannon.***Topics00:23 - Does SEO matter in 2018?2:23 - Where should a company start with SEO?3:53 - Who’s doing SEO well?4:48 - Why is Amazon doing SEO well?5:48 - How should you crosslink?7:33 - Paying for links7:58 - SEO don’t dos8:38 - Things that are no longer useful in SEO9:43 - Keywords13:08 - Reviews13:48 - Content17:48 - Images23:08 - Link building28:23 - How much time to give SEO?28:53 - Mobile30:58 - Ranking32:08 - International SEO34:18 - Translation37:38 - International search39:53 - GDPR42:13 - Hiring someone to do SEO44:08 - What to do when you organic SEO disappears46:53 - Metrics for an SEO hire48:53 - Tools for SEO50:23 - Getting started in SEO

Hey, how's it going? This is Craig Cannon, and you're listening to Y Combinators podcast. Today's episode is with Eli Schwartz. Eli's the director of SEO and growth at SurveyMonkey. He came in to answer commonly asked questions about SEO. Eli blogs about SEO and growth at EliSchwartz.com. And you can find them on Twitter at 5L.E. All right, here we go. All right, Eli, let's jump right into it.

So first question, does SEO still matter in 2018, or can companies ignore it? focus on other channels. So this is a question I get asked a lot, and I've been asked this question many years past, and sometimes the answer could have been maybe. But now I definitively think that SEO absolutely matters and will matter even more. And the primary reason is companies were using other channels to drive acquisition. Startups could start their company and think of, say, PR or paid media, they've raised some money.

They use paid media to generate buzz and generate new clients. but those channels have not gotten significantly more expensive. Brand CPCs on AdWords have gone up. Facebook has made it a lot harder to get organic reach, and now you're paying for traffic on Facebook. PR in our world where we want to measure everything, it's not as measurable. And then what happens is companies do spend a bunch of money on those different channels. They're maximizing their paid budgets on Google.

They're maximizing their paid budgets on Facebook. They may even go into Bing and say, hey, maybe about it. I have some potential market opportunity on Bing. And then they're like, what's next? And what they do is they look in their analytics tools. They're like, all right, we've got breakdown on the channels. We've got all this direct traffic. Don't know anything about that.

Don't know what to do to get more. We have referral traffic. Those are outbound deals. Some of those people linked us. Some of those things we did intentionally. We have our pay channel. Great. We know what we're doing there.

Organic. So this is an untapped channel. We don't know what we did. We followed some best practices. We read SEO for dummies. we did what we thought was right. It's driving business, very, very little investment. And then they think maybe if we invest in it, now we'll get so much more and we'll

intentionally be working on this channel. So I think that's where we are in 2018 where like this is the untapped channel. This is the channel which has been driving business for free. And now if companies put effort in it, it'll actually generate significant returns. And where should they get started? Say they're at just like the bare minimum. You know, they like have a little bit of content. maybe the blog or something, but where should they really focus their energy in the beginning?

It really depends on the kind of company. So if it's an e-commerce company, they want to think about architecture. And, you know, companies, any e-commerce company thinking about SEO, they should have thought about SEO before they chose their e-commerce platform. So they want to make sure the site is structured properly to scale. So what do you do if, like, you get a new distributor and now you're ready to add 10,000 products? And your website can't handle 10,000 products. And now you're going to load it up and each URL is going to be one, two, three, four, five, six, seven. Next one's one, two, three, four, five, six, eight. So that's not going to scale for SEO.

You're not going to have any way of putting in your right product descriptions and optimizing against likely competitors are selling the same products. So that's e-commerce. They should focus on good architecture or should have focused on good architecture long before. Media, that's kind of different. That's where they can look at what has worked for them in the past. And my favorite search tool is Google search. console because it's from Google.

It may or may not be as accurate as can be, but it's probably more accurate than other tools which are sort of projecting what the actual truth is. Using that tool, they can see what's working. Here's what we're getting traffic on and marry that data with analytics. Like, here's what's actually performing and converting for us. That's where they should double down. Like build more content there, build more conversion elements on that and expand. Who's doing it really?

really well. I think media companies have actually gotten around to doing this really well. Like the way they crosslink and just have every sort of page possible. So you take like an item in the news today, they've got all these cross links to things they've written in the past, which suddenly can become relevant to the searcher. It's not newsworthy anymore. But if someone searches it like, hey, maybe they want to hear about a cow that got struck by lightning four years ago. And like there's a display at. So I think media has gotten around to like really maximizing the value of their content. When it comes to e-commerce, it's really only Amazon that's doing it amazing. Amazon has like, you know, built a structure where every single product and they have billions of skews somehow is discovered by the search engines.

And if you manage to put in the right keywords that match that right product, you will find it. And what are the, okay, so let's like be a little bit more specific. So what are the exact mechanisms they're using to be discoverable? So for the most important thing when it comes to SEO and making sure that your site is somewhat optimized for SEO is that the site is fully crawlable. So no orphan pages. So Amazon has built this structure where no matter what the product is, it may be something that there's only two customers in the world. That product is somehow cross-linked. If Google drops their bot anywhere within the site, it will crawl through the whole site.

