May 10, 2023

Building an SEO Strategy w/ Ethan Smith of Graphite

Ethan Smith, CEO of Graphite a SEO, content strategy, and growth design agency has helped MasterClass, Thumbtack, and Upwork jump start into hyper growth. Learn how to build the right SEO strategy from one of the best in the industry.

Watch/Listen to the episode:

This episode is also available on these platforms:

The host

Nima Gardideh

President of Pearmill, ex-Head of Product at Taplytics, ex-Head of Mobile at Frank & Oak. YC fellow.

Our guest(s)

Ethan Smith

Founder, Graphite

About this episode

On this episode, Ethan helps demystify SEO and how to build the right SEO and content strategy depending on company stage.

Also, he breaks down keywords versus topics, how to build topic authority (even if you have none), how companies should design their content org, plus our hot take — what he thinks SEO agencies should optimize for but often don’t.

More highlights include:

  • What is SEO in basic, business, fortune 500 terms?
  • What are the three types of SEO? (Hint: Technical, programmatic, and editorial for growth and brand content)
  • Building a content strategy around topic themes
  • Measurement and timelines for SEO
  • What metrics should be tracked depending on company stage?

Follow us and our guests:

Want to share your hyper-growth story with us? Email nima@pearmill.com to be a guest.

The Hypergrowth Experience

Get notified every time when we release a new episode.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.


[00:00:00] Ethan Smith: There's a small set of few things that are really impactful. Those things are making sure that your cross-linking is set up properly, so you're linking across other pages, making sure that Google can even read what's on your page. If your page is empty, it's not gonna work that well.

[00:00:14] Ethan Smith: That's on the technical side. On the editorial side, write articles that are adjacent to where my topical authority is. So write stuff about LinkedIn pixel tracking. Don't write stuff about Instagram cause we don't have any authority there. It's like prioritization of the pieces of content that you have.

[00:00:29] Ethan Smith: Making sure that the outline is comprehensive, making sure that our conversion optimization is set up and people aren't visiting and leaving on the programmatic side, category pages and product pages. So making sure that those are set up properly. Google can read them. We've launched the ones that are impactful.

[00:00:45] Ethan Smith: If we're Thumbtack, we've launched all the different plumbers plus city combinations. Those are the and title tags. That's what's impactful. 


Transcript of the episode

[00:01:18] Nima Gardideh: Welcome to another episode of The Hypergrowth Experience. I am your host, Nima Gardideh. Today we're gonna have Ethan Smith, the CEO of Graphite on the podcast. He's an SEO expert. He's helped a bunch of companies like Masterclass and Netflix and Thumbtack scale their s e o efforts. And now he's built, a sort of like a tech enabled services company called Graph.

[00:01:18] Nima Gardideh: That is helping companies scale their SEO efforts. He's quite knowledgeable in this space. It was very, very fun for me to speak to someone that has spent their whole career, understanding such a craft like this in such depth. And I try to be as sort of like inquisitive as possible. , I don't know much about the, the specific, channel.

[00:01:41] Nima Gardideh: I've had experiences doing. , in my early twenties, which is over a decade ago now, and a lot has changed since in this space. So it was very interesting to just kind of get an understanding of what it takes to now build these programs. and Ethan's done really great work recently and, has been sort of going around on the, on the podcast world.

[00:02:00] Nima Gardideh: So you might have heard him speak on a few different podcasts. , we try to be a little bit different in the content. We, we ended up creating together. and we're doing some events in New York if you happen to see or hear this before that, enjoy the episode and we start by talking about the overall approach that he uses whenever he's, entering a business.

[00:02:24] Nima Gardideh: what don't we start with the simplest version of this, which is what is, what is an, what is seo? , if you were explaining it to a kid who does, doesn't understand marketing yet maybe they are already do in, in all these different ways with TikTok, but and, and all these different, different new apps that they're using.

[00:02:42] Nima Gardideh: But what is SEO in a, in a less layman's term, More

[00:02:47] Ethan Smith: Yeah, I would say, yeah. The way that I describe it is when you search for something, we are, we help companies show up in that search. So usually it's Google search, but to your point, TikTok search, other Utah U YouTube search, we help you show up when people search for things.

[00:03:04] Nima Gardideh: And if you were to explain it to a, if you had to explain what, what SEO is to a business owner that maybe has a set of restaurants or a small e-commerce shop, how would you explain what is the value of seo?

[00:03:18] Ethan Smith: We help you get customers without paying for ads.

[00:03:21] Nima Gardideh: That's a good one. , and then if you were to explain SEO to a Fortune 500 ceo would that change the definition?

[00:03:31] Ethan Smith: I would change it slightly. And I would say that a few of your competitors are growing dramatically for free, quote unquote, without paying for ads. And you're not, and you can,

[00:03:42] Nima Gardideh: I like that a lot. That's when you get, get into these hyper-competitive situations. You can, you can think of it in that, in that context. , one of the things 

[00:03:50] Nima Gardideh: that I think, sorry, go on.

[00:03:52] Ethan Smith: oh, I would say for a smaller company, for a large company there's a, there's a fixed addressable market and you have to compete and take from each other. And there's five players and Walmart needs to take, take from Target, and at the Target's doing something and taking from Walmart, that's really important.

[00:04:13] Ethan Smith: Whereas if you're a smaller business, the TAM is huge, and so you're not really competing directly with our restaurants, not competing directly with another restaurant per se, but when you're large in your Walmart versus Target, Little things that the other competitor's doing matters a lot. So that's why if a few companies are winning an organic, that matters a lot and you wanna, you wanna beat them.

[00:04:33] Nima Gardideh: And that's just cuz of the nature of search, right? There's only so many people actually searching for things at any given point, and your job is to just capture as many of those queries as possible.

[00:04:42] Ethan Smith: That, and that's just also the nature of large companies versus small companies, large companies there, there's a Target and Walmart are directly competing for the same people, whereas a small shop, you're not directly competing with other small shops.

[00:04:55] Nima Gardideh: one of the things I think I've always had issues with, and I think you've been pretty well spoken about this and I wanna dig into it more with you, is just how to think about SEO when you are trying to plan for growth, right? And there's a couple of maybe a aspects to that. One is just budgetary, right?

[00:05:16] Nima Gardideh: , do I how do I calculate a dollar I put in towards SEO versus a dollar I put on towards some other program? Like and then the other part is just sort of timelines, right? So how do I think about the return of this dollar that I'm investing in? , and I think you've been pretty well spoken about it, but can you walk me through how you would break those two things down?

[00:05:38] Ethan Smith: Sure. So in terms of the measurement there's a different curve for paid and for organic. So for paid, as soon as you buy an ad, you get the return on that ad, and it's easy to measure. It's easy to attribute. Whereas for organic, you can launch a page and it might take months for it to start ranking, and then it actually increases over time.

[00:05:57] Ethan Smith: Usually it's like three to six months curve, and it can even continue ranking or getting more and more traffic over time. So the Nerd Wallet article that was written 10 years ago, or 20 year or 15 years ago about best credit cards gets more traffic today than it did. Last year and the year before that, and they've modified it, but still there is a fixed investment that increases for years and years, 

[00:06:19] Ethan Smith: which is very different from paid.

[00:06:20] Ethan Smith: The, the, the ad that you bought 10 years ago is not returning any value. So it's a very different long-term curve. It's kind of like if you buy a house and it's appreciating over time, you, you, you don't care if you make a return on that house that you bought the next week. You care about year five or year 10.