It may take a long time, but it will find that page. that's where anybody that wants to do this right has to think about like what is this thing that I'm putting on whether it's an article or a product like what is it related to to to make sure that there is a path for that page to be found by a bot and that's where Amazon's doing an amazing job of making sure like there is a taxonomy for every single page every page is discovered and every page can be ranked with the right queries so in the context of a company that's not an e-commerce company would you recommend recommend, you know, using plugins to display relevant content to link to other pages or building your own. Like, how would you actually set up that architecture? So for media, the, you can use a plugin because all you're trying to do is create relevance. Yeah. Get links. Yeah. If users click it great. If they don't click it. Okay. But like things should be built for users. Okay. When it comes to e-commerce, there has to be a good plugin or something built internally because you're not going to want to have like say a red sweater cross-linked over to like a dog leash.

Yeah. Right. So like that's where a plug in that sort of, you know, correlates between hey, red and here's a red dog leash. That's not going to work. So you might want to build something internally. Again, media like what a lot of media companies do is they're just sort of looking for relationships. One thing I found to be very effective for cross-linking is just randomization. So if you randomize your URLs, you make sure that.

provided your randomization algorithm is correct, every single URL ends up being cross-linked. So if users click it, great. If they don't end up clicking, you still have achieved that level of cross-linking that Google can discover it. So just literally random pieces of content or products added in there?

Products not so much because there is a user experience to that. When it comes to articles, it will work. Really? Because, like, you know, again, someone's reading some article. They're reading about, you know, something that, you know, President Trump sent today. And there's an article about, you know, somebody getting struck by lightning.

They may click it and it could be relevant. And, you know, who's to say that we shouldn't, that's not another article that might be interesting. Yeah. Well, I think that will always be interesting. What about paying for links on other sites? That's an absolute no, no, according to Google. However, people still do it. Yeah.

And they're doing it through like a business development relationship. If it's done purely just to get the SEO link. then it's not advisable because, you know, someone could report that to Google and you don't want to go down that path of buying links. It seems like there are so many of these don't do, these no-noes. What are the other things people should absolutely watch out for? The big one is anything that you're not comfortable talking about maybe to your mother. So like, you know, hiding text at the bottom of the page so you're building a bunch of content. Like, you know, if it's not something you're going to be proud of, that's probably something you shouldn't

If it's something, the way Google really puts it, if you're doing explicitly to rank for SEO, then that's probably something you shouldn't do. They have a bunch of rules around that. But really, like, there's nothing you should be doing that's just for crawlers. It should all be for users. You know, maybe if there's some relevance for hidden text on a page for users, you might have an explanation. Right. Okay. And then are there things that were maybe quite useful in the past that are no longer relevant?

Like, I'm thinking like, you know, like meta keywords, stuff like that. Are there things that are kind of just like no longer useful? So meta keywords is a great example of something that's no longer useful. And the reason meta keywords existed is because Google and other search engines actually were relevant at the time didn't have a really good way of understanding text. So now we're like in the stage of, you know, development around search where Google announced today that Google Home understands Spanish. So like the understanding spoken words. and their understanding spoken words not even in English. So, you know, they don't need you to put meta keywords on a page to say what you're trying to rank for.

But back then, when, you know, search was early. They kind of did. Google was more advanced in other search engines as they were able to determine what the page is about. However, like, meta keywords was still a signal of like maybe what's on the page. Of course, marketers scam, they spam that by putting like what they thought would be the most relevant words, which are usually porn. Yeah. That's what people were searching for. And what about actually, so we, obviously, this is on YouTube, when YouTube asks for keywords, do you recommend people use those?

I mean, like we use them, but to a limited extent. So YouTube's different because Google's not great at really understanding what the video is about. Yeah. So by giving more information and it could be a transcription, it's those keywords that you put in, you're giving a better indication of what the video is about. With YouTube, though, it's really hard to guarantee your ranking. So the more information you can give, the more likely you have a chance at playing that game. One other place where keywords are really important and they're hidden is when you're trying to rank an app on the app store. So on Google, the way they rank an app is based on all the visible text.

So it's the name of the app, which is sort of like the title. The short description, which is forget how many characters, but a shorter page description of the app, which you see right above the fold. And then the long description, which is thousands of characters where you just fill. and whatever you want. So all of that goes into what Google's going to rank the app on Google Play. On iOS, there are keywords, and I think you have 100 characters hidden. So if you're trying to rank an app on like, you know, let's say selling something random, but you think that people are going to search porn keywords, that's where you would hide it.

No one would even know that's what you're trying to rank on. Oh, man. Okay. Gotcha. And so when it comes to actually, like, building out your SEO, like, what is the kind of content that's important? Because, you know, we talk about content marketing. We've talked about it in podcast before.

We've talked about it in the context of YC and other companies. And oftentimes people just kind of copy what other companies are doing. And, you know, it's not always the most effective strategy. So how do you frame, like, what's the most important content to be actually working on for your company, whatever that might be? So content's something that people have talked about forever and like they had this thing. Content is king and they used it for like, you. said content marketing. But really, it's something that we always need to underscore when it comes to SEO is because people stop using content. If you fill up a page with images or video or

JavaScript, you'd no longer have content. So Google, as much as they say they're really good at understanding and ranking JavaScript, they're not. Because in real time, when they're looking for your content, you are handicapping yourself by not having just raw text words on the page. So when it comes to like SEO content, raw text words on a page. And it's not really a truism, the more that more is better, but it is sort of. Because if you have a product description and you only have 100 words and somebody else has the same product, but they have a thousand words in different ways of describing the product, more than likely they will get more search traffic because as people search different keywords related to it, they become relevant because they've used the words. And you are not relevant because you have not used the words. So in that sense, more keywords are better.

More keyword or more words are better. More words are not better if all you're doing is filling up that words with emptiness. Like if you're describing a mug. And you describe many different ways like this is a mug that is a cup that you can drink or you can pour things in. And there's all these irrelevant keywords. That's sort of a waste of time. And it won't convert for you because users are going to read that. They're just going to be super confused.

So that's where people add in like reviews and that sort of thing for e-commerce. So reviews can be tough too because Google can understand that it's user-generated reviews, and it may not be as high quality. But that's something where, like, I don't know. Those are one of those things where you have to take that question and test it from the SEO standpoint. Like, if I had reviews onto this, am I going to rank more on different keywords and get more traffic? Or is Google going to look at that and say it's user-generated and not send more traffic for the content that's in the review? Got you. Okay.

So then what about, you know, your consumer company, it's not an e-commerce. company and you want to get into generating content that will generate traffic for you. Where do you start? And then like let's go through the actual workflow. So let's think of an example company. They're selling a service. So you want to come up with anything related to that service. And the, so the first kind of content you're going to need is the closest to that service.

Let's say the services is plumbing. You just go old fashioned. You're a plumber. So you offer different services that people need. like unclogging drains. So you want to have a piece of content that talks about unclogging drains. Now, when people search, and this is something where you don't really need to rank on unclog drains anymore because Google's very local.

So if you're in San Francisco and you're a plumber and someone searches unclogged drains, you now have a much higher chance of ranking on unclogged drains than Rotor. It's just national. By adding your address. You don't. Yes. Your address should be there. and you need to have some sort of local signal.

But Google's going to show for on-clog drains, that's a need. Yeah. Plumbing need right now. Google doesn't help to even show someone not in your area. Okay. So you want to have that piece of content. Now as you build out that content, you can move further up the buying funnel, which is, all right, now I have an immediate problem, unclogged drains.

But how do I make sure I never have a clogged drain? Now you can have best practices and how to maintain your plumbing. So as you build out this fuller site, it can even go into like, how do you, hire someone to renovate your house and make sure you have quality plumbing and what's the kind of piping you should choose and let's say you're going to buy it yourself. Where should you buy it? So there's many, many different avenues of content all related to original service, which can help you establish yourself as an expert in plumbing and pipes and clogged drains.

But how are you actually? So if the angle is like, it's just taste, like it's just opinion and you figure out like if you really understand your customer, you can know all these things. But is there a more programmatic way to be like, okay, customers that end up on our site are looking for X, Y, Z thing? How do you find out those terms, those queries? So you won't know until you start. So the best way is you're using Google search console. You can see the kinds of things people are looking for.

If you don't have that content, so say you had a great piece of content on unclogged drains, but everyone's searching about like unclogged bathroom sink. Like that's the specific thing they're looking for. And you don't have a piece of content on unclogged bathroom sink. Yeah. Time to write that piece of content. Now, before you start, what you can do is you can go to the Google keyword tool. I think that it's called Google Keyword Planner now.

They keep changing the name. So you go to the Google Keyword Planner and you're searching for keywords related to what you need. So you're looking for drains, you're looking for clogs, you're looking for unclog. And that should give you a bunch of ideas along with search volume about what people are looking for. You can also use my favorite SEPAID SEO tool, which is called AHRFs. They also have a keyword tool. So you're getting ideas. And from there, you want to sort of battle test those ideas by launching them and seeing what kind of traffic comes in and seeing how it performs.

Like maybe you have this great page on unclogging drains and all people do is learn and never fill out a lead or never call your phone number. And are you tracking that through Google Analytics or using other tools? What do you recommend? Analytics would be my favorite tool for tracking how their performance is going on, like as far as traffic coming in. Yeah. So even though you can't see the keywords people are searching on analytics. analytics, you can see the URLs. So you look at the URL that's generating traffic.

You see the conversions. Hopefully you've tagged your goals in analytics. And you see the conversions that come through. And you know what's your best performing content. Now you can take that same URL, put into the Google search console, and then see the kinds of keywords you're even getting impressions on. If their keywords are getting impressions on, but people aren't clicking, that's an area of optimization.

If your keywords that you're getting impressions on, people are clicking but not converting, another area of optimization. So it's very iterative. Like you're starting with getting ideas, maybe starting with keywords, building it, and then learning from what you're getting. So you said that it's important to have raw text on the page. That's the easiest to parse for, you know, Google and other things that are crawling your site. How do images affect all this?

And are they important at all? So the most important thing to consider within all SEO is that you're doing this for users. Yeah. You're doing this because users are going to use Google and then find your website on Google and then click through into your website and then hopefully experience your website and do whatever you want them to do. If you're generating revenue from ad impressions, you want them to click another page. If you're generating leads, you want them filled the lead form. If you are selling a product, you want them to buy it or whatever that is. So images help with that user experience.