[00:06:36] Ethan Smith: So that's sort of how to, how to think about seo. , in terms of how to measure it, usually what we'll do is, we'll look at the first, you, you don't wanna wait 10 years to see whether or not you're getting a return. So you want to look for leading indicators early on. And the leading indicators are you launched a piece of content and maybe you're ranking number 15 for some competitive terms, and you're getting no.

[00:06:58] Ethan Smith: , but you're ranking number 15 and then you're rank 12, and then you rank number eight. Maybe you're getting two clicks, still not that much, and then you're ranking number five and you get a few clicks, and then you're ranking for 20 keywords instead of one keyword, and you're ranking number five for that.

[00:07:11] Ethan Smith: And so it's sort of grows over time, but you can look at those leading indicators of ranking for a few keywords then than more than getting some traffic, then getting a lot of traffic, and you can kind of model that out on a per article or per per URL basis so that you can see within the first few months what do we think is likely to happen over time.

[00:07:32] Ethan Smith: So that's roughly, and then you can look at conversion rate. So I, of a hundred visits, the conversion rate is 5%, so therefore five conversions, and that L T V is $10, and so it's $50 for that piece of content. So you can kind of model it out over time. So that's, that's how to think about s e o measurement.

[00:07:53] Nima Gardideh: And sorry, before we go to the, the cost part, Did I understand it right, that you're looking at other per content pieces, so you're trying to get a sense of how much more this piece of content can rank for, for and what, for what set of keywords? , or is that at the sort of domain level or like the brand level?

[00:08:11] Ethan Smith: We are looking at, at a per r URL level especially if there's a, a cost per url. So if you're creating a thousand category pages, there's no cost per page, but if you're writing an article or creating a tool, there's a cost per, per page that you're creating. So just like you would want to know how much your ad costs to purchase, you wanna know how much your piece of content or your landing page costs.

[00:08:34] Ethan Smith: So the page costs $500 or a thousand dollars, and you wanna look at the ROI curve. And if the ROI curve looks good, which it usually does, then you wanna spend even more money. You wanna launch even more pieces of content. So you wanna look at the per u per U URL level. And then you also wanna slice it by dimensions.

[00:08:50] Ethan Smith: So article versus guide versus tool different categories, so like home goods, electronics and then you can slice that and then you can be really efficient with how you're, how you're creating more pages and make more pages around the stuff that has the highest roi.

[00:09:07] Nima Gardideh: Gotcha. That makes sense. , sorry, you can continue to the cost part, 

[00:09:10] Ethan Smith: Sure. And in terms of the cost part, do you mean how? How to think about your spend.

[00:09:15] Nima Gardideh: Yeah. There's a couple parts of it, right? Like how do, how do you allocate on. Maybe it sounds like different pieces of content you do have to produce. And then much more importantly as if I'm allocating a budget across a CEO versus paid what is an easy, is there like a formula I can use to compare the dollars?

[00:09:36] Nima Gardideh: , because the timeframe, as you mentioned, is very different. , so if I'm doing N P V even, right, and, and most companies are not smart enough to do net present value, but if I'm doing NPV type calculations, what is a timeframe to look at in comparing the programs such that I can give SEO what it's worth?

[00:09:55] Nima Gardideh: And I think the, the point you've made quite a few times in different podcast even on recently, is that people are under allocating capital to seo. And so I just wanna get a sense of like, if I were to try to use a methodology in order to solve that problem, what is the methodology I could use to compare to dollars?

[00:10:14] Ethan Smith: Yeah, so a sim there's a simple answer and a complicated answer. So the simple answer is we usually look at a, and it depends on the company. So if you need to raise money in six months, then you probably shouldn't spend most of your money in seo. And if you need to make a return over the next 10 years, it's a great channel.

[00:10:31] Ethan Smith: So part of it depends on what makes sense for the company, but the, the, the simple answer is we'll usually benchmark out for three years. So it takes a few months to come up with a strategy, build the pages, launch them. And then traffic starts curving up at month six and maybe you'll start making a return at around month nine.

[00:10:50] Ethan Smith: And then that value curve is really compelling in year two. , that, that's kind of the simple way to think about it and we'll, we'll benchmark based on, if I'm Wayfair, I look at Target's, traffic, and a few other Kohl's. And then what is there traffic, what's the gap? And so that's sort of, we'll, we'll create milestones and then we'll timeline it out.

[00:11:11] Ethan Smith: , that's the simple way. We'll me measure it and then also look at conversion rate and L t V. So I already have some traffic. What's the conversion rate on these different page types? Apply that to the traffic that I modeled out, L t V, and then I can model that out, model it out over two to three years.

[00:11:25] Ethan Smith: That's the, that's the simple answer, 

[00:11:28] Ethan Smith: but there's a more complicated answer. So each of these can be broken down and, and, and be more complicated. So the first thing is attribution, multi-touch attribution. So for something like a Target last touch is totally fine, but for something like peer mill or for or for script, you don't just search for ad agency and click a button and then hire peer, peer mill, or, or for, for notion.

[00:11:54] Ethan Smith: You don't just type , project, project management software and then spend thousands of dollars. You've heard of notion over many, many touchpoints.

[00:12:02] Nima Gardideh: Mm-hmm.

[00:12:03] Ethan Smith: Or we work with BetterUp, which is a coaching platform. , most people who use BetterUp, many of them had never heard of BetterUp before and didn't even know that coaching virtually was a thing, but they learned about it over many, many touchpoints, ads, email, organic tv, word of mouth, et cetera.

[00:12:20] Ethan Smith: And so you need to be able to wait the each touch. This is obviously challenging to do, but , ideally you can do that, and that's a whole other conversation about multi-touch attribution. But that, that's one thing. , the second thing with attribution is to your point N P V or discounted cash flow model.

[00:12:42] Ethan Smith: So I launched this piece of content and for an ad, I get the return immediately. And for organic, it appreciates over time. So how do I value the, the value of a house that I bought that is not making a return the week after I buy it? But it's gonna make an amazing return over the next 10 years. So you can use an N N P D or, or discounted cash flow model.

[00:13:04] Ethan Smith: I've never, I've never seen anyone actually do that. But you can do that. And so if, and I think you, that you should, and I mean the, the the measurement error of not doing that is dramatic for a nerd wallet or TripAdvisor. If they said, well, we wouldn't get any traffic next week, so we should just buy a bunch of ads.

[00:13:24] Ethan Smith: The fact that they don't have to buy a bu the fact that TripAdvisor gets 80% of their traffic from, from seo, it is why TripAdvisor exists. So how they got all their app installs, so versus another company who's still buying ads. So the, the measurement era of not looking over a longer period of time is dramatic.

[00:13:42] Ethan Smith: So using something like an MP or a discounted cash flow model will give you a dramatically better measurement, which means that you can Ha have a huge competitive advantage in terms of efficiency of your spend on organic versus paid versus other channels.

[00:13:58] Nima Gardideh: But what is, what is, what is the timeframe in which you think N P V would work is like 12 months enough where it would compete well against like a paid channel. And I think you say one week, right? And I, I can argue that it takes something like 90 days for paid programs to also be useful , with, you have attribution that are properly and, and you're thinking about it in that multi-type sense as well.

[00:14:20] Nima Gardideh: , so that, that's kind of the timeframe that we look at things like 90 days. What is the sort of actual return on the, on the cash that we're putting into these programs? , I guess I'm just throwing 12 months out there as a random number, but does that make sense? You, you mentioned three years before, which is actually not bad.

[00:14:39] Nima Gardideh: Three years is, is harder to model out, obviously, but there, there's something there. , what, what makes sense to you?

[00:14:46] Ethan Smith: 12 months is too short. I think that two to three years is what I would look at that. And that doesn't mean that you have no information for two to three years, or that you will not get a return in two to three years, but it does mean that your return, the return in year two will probably be twice year one, and the return in year three will probably be two to three times year one.

[00:15:09] Ethan Smith: So I think it's important to look far out. But again, look at leading indicators so that you're not waiting so long before you have any information about whether or not it's.

[00:15:17] Nima Gardideh: Yeah. Well, those numbers were interesting. It felt non-linear. Do you think there was like a. What does the curve look like usually? Is it, does it feel non-linear? Then it's just like, there are some of it, but then all of a sudden it come, it becomes a lot more useful. Or it's just like kind of doubling every few years until you hit sort of saturation, like, does it look like?

[00:15:39] Ethan Smith: It's sort of an S-curve. So when, when you launch a piece of content, Google doesn't have information about the engagement of that url. So you test on a small number of keywords, look for engagement, and then as that engagement on per keyword, it's kind of like a quality square of an ad. Once you have some data once Google has some data, then they start ranking you higher.

[00:15:59] Ethan Smith: , so it curves up and then eventually it'll, it'll saturate and, and get most of the share of that set of keywords, and so then it'll kind of cap out. But that can take over a year. , the other thing is that search generally is increasing. So just the, the, the number of people searching for any given keyword is increasing.

[00:16:21] Ethan Smith: And so it's sort of an s curve, and then it kind of continues to curve up or not, not curve up, be linear at the, at the end of the S curve.

[00:16:28] Nima Gardideh: gotcha. And then you introduce new keywords and new keywords, so then you compounding over time like that. 

[00:16:32] Ethan Smith: Yes. And then the other thing is for, for your campaign overall, you launched a single piece of content and the, the curve for that is an S curve, but you're launching more and more content. So it's all these culative s of the various SSC curves.

[00:16:45] Nima Gardideh: Gotcha. Yeah, that makes sense. I, this is a bit of a tangent, but I'm just curious what you think, why, why do you think. Marketers don't do more N P V S analysis. Like what, why, what, what's the blocker do you think? Because I, I deal with this a lot and I want everyone that we work with to do it. , I would say 10% of the companies we work with barely do it get like anywhere close to doing some more of a sophisticated analysis.

[00:17:10] Nima Gardideh: Why, why do you think it's not the norm?

[00:17:13] Ethan Smith: Two reasons. So the, for the first reason, a lot of companies don't even have event tracking set up. So the idea of having a really complicated model. Let's just get our traffic and our conversions set up properly to start and know which channels drove, drove the conversions. , for some reason, tracking and, and, and metrics are very challenging.

[00:17:38] Ethan Smith: I'm not entirely sure why. I think it's probably because there's no single owner of it. It's sort of spread. The responsibility is spread across multiple departments and so it's no one's primary role. And so things just tend to be broken. And if they're broken, then why would I do a sophisticated model?

[00:17:55] Ethan Smith: So I think that that is number one. It's just lack of good data. , and then also data is in Looker. And then I go into Looker, but I don't know how to use Looker. And so then I have to use as the engineer for this thing. And that it's this very high friction metrics are very high friction.

[00:18:15] Ethan Smith: So then once I have the data, I want to do all this modeling. And I think that marketers are not, most of them are not super deep in statistics. So doing a net present value is something that they've never heard of before and it sounds intimidating. And , and so they just don't look at that. And then the last thing is, I'm probably not gonna be working at this company in two years, so why would I model out metrics where I'm not even gonna be here? Why would I make, why would I make a three year investment when I won't be here in three years?

[00:18:50] Nima Gardideh: It, it, it, yeah, that, that part is an interesting insight. I think like the first, I, I commonly deal with the, the, the parts where I feel like the companies that, that, that do it it tends to come from way higher up. So usually like the founders or the CEO of the company wants to look at programs in that way and, and enforces it into the whole org.

[00:19:15] Nima Gardideh: And that's kind of what's what I've seen work where they just, they want a spreadsheet and they want to see the MPVs and they want to see the reasoning behind it. And they wanna push back on your assptions that before they sort of like, like you go through with it. , and they tend to do better anecdotally speaking.

[00:19:34] Nima Gardideh: , but yeah, you're right. Like the, the tracking part is, is a mess. , I don't think I've been in a single company where, That's not a problem. , we have some like things that we talk about on how to, how to solve it, but it just seems like a people problem more than anything else. , and or a problem as you mentioned.

[00:19:53] Nima Gardideh: So yeah, it's, it's curious. I I, I, I want to live in a world where that neither of those problems and also like I, just feels like you can, you, there's like a big optimization to be to be doing in terms of how teams are run. Like we talk a lot about experimentation and running the process and these types of things in the industry.

[00:20:15] Nima Gardideh: , but this metric that is very well studied and totally different parts of company building and, and well, really interesting, the capital allocation world is not being used in, in this part of growing companies. But it has so many sort of parallels anyway, 

[00:20:34] Ethan Smith: Yeah. to summarize, I think that, or it's org design and prioritization. So org design, again, it's not a centralized function. It's distributed across many departments since it's hard to actually have a centralized metric system that works well and you have perfect data, and I think it's a prioritization problem because if someone says on a board slide, My plan for Q2 is to have better measurement.

[00:20:58] Ethan Smith: That's not a really compelling slide versus a slide where we're gonna launch personalization. That sounds a lot more exciting and, but better data means you that you or you're way more efficient with everything. So I think it's a prioritization thing to your point. Some companies prioritize it and most don't.

[00:21:14] Ethan Smith: And the ones that do are do a lot better.

[00:21:17] Nima Gardideh: Speaking of org design, I, I'd love to walk in through maybe a, an SEO team. So there's maybe two questions I have for you is, one is how would you, if you were to sort of build an internal SEO team, How would you design that team? What types of roles would that would be on it? , would you always work with an agency or not work with an agency?

[00:21:39] Nima Gardideh: , and then I'm just curious about graphite and how you guys design your team, your teams around clients.

[00:21:47] Ethan Smith: Yeah, so it depends on the type of company. So I'll define the three kinds of seo and then I'll describe how it's set up a team. So the first kind of SEO's, technical seo, technical SEO is the infrastructure layer. So crawl optimization and your tags and indexation logic and things like that.

[00:22:10] Ethan Smith: And usually technical seo. People spent too much time on technical SEO and not on other things. It matters, but it's not the most important thing. So technical SEO is the first kind of SEO second kind of SEO's programmatic seo, which are pages that are programmatically generated. So a category page like cameras, and there's a grid of cameras or product pages individually.

[00:22:32] Ethan Smith: The Nikon E 40 page or profile pages that's programmatic seo. So where these pages are generated programmatically versus a han sitting down and writing them. And then the third kind of SEO's, editorial seo. An editorial SEO is where someone actually sat down and write something like an article or a guide or a listical.

[00:22:53] Ethan Smith: And all companies can benefit from editorial seo, but only some can benefit from programmatic. So for peer mill or for BetterUp, they don't need programmatic seo. They just need editorial seo O and they don't really need much technical SEO o either. It's mostly around the editorial side. And so that team's gonna be very different than the team that works at Target.