So all you're doing is putting text on the page, none of us, like, like to look at raw walls of text. So the images in that sense should fit in with their user experience. But there is an SEO aspect to that. And that is people like Google Images is one of the largest, if you want to think about it as a separate search engine, it's one of the largest search engines after Google. I think YouTube is number two, but Google Images is definitely out there. So if you put images on the page, one thing you can do is make sure the images are titled.

so the actual JPEG file or PNG file are tagged with the keyword. So say you're selling coffee mugs, you want to call it like coffee mug.jepac. And not just some random file name that came from the camera. You also want to put alt tags or alt text around the actual image, which is still needed in today's day. But Google eventually will figure out what images are. And they're going closer to figure out images are. Alt text is the version of meta keywords like meta keywords used to be. So Google didn't know or doesn't know what images are.

So you're saying what the image is. Yeah. And that helps your rank. The actual ranking formula for Google images is really, really hard to discern. But if you follow best practices, like you have this chance of, you know, people searching. And then there's this additional avenue of people finding content. Yeah. Well, it seems like Pinterest has been the one who unlocked that, right?

Like many things I search for and I click through, I end up on Pinterest and I'm forced to log in. So that's essentially them doing a great job of SEO around the images and then they have these unique images and they show up on on Google. However, there is something interesting about your Pinterest question, which is Google specifically does not like search within search. Google would much prefer that you don't go to Pinterest, that you go directly to where that image was. But maybe that site didn't do a great job of optimizing the image, but Pinterest did. So now Pinterest is that layer in between. So if you just sort of follow best practices like Pinterest does, now you're skipping at the middleman and they're coming directly to it.

And are there best practices around, you know, like technical aspects, like literal resolution of the image, anything like that? Not as far as I know. Okay. It's really raw and sort of like put your all text in. You can put a title in around the image. You can caption it. Yeah.

There's also relevancy. So like the kind of page that the image is on sort of gives more signals. But for now, I don't know that Google is. really looking into the quality of the image and they're not yet, I don't think, they're out of place where they're understanding what the image is and then choosing to rank it. Got you. Okay. And so say, all right.

So now we've built out our content site. We're keep with like San Francisco plumbing and we've made, you know, 20 blog posts about whatever we think someone might be searching for. When it comes to distribution, are there any like non-intuitive pro tips that you would recommend? Are you suggesting people start getting paid traffic through their sites or or do they just leave them up and hope for organic? You can't hope for organic. You could wait like years for the organic traffic to show up. You need to be proactive about that. However, in the meantime, and actually like when you, your organic strategy should start when you put up the website. So in the meantime, you can be generating paid traffic and then organic has time to grow until organic becomes an important channel for you. The we're sharing things on Facebook or, you know, sharing things even with friends and family helps is that if you've created great content, let's say, again, on the San Francisco plumbing, you've created the best guide in the world about how to unclog your drains and you share that on Facebook. Now, other people may share that to, you know, to their friends and people see that.

And then someone out there is going to decide that, you know, they're writing a blog about do it yourself and they just have to link and recommend this blog post that you've written. article that you've written about on clogging trains. And that's how you're getting links. And you're starting the distribution of that for organic. And from there, Google discovers it. You can also go and, like, you know, as you build your website, so, you know, it's a service website, get yourself listed on all the directories. So, you know, Chamber of Commerce and other directories for your neighborhood.

So the directories don't really give links, but they give avenues for discovery. And then as you get discovery, you can generate links. So a pretty common question and word around SEO is links. link building, all that kind of stuff. Is it important? And if so, how would you go about doing it? So links are what differentiated Google when Google first came out. So at the time, there was Yahoo and a bunch of other crawlers and indexers and search engines, if you want to call them that.

And what Google did differently is they ranked a site based on the quality of the links. And it was sort of academic. Like if you were an academic and you've come out with something interesting, other academics are going to cite you. So Google determined that if you were, and they had this page rank patent and algorithm, that if you were quality, other people, other websites would link to you. So that was how Google sort of grew into that. And then, of course, people spammed it. They bought links and they, you know, they did a bunch of spammy things to get links and traded links and all that. Links are still very important, but they're not important in the sense that they used to be where people measure them by quantity.

Like my competitor has 100 links to this page. I need 101 links. And if I get 1001 links, then I will get. Rank ahead of them. So that doesn't work anymore. What you need are quality and relevant links. And it could only be a handful of links. That's all you need.

And what Google is doing is this is the algorithm borrowing a sort of link equity from the pages that link to you to make your page more valuable and relevant for all the other queries that people might find you for. So it doesn't have to be like this page, say it's Stanford is talking about metaphysics and you're talking about plumbing. Therefore, you can only rank on metaphysics because metaphysics is. linking to you, but like that link value is helping you rank better on plumbing because of Stanford thinks you're a good plumber. You might be a good plumber. Right. Now, where I like to do link building and where it's the most effective is to create

interesting content. And this is sort of where SEO goes hand in hand with PR. So you're creating interesting content that people might want to talk about. And then from there, people might also want to link to it. And one good metric of measuring if you're on par is do people share about it on Facebook? Like, do they like it? Because if they, the motivation to like something is the same as the motivation to link in something on if you have a website. So the most successful link building campaign I ever did was really an intentional link building campaign was I, when I was living in Singapore, I did a survey with a company or an agency that measured the cleanliness and satisfaction of people around bathrooms.

I recall the restroom association of Singapore. And we created this survey, and I created a bunch of interesting questions around people's bathroom habits, which is something people are always interested in. And from that survey, I turned it into infographic and a blog post and then just kind of put it out there. And I basically got links from every single media site within English media site within Southeast Asia. And it even got translating to other languages. And that was linked. But they were mentioning it and they were just providing due credit back to the original source with a link. So if you're doing creative things and you're doing good PR, that's sort of like a good SEO effort because you're getting that link to that content.