[00:23:17] Ethan Smith: So on the editorial SEO side, the way that I would build that is it would be it would primarily be on the content side. So it would not, you don't need technical expertise. You don't actually even need to know that much about seo. What you need to know is, is how to write really good content and you need some sort of strategy around what to write about and how to write it.

[00:23:37] Ethan Smith: But it's not deep technical SEO expertise. And so at Masterclass, for example, the team was excellent at, at content and they didn't have a lot of SEO experience. And that was totally fine, and that was actually better. So all they needed to know was what to write about and what the structure looks like, and then they can do really well.

[00:23:54] Ethan Smith: So more specifically, how I would structure that team would be, it needs to be separate, but aligned with the brand content team. So the brand content team, a, a common issue is that the brand team needs to be really careful about making sure that the brand is always described in a particular way consistently.

[00:24:15] Ethan Smith: So you want to be cautious. You don't wanna describe yourself, like if you're Grammarly, you don't wanna describe yourself as a spell checking tool. It's a writing enhancement tool, and you wanna be really careful about that. , but for, for editorial, SEO and growth, you do not wanna be careful. You wanna be very aggressive and you want as much traffic as possible.

[00:24:35] Ethan Smith: Possible. You don't wanna be too cautious about, well, we don't want to write. For better up. We don't wanna write about five year plan or something else cuz we don't have five year plans, but that's fine. We want to get as much exposure as we possibly can. So there's brand content and growth content and neither is better.

[00:24:52] Ethan Smith: They're just different. And so growth content should be aggressive and brand content should be really cautious and careful. so we'll usually say, well, let's generally be aligned on how we talk about the brand. But the growth content team is a separate track and it should be aggressive. And so that team is gonna be a content manager, editor type of a person, content strategy person.

[00:25:15] Ethan Smith: And that's basically it. So that person manages a group of writers, you're prioritizing which topics to write about, outlines, optimization of title tags, measuring that, but it's more of a content role than a pure seo. That person can be in-house or external. I think that the, key thing that the content person needs is just to know which pieces of content we should create, which topics and the outlines, and that could be internal or external.

[00:25:42] Ethan Smith: , so that's how I would set up an editorial team. And, and, and again, that could be with or without, with her or without an agency either. Model's fine. Programmatic SEO is completely different. And, and, and for, for the editorial SEO side, none of this is touching the core website. You don't need to meet with other teams to , change the product page and change the category pages and getting to the core code base, this can be completely a completely separate autonomous track, programmatic seo, completely different.

[00:26:11] Ethan Smith: So if I need to change the category page on Walmart, I need to have multiple meetings with multiple departments asking question. This touches many parts of the org. I can't just. Make my own category pages and do whatever I want. So now you need to be part of the, the core product. And , the common problem with this is that the product team for the category page is not necessarily doesn't have their goal around SEO traffic.

[00:26:38] Ethan Smith: So if the SEO team says, I want this thing, that doesn't necessarily get prioritized because the, the product team has a bunch of other priorities. And so frequently it's just hard to get things prioritized. Or there's a conflict with, I wanna show 50 products, but the core product team wants to show 20 products because the page is faster if there's 20 products.

[00:27:01] Ethan Smith: So you have competing priorities. , direct conflict of priorities and also just hard to prioritize that stuff. So how to do that? Well, I don't think that there's a single correct answer, but option one is that Dec core product team has a goal around impact and results and conversions. Like, I want more products sold at Walmart and therefore, as it would make sense to have SEO be a part of that.

[00:27:25] Ethan Smith: And so that could be just that product team owns seo or there's a separate SEO specialist team that says, here's how to do this. , show 50 instead of 20, show more tags, add links, and then it gets prioritized. Well there's no single correct solution, but that's typically how I see it.

[00:27:42] Nima Gardideh: Is there a version of it where the category pages that people see coming in from SEO would just be different than ones that you see if you're logged in, for instance?

[00:27:55] Ethan Smith: Yes, you can definitely do that, and we've done that before. , where the the, the internal product page, an example would be the internal product page is rendered with JavaScript and it's asynchronously loaded, and different components are being loaded at different times, and Google can't see any of that.

[00:28:13] Ethan Smith: The page is 

[00:28:14] Ethan Smith: empty, which is, which is not uncommon. So we could either say, please engineering team, please completely rebuild this page just for our SEO team, which is a, which is a hard ask. Sometimes it happens, but it's a hard ask or SEO team could say, well, we're, we'll just make a static version of this, or we'll make a separate SEO landing page version of this.

[00:28:35] Ethan Smith: And and that can work really well. But then you need to talk with the engineering team and say, Hey, we're gonna duplicate the exact same thing that you already did with this different code base just for Google. That could work. But it's also sometimes a difficult conversation to have. But that, that is a good solution.

[00:28:53] Ethan Smith: And, and we've done that before, so I, I do recommend that as something to consider.

[00:28:56] Nima Gardideh: And would you say, so that's an interesting thing around the, the loading part that is, is Google still, and it's 2023 when we're having this conversation, not waiting to render JavaScript based , pages. They're still looking for fully static page.

[00:29:14] Ethan Smith: No, they look at JavaScript, but they, they struggle to index everything on the page frequently. So there's two passes of the crawler First crawl. , first pass is without JavaScript server side second passes with JavaScript. The common problem is that for the second, and, and if your page is completely in JavaScript, that could be totally fine.

[00:29:39] Ethan Smith: It oftentimes is like, I think Pinterest was and probably still is completely JavaScript based, and that's totally fine and everything's indexed properly. There's two problems with that happen with that though. So the first problem is hybrid rendering. So this is where. , part of the page is server side, and part is JavaScript and the crawler.

[00:30:00] Ethan Smith: If they think the crawler thinks that they got everything that they need with the server side rendering without JavaScript, they will not do the second pass. And so if you have a an article, we we actually saw this specific example. So if you have an article that is that we had an article and, and supplementary content below so that the article was, let's say 20 words, it was a poem or a haiku poem, it's 20 words.

[00:30:30] Ethan Smith: Google will say, I didn't get everything I needed. I'm gonna look at the rest of the page and index that. Now the article is 300 words. Then Google says, I got everything I need, and I will not look at the supplementary contents, like the cross-linking of different articles. So the length of the server side rendered.

[00:30:46] Ethan Smith: Matters. The shorter it is, the more the crawler says I need more, the longer it is it says I don't need it. And it is common to have part of the page, especially the top part indexed. And then all this cross-linking and supplementary content is ignored. And so now you have a product page with no cross-linking.

[00:31:05] Ethan Smith: So you have the description on the price, but there's no crosslinking because of this 

[00:31:08] Ethan Smith: hybrid rendering issue. Or, or, or reviews are not crawls like reviews are asynchronously, loaded in and they're not seen. And we see this frequently. , so hybrid rendering is problem one. Problem two is everything's on the loaded on the JavaScript side, but it's asynchronously loaded and maybe the bot, maybe the reviews are not loaded until there's a user scrolls to save on performance, which is very reasonable.

[00:31:35] Ethan Smith: Google can't see that then. And so we see this happen regularly. So Google can read JavaScript, but it depends on how you're loading it. And you can QA this by doing, by taking your URL and then doing site colon url and then taking different parts of the page and then having an exact match snippet.

[00:31:55] Ethan Smith: So if you want to know if your reviews are getting indexed site Co on that url, and then take a snippet from a review and then put it in quote marks, and then look to see if Google has it. And if they have it, then it's indexed. And if they don't, then it's not.

[00:32:09] Nima Gardideh: That's cool. So within the Google search bar, you're doing site call and putting the full url. 

[00:32:14] Ethan Smith: Yes. And you cannot, you cannot just look at u URL inspection or the cash version, because the cash version does not fully represent what's seen. And so sometimes the cash version is empty, but it's actually all indexed. And the URL inspection tool in search console is good. But it does not tell you this.

[00:32:31] Ethan Smith: The only definitive way to know whether or not things are getting indexed is this cycon quote exact match of it, which is obviously , an odd it's not a, it's not a clean analysis , way to do it. But I mean, it, it does work. And that's the 

[00:32:46] Ethan Smith: only, that's the only, definitive way to check that.

[00:32:48] Nima Gardideh: yeah, cause my understanding was that they had started crawling JavaScript stuff, but it sounds like there's a lot of nuance there. , would you, if, if you were to start from scratch, would you just not do JavaScript and still recommend static? Or as long as  about these things and , it sounds like there's a couple of like obvious routes to fix these stuff up, this stuff up.

[00:33:09] Nima Gardideh: , you would be okay with JavaScript based pages.

[00:33:13] Ethan Smith: Yeah. And just to be clear , I, I, I should, I should clarify client side rendered versus server side rendered. So would I not do client side rendered, rendered? I, I would not recommend. I, I would not recommend against client side rendering. That's sort of a broad stroke against a pretty, pretty good technology.

[00:33:32] Ethan Smith: So I wouldn't just broad stroke. Don't use this modern technology for seo.  build, build, however whatever you think is the best way, and then QA it and if it's not working, then, then fix it.

[00:33:44] Nima Gardideh: obviously, there was like a sort of a rush of libraries that came about just to solve this. I think next JS was a maybe a popular. , coming in to help people produce pages on the server side so they're not re being sort of re rendered on a client's side.

[00:34:00] Nima Gardideh: And, and now their new version kind of just as a hybrid which I like a lot. , cool. So that, this makes sense. So there's programmatic and, and in that team it sounds like there is like a product lens. So are these product managers that are doing it, is it like a SEO specialist that's embedded in a product team?

[00:34:18] Nima Gardideh: Like what, what, what are the rec what you, what you seeing and what do you recommend?

[00:34:22] Ethan Smith: Ultimately the decision will come down to whoever the product person is for that category page. The SEO manager cannot make that decision. So that decision is from the product manager. So I ideally that product manager who does not need to be an SEO expert, but, but they, the only category page is they're product expert.

[00:34:42] Ethan Smith: They're not an SEO expert. Their goal needs to be around, their goal needs to be SEO needs to be a part of their goal. Probably something around conversions. So the, the goal and the incentives need to be aligned with this. And then I would as embedded as possible, ha having that SEO o person.

[00:34:59] Ethan Smith: So that could either be an internal SEO person who's on that team, or an external group who's giving the strategy to that product person either can work. But the decision was is not what the SEO manager, it's what the product manager and the goal that product manager needs to be aligned with impact and SEO in order for this stuff to get prioritized.

[00:35:20] Nima Gardideh: cool. So let, let's go into sort of, okay, let's say you have a team now and you want to go after a, a bunch of keywords. Do you think in keywords, I heard you talk about topics like, first of all, what is the difference between a keyword and a topic? , and how do you get started?

[00:35:37] Nima Gardideh: Do you do, do you just like into it? Or If I were to search for this product space or like the product that I'm trying to sell, Here's the set of keywords I would search for. Like how, how do you go for like step one of this.

[00:35:51] Ethan Smith: Yeah, so the difference between keywords and topics, keywords is a single individual search query, like five year plan and topic is that concept. And the concept of five year plan has 500 different keywords that are minor variations of each other, like five year plan, example of five year plan template, personal five year plan.

[00:36:09] Ethan Smith: There's all these different variations of how to describe the concept or the topic. So the topic is a one-to-one map with that url. So this r url, this single URL or single topic is targeting many keywords. And so that's, that's the difference between the two. And your second question was how, how do you think about the strategy?

[00:36:29] Ethan Smith: So, I wanna rank for five year plan, or let's say I wanna rank for five year plan template. Then I look at the search vole. The search vole is a hundred searches a month. It's not that many. So I say, I'm not gonna, I'm not gonna make this piece of content. But in reality, the search vole is 50,000 because you didn't know all of the different keywords.

[00:36:47] Ethan Smith: And the, and, and the search vole for all of them is much larger. So we have internal tools that do this, but essentially what you'd wanna do is take five year plan template, map that to a set of URLs. So there's 50 different articles about five year plans. So you look at all 50 of those, and then you look at the keyword overlap of all those, and then that is your keyword cluster.

[00:37:12] Ethan Smith: Then the vo, the opportunity is the s of the vole of all of those keywords, the 500 of them. And number one. number two is that the piece of content needs to target this entire set of 500 different keywords. So, templates, examples, personal versus professional. , so that's generally how we, we do topic research.

[00:37:34] Ethan Smith: , that is obviously a lot of work and it's imprecise. So when you hire an SEO person, usually their job is to go into s scm, rush and download a bunch of keywords, put it in a spreadsheet and make guesses about what keywords go with, with what. And they can't go through hundreds of keywords for one, for one year url.

[00:37:54] Ethan Smith: You can't go through 500 different keywords manually, and you don't actually know what should be grouped. So maybe professional five year plan should be a separate article and you don't actually know that. And so you might write. Five articles when you only needed one, or you might write one article when actually you needed five and now you can't compete for a bunch of the keywords because you didn't know this.

[00:38:14] Ethan Smith: So the measurement error again, of this keyword clustering is very large and it's impossible to do manually. So on our side, we built internal tools to be able to do this clustering. But essentially the clustering is, again, for five year plan example, find the 50 different URLs that are ranking for this.

[00:38:32] Ethan Smith: Find out the keywords that they're ranking for, and look at the keyword overlap of all of these. And that's essentially what your keyword cluster is.

[00:38:38] Nima Gardideh: is this data from Google itself or you're pulling this data off of like different sources and I asse there's some of that information in, in Google's APIs alone, right? , where they help you with keyword 

[00:38:52] Ethan Smith: , no, there's none of, none of this comes, none of this comes from Google, voluntarily giving you the information. , the way that this comes is with stir scraping. So you take a, a URL for a five year plan, and then you scrape, you, you get a bunch of keywords. The keywords would come from, it would probably come from , Google Ad Planner, and you have your own internal keyword data.

[00:39:14] Ethan Smith: So then you take all these different keywords, like five-year plan template, five-year plan, example. Then you scrape Google as s scm rush or AHS do. So scrape Google, look at the different URLs that are showing up, and then do this for all the different keywords that you can find or that you have.

[00:39:30] Ethan Smith: And then again, you're looking for which URLs are overlapping with each other frequently. So for five year plan template and five year plan examples eight of the top URLs are, are the same. Therefore, it's a group. Therefore, these two keywords are related to each other. But, but none of this is coming.

[00:39:46] Ethan Smith: Almost none of this is coming from Google. It's coming from THP scraping, essentially.

[00:39:50] Nima Gardideh: And SE stands for.