How many links you need? Still, we wouldn't know. Where do those links need to come from? Like that, you know, you can't know. But if you're working on these processes and making sure each page is linked, you're doing what's necessary to get that link value into each page. And are you spending time and energy and money on PR itself? or do you just mean like hopefully organic traffic? Yeah.

So I'm thinking of this is like PR-mindedness rather than actual PR. I, from my experience, the PR people I've worked with, they don't necessarily do things for SEO. So I've even worked at the PR agency where like they were proud of an article that showed up or an article that was in print and they gave me the PDF. That did nothing for SEO, obviously. I mean, they just use PDF. But if they're doing something where it's actually an actual link or a.

Or I mention with like, here's a credit back to the source, that would be a link and that would be valuable for ICO. Okay. And then as an individual, like, thought leadery type person thinking about link building, do you recommend people go out and like, you know, blog on other sites? Is that still an effective strategy? It is an effective strategy as long as it's not spammy. And I got an accidental sort of link, probably best link ever got to my blog, is I was working with a professor at Stanford. he invited me to his class, an interesting seminar. I was passionate about things related to Asia.

It was an interesting seminar. And he asked me to do him a favor and take notes. And I got this idea that if I took notes, maybe I could turn it into a blog post. And if I turned it into a blog post, maybe he'll publish it. And I asked him if that was something he'd be interested in doing. And he did. And now I have a blog post on Stanford's website linking back to my personal blog. So that is a guest post, but there's nothing spam.

me about that. Like, this was content the professor needed. And that's something that can be repeated anywhere. Professors are like anybody, they can be baited by their egos and you can write content about them and they're very likely would link from their classroom web pages. Well, also, if you just make something good. Like, this is something that is overlooked time and time and time again. It's like, make something good. Yeah. That is the underlying ingredient for anything related to SEO. Good content will get good links. Yeah, perfect. And so how much time do you give yourself to really test an idea and test to see that something's working? It depends on the scale of idea.

So when it comes to a piece of content and the website's brand new, could take a really, really long time to know if that content's resonating to get ideas around it. Does that mean a year? What does that mean? It could be a year. Okay. But if like, you know, it's a going concern. Say, you know, it's Y Combinator.

You've already got a ton of traffic. You know that you put something up. It's instantly going to get found. people are going to search for it. That could happen in a matter of weeks. Okay. And what about all of this in the context of mobile? I know the habits are very different.

So there's a lot of talk about mobile first and Google moving to mobile first. And within that, what people forget is that there's still only 24 hours in a day. So yes, we're on our mobile devices, all of our waking hours. But eight of those waking hours or more than eight of those waking hours, we're actually out of desktop or a laptop. So we're on a larger device. So it's not that we shouldn't optimize for the desktop anymore. It's just that you need to consider mobile and you need to consider user behavior. There are many, many products that are probably never going to be bought on mobile. Like probably no one's going to buy enterprise software on a mobile device. And I highly doubt,

at least in our first world economy, people are going to book their entire vacation on a mobile device. You'll probably, like, you know, do a little searching and you're, you know, while you're sitting on the train, where should I go? What should I see? but before you shell out thousands of dollars, there's a desktop involved. Or if you're doing, let's say, Black Friday shopping, you may read some emails around Black Friday deals. You may do some searching on your phone.

But before you go and buy like a brand new TV, you're going to go to a desktop site, look at that and look at it as many images as you can, read reviews, possibly even purchase it on the desktop. So, yes, mobile is growing. We need to take it. We need to consider mobile users, but not at the expense of desktop users. So with that in mind, what are the best practices to account for both?

When it comes to SEO, not much. It's really taken to account that there's a smaller screen and people are clicking through on search from a smaller screen. And then the biggest thing that you really need to keep in mind is that it's not about ranking in the top 10 results. You actually have to rank in the top three results. And they just try harder because the mobile devices, you know, there's just less results. Yeah, I mean, that does seem to be the perma challenge, right?

Like getting from the second page to the first page and then to the top of the page. Do you have any, I mean, obviously, like, you've written about this pretty extensively online, but like, what are your pro tips? So it's really matching user intent. So as we discuss, like, the user is the most important. So understanding the user and getting in the user's head and finding the queries that the user will search for to find your content. So it's not about like, what do I need to do to be in the user's head? the first page because that's like a very vague goal. You want to be in the first page for what your specialty is.

So our plumber's specialty is unclogging drains that are very old. Yeah. So they, that only matters. And they can chart, let's say that plumber can charge a lot of money for their specialty. So that's what they want to be on the first page for. Write a lot of content for that specialty. Make sure they're found mobile devices in the first couple of results on the desktop and for sure the first couple of results. But just generally ranking for like this broad term.

that's very hard to do. And if you're not a big brand, may even be impossible. Right. And if you can't spend against it, like a lot of cash. Yes. And in the international context, how do you think about getting to that? Assuming that you have a clear intent, say you're going to franchise, your San Francisco plumbing business is going to open up in Beijing.

What should you be thinking about? So international is the untapped opportunity in SEO. But it will probably remain untapped for a long time because the product. and services most American-centric or English-centric websites offer aren't necessarily relevant for a global audience. You can take a couple steps into becoming relevant. So say your e-commerce, if you partner with a company that can enable shipping to other countries.