[00:39:52] Ethan Smith: , CERP stands for search engine result position. So this just means that this just means for five-year plan examples, which URLs are showing up in what order. So this article shows up number two. So their crp position is number two.

[00:40:07] Nima Gardideh: Gotcha. So you're basically creating these topics or like The idea is basically there is like if I were to maybe abstract it, there are these words that correlate a lot with these sets of other words to Google. And so when I'm searching them I'm kind of finding these sets of different URLs and if you wanna compete on that general, Topic, then you should be producing content that doesn't just address these like obvious keywords, but there's all these like cluster of ideas and keywords that are related to it.

[00:40:44] Nima Gardideh: And so are you mentioning all of those 500 different keywords in one article or is it just more like you're alluding to them? Like what is the content strategy from that idea?

[00:40:53] Ethan Smith: Yes. So this comes this, this gets into content outlines or topic outlines. So for five year plan, we have 500 different keywords. So the simple idea is make sure that every single word in all 500 is mentioned. That does not lead to a good piece of content that's useful. And it's also not how the, how, how the algorithm works.

[00:41:18] Ethan Smith: So what you actually want to do is you wanna synthesize all 500 and look for themes and write about those themes. So users are generally looking for examples of five year plans and templates and professional versus versus  personal.

[00:41:35] Ethan Smith: Those 

[00:41:35] Nima Gardideh: and this specific topic cuz you've, you've already done this work about this five year plan thing.

[00:41:39] Ethan Smith: Yes. , but when you go through your 500, you're looking for general themes of subtopics, and this informs the outline. And this is the second job of the SEO manager, or, or, or the content manager is, tell me, tell the writer the themes. So if you don't give examples of a five year plan and people are looking for that, you have a gap in your content and it's not comprehensive.

[00:42:01] Ethan Smith: So part of how, how the algorithm is looking for good content, quote unquote, is comprehensiveness. Did you cover everything or did you not, did you need to go to another result because you didn't have this thing that the users wanted? Or did you have everything? And so the content score is basically looking for this comprehensiveness of, of all the different subtopics.

[00:42:22] Ethan Smith: And again, it's very hard to go through 500 of these. And so people don't do that. , and the result is that they're missing things. And if you do do that, you get a better content score. And so it can be a huge competitive advantage. , so that's one thing about comprehensiveness. And then the second thing that Google's doing is they're looking for co-occurring terms.

[00:42:43] Ethan Smith: So by that I mean basil and lemon co-occur disproportionately on the internet. So if you have a recipe about chicken piccata, make sure to mention lemon or make, make sure to mention basil because these things co-occur. The, the, and this is to how some of the tools like a clear scope or a market muse work, which is they're looking at given a keyword.

[00:43:09] Ethan Smith: What are all the dis , overrepresented terms that are, that are appearing? So for five year plan examples is , co-occurring. So make sure to mention that, but you don't wanna look at it at the keyword level. You want to, you want to abstract it at the concept level. So don't just a add this word, cover the general concept of examples.

[00:43:30] Ethan Smith: And Google actually has said that they specifically look at the subtopic and paragraph or concept level, not individual words. So again, you don't wanna just add words, be because , add the word examples five times you wanna cover the concept generally.

[00:43:45] Nima Gardideh: It's almost like the information about what is going to make this thing comprehensive and so like, yes, lemon being there really just means like make sure the full recipe is there and you're describing how to make it and not just put the word lemon in there cuz that's gonna make it Make it 

[00:44:01] Nima Gardideh: work. 

[00:44:02] Ethan Smith: a and so this is, this is how to get a good content score with the algorithm, but it's one-to-one with what users want. So it's also helpful for users. Users want examples, and if you have examples, then you're gonna get better engagement. Then if you don't have examples,

[00:44:16] Ethan Smith: therefore your engagement score is higher, therefore your rank is higher.

[00:44:20] Ethan Smith: Engagement is a, is a high weight input as well. So for both the content score and for engagement, you want things to be comprehensive and you want them to be useful and high quality. If you have thin, non-useful in content, maybe your content score will be fine, but your engagement will be bad, and then you'll, you'll, you'll rank less.

[00:44:37] Ethan Smith: So having high quality content that's comprehensive and useful , directly results in impact.

[00:44:43] Nima Gardideh: Yeah, that makes a lot of sense. It's similar to actually, like, I think ads, we have quality scores and it's like similar thing if you're ending up converting, that's a very positive thing and it's a very positive signal for your ad and, and the quality of it. , even both, almost every ad network basically has the same approach.

[00:45:03] Nima Gardideh: , so, so I'm quite familiar with that. It makes a lot of sense to me. 

[00:45:06] Nima Gardideh: , well, so, okay, so now, now that we've done this and people are arriving on this site, And they're getting the value. So they search for like here's an example. I want a five year plan. And you, you go in there and there's like a very clear template for how to build a five year plan.

[00:45:23] Nima Gardideh: , I personally struggle to then build that into a conversion. Like I, I, I think whenever I've thought about these SEO pages, it's on bit unclear to me how to organically take that into telling the user now who we are as a company and you should use us or work with us, and so on and so forth. So are there like heuristics you use on, on thinking about that as does the topic matter is different per maybe like the concept and how how the intent of that topic is to the, so to the product that the, the, the company's providing or the service the company's providing? Or is it the same, it doesn't matter, just put the signup button somewhere and usual cro things that it's gonna work. Like, how do you think about that part?

[00:46:12] Ethan Smith: There's a measurement piece and there's an optimization piece. So the measurement piece is multi-touch attribution. So for Target, last touch is fine. You can just look at the conversion rate for last touch, the people who came in, how many, how many bought a product. And that's pretty straightforward. For the BetterUp example, you need to do multitouch ultimately to have proper, proper attribution.

[00:46:35] Ethan Smith: , but you can do a lot to optimize. So yes, to doing typical cro o things, but most the highest impact stuff on cro R is not obvious and it's typically not done. There's a a ton of upside. , high quality conversion optimization. So essentially what you're doing is you, you wanna build the intent of the user over time.

[00:46:59] Ethan Smith: So for BetterUp, it's not, I need to get them to subscribe and pay right now, but I generally want to build their intent over a long period of time. I do that and, and, and a lot of SEO content does not do that. So it's common to search for something, land somewhere, get useful information, leave and have no idea where you read that, and that V view had no value at all.

[00:47:21] Ethan Smith: But you wanna build the intent of the user. It's kind of like if you have a TV ad and people love the ad and you never mentioned the company that would, that the ad was for, that would not be a useful useful tb. O you want to, you want to tastefully remind the user and have them remember that the ad was for Ford trucks versus it's beautiful mountains and it never mentions Ford.

[00:47:44] Ethan Smith: So the way that you can do that the, the bad way you can do that is put a big popup and say, you need to subscribe to my email newsletter. , people don't like that, so we, we don't, we don't recommend that. But you wanna understand the intent of the user and try to have conversion elements and intent building elements that are related to that intent.

[00:48:04] Ethan Smith: So again, newsletter, if I'm searching for five year plan newsletter is not my intent, but my intent. Well, I would like a template that I can download four or five year plan, and then I can start creating my plan. And BetterUp is mentioned as a there's the logo and the download of the template.

[00:48:21] Ethan Smith: Maybe we get your email when you download it. HubSpot does a good job of this. They have a lot of templates that you can download, artifacts that you can download on each page. It's branded as HubSpot. , you sign up as a in order to get that, I want to sign up because I want this template. It's not because of a newsletter I don't want, I have a goal of actually making something.

[00:48:42] Ethan Smith: So I do want to download this so that the, the, the conversion element is directly related to my intent. , so downloadable artifacts can work well. Quiz flows can work well. , widgets like a mortgage calculator videos to just have a video where you can watch it and see what BetterUp generally does.

[00:49:03] Ethan Smith: , images with a logo on it. So rather than an image with no logo on it, put the logo there. All of these things are related to your intent, and they're tastefully done. They're not, they're not intrusive. And now the val, the value of that visit is substantially higher. It could be 500% higher, a thousand percent higher, just like, again, the TV ad that never mentioned Ford trucks.

[00:49:22] Ethan Smith: If you mentioned Ford trucks over and over again, tastefully, the intent goes way up

[00:49:27] Nima Gardideh: Yeah. And then that number you just said like 500% higher, it'll show up in a multi-touch attribution model. If you show that you, you're making these changes in a way that brings the value of these articles up in terms of conversion and not just in terms of getting people to come to these pages, it sounds like.

[00:49:46] Ethan Smith: for sure. And then the other thing that I would say is, so, so this is intent building on the page. But then you can create a, a loop. So the loop would be, I got a visit that comes in on seo, now I can pixel them and now I can retarget them. So I didn't have that user pixeled, and now I do, because of this SEO event, or I got an email I, I got somebody to sign up to or give us their email to get this template.

[00:50:11] Ethan Smith: Now I can market them with email marketing. So if you can create these loops, or, or if somebody shared my content because it was useful on social, now I get some traffic from social. So not only do you wanna build the intent on that view, but you wanna create loops across other channels.

[00:50:25] Nima Gardideh: Yeah, that makes a lot of sense. You just put, wanna put it in the whole ecosystem of how you're trying to market folks. , one question around this. So I think generally the process makes sense to me. 

[00:50:36] Nima Gardideh: Now if I'm early in this process, I've just hired an agency, or I'm just like working on this, or even maybe in later stages where you're trying to change the dynamics of your organic traffic.

[00:50:50] Nima Gardideh: How do you, what, what are the sort of metrics you would use to keep the team accountable on something like this? Is, is it just production? Are you looking a little bit deeper than that? Is it just top line at the very list, least visits or, or, or revenue. How would you sort of judge an SEO team's performance?

[00:51:12] Ethan Smith: For late stage, you can just look at conversions. So for Target, I want to know how many products we sold. , and, and that's how you would judge that. How many products we sold from this channel. For an early stage company, that's not possible. So we usually say do, do not focus on conversions from the beginning because you didn't have time to even build a conversion.

[00:51:37] Ethan Smith: On the page, focus on visits first. Once you get the visits, then focus on conversion. , and before visits come, it's the leading indicators that I described. So if I'm running if I'm doing SEO at Pier Mill, and we haven't done SEO before milestone one is I want a few pieces of content that are ranking number eight for, for a set of terms.

[00:51:59] Ethan Smith: And maybe there's a very small amount of traffic, but I have these leading indicators that with a few pieces of content I can rank competitively. So that's milestone one is, is ranking for key terms for a small set of content that's milestone one, milestone two. Ranking for more things. So I had five articles, milestone one, I have 50 articles or 50 landing pages, milestone two, and now the traffic is starting to curve up.

[00:52:27] Ethan Smith: And now I know if I go from 50 to 500, I can get even more. So milestone two is, is getting some amount of traffic 50 pieces of content. Milestone three is traffic is now growing. I have 500 pages. Now I can start working on conversion. And now I wanna start looking at total number of conversions, probably total n, number of conversions per piece of content or, or, or overall.

[00:52:51] Ethan Smith: , that's milestone three. And then milestone four is just growing that as much as we possibly can.

[00:52:56] Nima Gardideh: And in the first milestone, you mentioned ranking for a set of like sort of high value keywords. Is it ranking at all that were showing up on like page two, or are you saying like competing at the top, like five of the results? Like what is that feasible in the early days of, of this process? Or is, is that something that comes after you're sort of getting more authority over time with Google?

[00:53:22] Ethan Smith: It is possible to rank on page one for competitive terms, even if you're very early. So Pi Mill, for example, I was looking at some of your keywords in your ranking competitively with, with very little authority for LinkedIn Pixel Helper 

[00:53:36] Ethan Smith: which actually has a, a, a decent amount of search vole

[00:53:39] Nima Gardideh: Mm-hmm.

[00:53:40] Ethan Smith: and you're, you've done very little seo.

[00:53:43] Ethan Smith: So we also work with a company practice.do, which is a a coaching software platform. So if you're a coach, you can sign up on, on practice to be an online coach and run your business. And we started with them. Right when they started, they had very few links. They actually still don't have that many links and they're able to rank number one for coaching sessions and top five for some very competitive terms.

[00:54:08] Ethan Smith: , so if you're only ranking on page two, I think that that, that's not a great goal. You wanna rank on page one for some competitive terms, and that is possible early on. It's hard, but it's possible.

[00:54:18] Nima Gardideh: Yeah, that makes a lot of sense. Yeah. Going back to as , I'd, I'd love you to go through more of what you found cuz , you did some work on para and sort of your thought process and what you discovered and, and just to close the loop around that LinkedIn pixel he helper thing, I was basically a Chrome extension.

[00:54:36] Nima Gardideh: We rebuilt to help people find out what With pixels they have on their site for LinkedIn, and it was recently actually brought down by LinkedIn. They didn't like that we would lose, use their name in the Chrome extension. , which we might, we might relaunch it, but I think we, it has something like 20,000 active users still just because of that page that we put up in organically for hit.

[00:55:00] Nima Gardideh: The, one had this product, so it was just a blue, blue ocean thing and it just, but a lot of people wanted it. , and it immediately became quite big.

[00:55:08] Ethan Smith: Yep. And then your question was just how, how would we generally use that and do more with it?

[00:55:13] Nima Gardideh: Yeah. So now, now that you, you've seen the, this, this, this set of keywords are working, we have all this other content that we've produced without really thinking of it in this way, is the way we've thought about content production historically has just been in the sales process. What pieces of content.

[00:55:30] Nima Gardideh: Could we use to keep folks warm? Right. And, and for context , we're going after tech companies and, and these slightly bigger companies, like the sales cycle, as you had mentioned earlier, for a company like ours is very long. , and we, we don't want them to dis, ideally maybe they discover these articles by themselves, but a lot of the work we do is just our sales team will send the articles to the leads in which the article is very relevant to and then use that to show our authority of how we are good at growth and good at C R O and production of creative and so on and so forth.

[00:56:06] Nima Gardideh: So we haven't really looked at production of content from an SEO lens. , but that probably means there's all this information sitting there that we could use to rank for some keywords that we would want to be ranking for.

[00:56:20] Ethan Smith: Yes. And what I would do is because Pure Mill hasn't done a lot of seo. You don't have a lot of authority, but you wanna look at where you have the authority and you wanna do things that are adjacent to that which, which gets into topical authority. So currently peer mill ranks for LinkedIn, pixel helper, LinkedIn insight tag, Chrome extension.

[00:56:43] Ethan Smith: And so you're gonna rank for things related to LinkedIn Pixel generally. So beyond just pixel helper and tracking or, or an extension specifically to do that, you'll generally rank for things related to link LinkedIn pixels. So now you can create a guide on how-to content about LinkedIn pixels or we want to get people who are really high intent and who would want to use Pyramid.