So you don't actually need to do the shipping yourself. You just need to get yourself listed or make whatever integration you need necessary in order to get things out to other countries. If customers want to buy, at least you've enabled that. You can also take other sort of payment, like work with a payment provider that may be able to take other currencies. It may come into you as dollars, but you just want to work with that gateway that can take the pound or take the euro. Because if you're just working with the standard payment gateway, like if someone doesn't have a U.S. credit card, they can't buy them. So that's an opportunity to just expand a little bit.

Maybe if you have an audience in another country, that's where you can take one step forward into SEO where say you have that audience in Beijing. you might want to create one page about your product in Chinese, and then people will discover it, click through, and that's an experiment I've tried in the past where, like, you localize a page, you have your CTAs in that local language, click the CTA, and then bang, you have English. So you're obviously going to filter out a lot of people, but you're also going to get people through that

where, you know, if they may be a potential, you know, customer of your product, they would still go through that funnel. So you've broadened your appeal. you've broadened your potential base by having your content translated, and you can potentially get more customers at the bottom of the funnel. But in general, international, huge opportunity, but probably not relevant for most companies.

Okay. And so let's use YC as an example, right? All of our content is in English. It's at Ycombinator.com, right? Do you build another website with a separate top-level domain, or do you add, you know, Chinese translated content to your main blog? What do you do? So it used to be best practices where you would build another website for customers in another country.

I no longer think that's the best practice. And I've tested this a few times. And the reason behind that best practice is that users trust something local more. So a UK user is going to trust a dot code at UK. Google actually hides URLs in search to begin with. And you can't, you know, they test it. Sometimes it comes in, sometimes it comes out. On mobile, you don't see the URL at all.

So that doesn't matter. Users don't even know that your dot com versus dot code at UK. So with that in mind, you have a great website already. Why would you split your efforts and create a new website? So what you're going to do is you just want to open up a new directory that's very explicitly another language. And you want to follow the two-letter naming convention or the three-letter naming convention. So Google understands exactly what it is. So if you're going to Chinese, I think it would be ZH.

for Chinese or if you're going to another country. So obviously the UK would be UK or GB. So Google understands that localization aspect of like, hey, this is hard to get it somewhere else. You wouldn't want to go to like, you know, say Spanish and just, you know, do SPAN. Yeah. Because it would be hard for Google understanding. And with that, you're doing a direct, a subdirectoring, not a subdomain. Okay.

And then you are just having your content machine translated. Are you paying for translator? what are you doing? Again, user-centric. If content is machine-translated, the machines have to be very, very good. If it's very good and it's not that important content, then that should probably work.

If you really want to get users through the funnel, you want to make sure there's no misunderstandings, you should hire local people to translate it for you. Okay. So, you know, in the context, we can just use YCN as an example. It's pretty easy. you would just take say like our top 100 blog posts pay for a translator who's like native in Chinese or Spanish or whatever it might be post it on the act of Ycombinator.com site in a new

subdirector in dedicated just for Chinese right okay and then meaning like a new subdomain on the blog not subdemean so sub directory within the blog right okay and link that through the nav on the blog? What is the best practice there? So you could do it through the nav on the blog, but it may confuse your existing English users. What most people would do is have some sort of language selector at the top or the bottom of the page. Again, with the user in mind. And the language selector should be an HTML link, which switches over to that new directory. What you don't want that to be is a JavaScript link where it sort of reloads the page and brings up the new content because then Google possibly couldn't follow it.

Hmm. Okay. And so all the content is actually there. All the content, well, so all the content you've translated is there. Right. So you don't want to have any sort of automated like I change, I'm changing, let's say, E.N to automatically ZH or something like that. Okay.

Gotcha. And then in terms of appearing on separate search engines across multiple languages, what are the best practices for that? So there really are no other important search engines in the world. there's, I mean, they're locally specific. So there's Baidu for China. So if you care about China, then it's Baidu. If you care about Korea, it's neighbor.

And other than that, I mean, it's Bing. And wherever Bing has penetration. So if you're following best practices for Google, that should help you rank on Bing. That doesn't necessarily help you rank on Baidu. And that's for China specific reasons. In order to rank on Baidu, you actually need Chinese hosting. And the reason for that is obviously China has censorship. And Biden's concern is that if you're not, if you're not local, they don't want to take responsibility for your content.

If you're local, you're already complying with laws on what kind of content you can have because you're local. So if you're not local, you're going to see a huge rankings hit. Your website will be significantly slower, obviously, you know, getting behind the Chinese firewall. Yeah. So if you care about China, there's a lot of things you need to do. it may be even advisable for China. Again, this is China-specific, not Chinese language, to have like a separate Chinese domain

because to have that Chinese domain, again, you're complying with the loss. And then Baidu would rank you. Korea is different. Korea has a neighbor as a search engine. I don't remember the breakdown of neighbor versus Google right now. But Google used to not be dominant in Korea, and they've been increasing dominance

through their Trojan horse of Android. So Google's obviously very closely partnered with Samsung and LG, Korean companies. And as Korean users use these products, they use Google. Prior to that, Navor, which is, it was search, but it was also sort of like an AOL where they offered everything. From forums to chat to blogs and search, again, their ranking algorithms were very different and sort of rudimentary when it comes to Google. So less Korea is a specific target for you, where China's a specific target. target, Google's all you're going to need. Right.

Okay. So if you kind of just like optimize your Google strategy, dial that in, you're going to be in a really good spot. Absolutely. And so thinking again internationally, like GDPR just happened, what should people have in mind? And like, why do you think SEO is important in the context of GDPR? So I'm in the middle right in a blog post about GDPR and the benefits for SEO.

I think GDPR is one the greatest things that has ever happened to SEO in the last five years. For American companies? For everyone. For everyone. Okay. So one of the requirements of GDPR is updating privacy policies. It's why everyone got an email from every company they have ever engaged with about a privacy policy update. Privacy policies also have to be very clearly defined and linked on websites.

So now what you have is one page privacy policy that appears on every single website or every single page of a website. To me, that's the most amazing way of cross-linking through an entire website. Whereas in the past, you may have had a tough time of like, How do we get users from a home page over to this deeper page? Now you have this deeper page, which is linked to privacy policy. And if you, you know, you can involve a site map in it to make sure that privacy policies there. You have one commonality between everything, every single page on a site, privacy policy. So that's one advantage.

The other big advantage is that paid budgets are now going to take a huge hit. Because of GDPR, we've gone to this nightmare scenario where you don't know if your paid budgets are working because you can't track users. until users agree to be tracked. So if you reduce your paid budget, where are you going to put it? The only place you're going to go and put that is an organic, and you're going to start ramping up organic efforts. And then the other thing that I think SEO really benefited from this is paid has always sort of cannibalized SEO because it's everywhere.

You may have discovered a website via organic, and then you got retargeted and maybe you end up converting with paid. With the reduction of paid, SEO might actually get its due. Like we're going to see that full funnel conversion SEO all the way through SEO. And is this affected how you do like any SEO practices? Not necessarily. And the other benefit to GDPR and SEO is it's one of the few channels that works without GDPR, like without any opt-ins.

Like Facebook, if, you know, I don't know that the stats are out yet, if people refusing to use Facebook because they don't want opt-in through their privacy policy, maybe they're not on the platform anymore. But you don't need to agree to Google's privacy policy. and cookie rules in order to use Google. So you go on Google, you won't agree to anything, any tracking. You're still now going to search and find websites. Okay, great.

So if someone wants to employ all of these tactics, who should they think about hiring? Who should be brought on to the team? So the people that I found to be the best at SEO are incredibly creative and are very curious and are able to develop a hypothesis around what should work. And they're basing that on other things. they've seen, things they've read, things that Google has talked about. Google has a whole guide on SEO best practices. They put out a guide on best practices, as well as a guide to their search quality raters of how to evaluate a website. So you can see what Google's thinking. So you want someone

that can take all that, come up with ideas, but also is incredibly analytical so they can understand what works and learn from past performance of websites and measure their own testing. So how as say you're a startup founder, right? Like maybe you have 10 employees or something like that. How do you even evaluate the applications you might get? It's sort of hard. You want to look at someone like I think someone that was entrepreneurial in the past demonstrates that creativity.

But this is something where I think you're the on-site in-person interview or phone interviews or ask a lot of questions. It's really where you're going to flesh out their creativity and that analytical ability. Looking at a resume, I've looked at many resumes of people that have done SEO and I've discovered they don't really know anything about SEO. They've just done SEO. When you talk to someone that would be good in this seat and it doesn't even have to be as someone that's ever done it before, you're asking them creative questions. Like, how would you solve this? Or like, what do you do if you woke up in the morning and your organic traffic has disappeared? Like, where do you go first? How do

you troubleshoot this issue and just understand their ideas? They'll make mistakes. Everyone's going to make mistakes. But you want someone that, you know, has that natural curiosity and creativity, along with the smarts to really be successful in this. Actually, what is your answer to that? That's a great question. Your organic disappears overnight. What do you do?

That's happened to me multiple times. Okay. So the first time it happened to me, I was working for a startup, and I remember the exact date. It was February 24th. I don't remember the year. And it was the day Google rolled out the Google Panda update. So up until this date, which I think it was in 2010, the best way of writing content was to write as much of it as possible and come up with every keyword and rank on everything that you could think of. And these are the days of EHOW where you actually had guides on how to pour water or how to boil water.

And there's the crazy stuff on EHOW and may still exist. So we had the site I was working with and the company I was that, we had content for every possible iteration. we was an automotive review site, and we had every car, we had every review, we had every picture, and we lost 65% of our traffic one morning. So at the time I was measuring rankings, I no longer believe in ever measuring keyword rankings, like as we discussed, we're focusing on a user. So it's not just generic rankings. Started with looking at our rankings, what we'd ranked on number three the day before.

We were now at number 80. So it started with that, then going into analytics, looking at the URLs that got hit. And then after that, oh, crap moment, it was, it looks like this is a Google penalty. And it wasn't until, I'd probably say midday until the blog posts had been out of like, hey, you know, we've, Google first put out the blog post. Like we've, you know, we've gone out and destroyed half the internet and changed the way things work. Man, okay. That's, and then what did you do to fix it? So we actually had to rip the bandaid off.

So we lost 65% of our traffic. We ripped off. We took down tons of content, brought it down to like 80 to 90% complete loss of traffic. So we took it intentionally took away organic traffic and rebuilt up our high quality content, rearchitected everything we were going to do. We ultimately recovered. So we recovered with a 20% gain on what we had in the past. And how long did that take? That took only six months.

We were lucky that like we ripped the band-aid off and were able to do it. Nowadays, it could happen a lot faster. At the time we were waiting, we were waiting for Google to rerun that algorithm. once they re-ran it, we were back in the good graces. Now, like, things are more automated and algorithmic where if you got hit by the penalty one day, you cleaned it up, did everything possible, maybe filed a reconsideration request, you can be back in a couple weeks. Oh, okay.

It's not so bad. So going back to this person that you're evaluating to be hired for your company, say you pick someone, you think they're good, you hope they're good. How do you set metrics for success? How do you set a goal and then how do you measure it and how do you know if it's, working like what should you be looking for and and really like how much time should you give this person so if they're creative they should be able they should be coming up with creative plans right away so coming up with a strategy of how you're going to grow and they're also going to require a lot

of investment yeah if they're coming up with strategy around content you got to buy the content so that's the first thing you're going to want to look for is like what is our strategy how how are we going to grow how are we going to increase your SEO traffic or get SEO traffic if we don't have any to begin with. That's the first metric. I'd say within 90 days, you want to have that plan. Moving forward, you're going to measure against that plan. Like, are we getting traffic?

Did we make any mistakes in our forecast and our estimates? And then, again, assessing your creativity, how quickly can they iterate and pivot if they're wrong? So you're looking at the growth in traffic, like that's your primary metric once the traffic starts coming in, put some goals in that. If it's a startup, you can have things like 100 to 200% growth. If this is an existing company with already has tons of organic traffic, maybe it's more manageable to have 10 to 20%. But those are your metrics you're going to apply to this person. Okay. Got you.

And then I would look at traffic and dollars coming from this channel. And what you never, ever want to measure for anything related to SEO is rankings. Like how many page one or how many position one rankings we have? Because those aren't yours to have. those are you know Google could change their mind or users could change their minds and like that is the wrong way of of aligning your SEO efforts.

Yeah, I mean that's kind of true across the board. Like all of these systems, if you engineer the employee incentives for them to just game it, that's what's going to happen. And if it's not directly tied to your bottom line, you're wasting everyone's time. Absolutely.

What about these tools? So you mentioned AHRFs. You mentioned Google Analytics. What are the other tools? people ought to be considering to like you know say you brought this employee on right like what do you need to install the first day to ensure that you're tracking all the right things so tracking can really be done with age refs can help you track keywords again i'm not a fan of tracking keywords a hrefs can also help you track links so if you're building links it's just a good way of seeing

what google's recognized not google but like a h refs and hopefully google's recognizing it too google search console that's free the other piece is the technical piece. So as you're building out a site or as you're, if you're working with an existing site, you need to use a crawler to try to imitate how Google would actually crawl the site. So my favorite tool there is called Screaming Frog, SEO Spider. There's a similar one that only works on Windows called Xenu Link Sluth. So you're using these tools to really crawl a site and you can get back all of your

SEO metadata. You can see all your redirects. You can see all your canonical links, which is pointing to duplicate content. pick up all your image URLs, find all your JavaScript. So really just like recognizing whole site. Not as relevant for a small site, but just an interesting tool to have in your tool set. The other thing you can do with that tool is crawl your competitors. Now, I wouldn't recommend crawling your competitors from your company IP address

because if the competitor savvy, they may see that show up in their log files like, hey, why is this IP address tied to my competitor now crawling me? and they couldn't make them upset. So, you know, use at your own risk. Okay. But this is a great technical tool to just understand how sites are structured. Hmm, cool. And what about if someone just wants to get into SEO?

Like, where do they, where do they start? What courses should they take? What would you do? So, no courses to take, really. It's a creative pursuit. Yeah. You have to, you know, start drawing art in order to be an artist. Yeah.

But the way I learned this was I was working for a company and I was managing affiliates. And at the time, and it might still be true, the most expensive AdSense keyword and AdWords keywords was miso thiloma, the lung cancer related to asbestos. So I decided to see like, well, I needed to write content in order to get content matched to this AdSense keyword. So I decided to write content related to misoiloma and see if I get ranked. And that was really my first SEO project. and in the process I wrote something about John Edwards and I don't remember the connection to it. And I, this is when you could still see all the keywords coming in. I saw keywords coming in related to John Edwards.

I'm like, well, if I write more content on this, maybe I'll get more traffic. And I did. So like, that's where you start. You start with like an idea. Maybe it's a personal blog. Maybe it's something you're passionate about. Create content. Get the traffic in and learn from it.

And they, you know, if that makes you excited to try more, then this is something for you. If this is like, well, this is taking too long and this is really boring and I'd much rather be like a Instagram influencer. Then that's where you should go. Awesome, man. Well, thank you so much for coming in. If someone wants to reach out to you and learn more, how can they connect? I like LinkedIn a lot. I respond to all my LinkedIn messages and I've done a decent job of making sure I can be found on LinkedIn.

Man, that's great. Okay. That's also rare. And yeah, it sounds like effective strategy. Yeah. Thank you. All right. Thanks, man.

All right. Thanks for listening. So as always, you can find the transcript and the video at blog. dot ycombinator.com. And if you have a second, it would be awesome to give us a rating and review wherever you find your podcast. See you next time.