[00:57:06] Ethan Smith: So people who are looking for specific niche problems around pixel tracking issues is probably very high intent. And so now we could, and we can rank for those even though we haven't done much seo, 

[00:57:18] Ethan Smith: , because of this tool, even though it doesn't exist anymore. So now let's find topics that are adjacent to LinkedIn.

[00:57:24] Ethan Smith: Pixel helper broadly, which is LinkedIn pixels. , Topics in general. So let, so let's go through all the different problems with tracking pixels. Let's generally go through LinkedIn tracking as a general subject, and we'll probably actually rank quite well for that. And then if we want to go into Instagram ads, then we can have an Instagram tool and get some topical authority there because of the tool.

[00:57:48] Ethan Smith: And now we can have how-to guides about various issues around Instagram tracking. , and then if we want to go into TikTok, same thing. , again, because we're early, we don't have a lot of authority. Good way to build authority is with these, these tools that people share. And then you do stuff that, you do articles and guides that are adjacent to those areas.

[00:58:09] Nima Gardideh: Gotcha. That makes sense. So it's almost like there is the beginnings of a fire in there. Let's just pour. Gas there and just go around it and use that initial authority that we built on, on a set of keywords and then go into the tangential topics. Well, first it sounds like subtopics and then next the tangential topics around those, so then we can continually expand the authority around that.

[00:58:36] Nima Gardideh: , how, how does authority work if we're just to double click on authority in general? , it sounds like there is some of it as just people coming to your site and getting value. So that, I understand how Google will track that if they click on some search result and they end up not having to search more around that topic.

[00:58:55] Nima Gardideh: They clearly have like a signal that this was a very useful article.

[00:58:59] Nima Gardideh: I remember there was like a craze for backlink building. Is that still something that goes on? Are there other things I just don't know about? Like maybe. , video content being about your product or something like that, that, that Google has put in the ecosystem on how they understand and put authority behind your brand.

[00:59:19] Ethan Smith: Yeah, there's two different pieces. One is a direct input of authority, and the other is engagement. So on the engagement side, when people search for LinkedIn pixel helper and they click on you, if they go back and go somewhere else, that's a bad engagement signal. But if they stay on pyramid and engage with it and maybe download something, that's a good engagement signal.

[00:59:39] Ethan Smith: And so the more people have good engagement, the more that you'll have engagement authority for that one keyword. And then they'll te Google will test it on things that are adjacent to that, like LinkedIn pixel tracking. They know that those are similar or the, the distance between those two keywords are is, is short.

[00:59:56] Ethan Smith: So therefore they're gonna test you on that. And then if people click and stay on you disproportionately, now you're gonna rank better for that. Then they're gonna test you for LinkedIn tracking generally, and now you have good engagement for that. And so that your engagement authority score is based on people clicking it on you.

[01:00:13] Ethan Smith: And staying and getting the answer. Then there's a separate set of inputs, which are things like backlink shares, branded search that also feed in branded search is probably the, the largest, most impactful authority signal other than engagement. And so that would be peer mill LinkedIn or peer peer mill Instagram.

[01:00:34] Ethan Smith: So the me the inclusion of your company name and then a modifier. And that's essentially like a backlink that says Instagram for branded search. And the more people that type in it, let's say the 10,000 people typed in peer mill LinkedIn, you're gonna rank really well for LinkedIn within days. And this is again, people going to Google and typing in peer mill LinkedIn.

[01:00:56] Ethan Smith: You're gonna rank for things anything related to to LinkedIn. , so that's branded search. Back links would be the new or u s A today linked to peer mill. And it said LinkedIn, you're gonna rank better for things containing LinkedIn. You wrote an article that went viral about LinkedIn tracking, and it got shared a lot on other platforms like Twitter and, and Pinterest and Facebook that's gonna give you authority.

[01:01:20] Ethan Smith: , one interesting example I'll give you, we we, we on the engagement side, so I asked the team, we, we haven't actually ironically done any SEO for, for graphite. , so we, we just built a website in December December, November. We, we basically didn't have a website before and I wanted to rank for SEO agency.

[01:01:44] Ethan Smith: And so the title tag before I just set graphite. So the first thing we did was at SEO agency, did the title Tech. So now we've started ranking, but we ranked number 70 

[01:01:55] Ethan Smith: for SEO agency, which is not very good. So then I, I, I tried to test where I Ask the team to go to Google and type in SEO agency and go all the way down to page seven and click on us.

[01:02:11] Ethan Smith: And we had about, let's say 50 people on the team do this. And within a few hours we went from number 70 to number four on Google for SEO agency across the entire United States ahead of Neil Patel and companies with tons of authority. We probably have like 20 back links almost nothing but, but basically because the clickthrough rate was disproportionately high for number 70, like nobody goes to number 70 and clicks 50 times in a day.

[01:02:42] Ethan Smith: The signal said the clickthrough rate is really high, so I need to rank them much, much higher. And this is a very competitive term with SEO agencies who are specifically targeting this, who are good at seo. And we were able to get to number, number four. I think we actually maybe got to number three. And we stayed there for about two days and then we dropped all the way we dropped down on like number 30.

[01:03:04] Ethan Smith: But this is just a, a very tangible recent example where click through rate on Google has this, has a very high

[01:03:10] Nima Gardideh: Very high. Yeah. Wow. That's quite interesting. And that it feels like an easy sort of like, let's boot this up. Let's just see what our. , numbers are gonna be after we boost it up a little bit by just getting a bunch of people clicking on it. , and do, are they smart enough where they're on the same office or you have a remote team?

[01:03:30] Nima Gardideh: And so it, it helped that people from like all around the US were clicking on something like this. I asse they're smart enough to realize everyone in the company is clicking on it in, in this one IP address. , they're not going to sort of put that and use that as a signal.

[01:03:46] Ethan Smith: Yeah. Google's very sophisticated about looking for actual clicks and I mean, these are actual people around, around the US and Canada who are searching for it. It's not like we're. , we're not fooling the algorithm. Like the, the, the algorithm would wanna know if, if real people are searching for something and clicking on a result, it should rank higher.

[01:04:06] Ethan Smith: , it's very hard to manipulate and to do with bots it won't work with hiring a bunch of people used to do this with Amazon Mechanical Turk, where they have tried to manipulate click through rate, and it did work for, this is 10 years ago. It, it did work briefly, but they shut that down.

[01:04:24] Ethan Smith: So they're very sophisticated in looking at things like the IP is the same. So I wouldn't suggest invest time trying to figure out how to spam it because it'll be. Al nearly impossible. What I would say is get real people and, and, and again, our team did this, and then we dropped back down in, in two days.

[01:04:45] Ethan Smith: So you need to keep on happening and we're, we're not gonna ask our team every single day to search for SD agency and click on us. , so you want real people doing this if you want this to actually last. And the way that you can do that, number one, is to just have good content that people would wanna read and you had comprehensive content that was useful.

[01:05:04] Ethan Smith: And then the other one is to have a compelling title tag. So this is why people do 10 best and things like that is the clickthrough rate goes up if the title tag is really compelling. So spending a lot of time optimizing your title tags and having them be clickable same with ads. , having that be clickable will increase the click through rate.

[01:05:22] Nima Gardideh: Interesting. And so for that, something like that you're just looking at the click through rate of keywords on your search console. If you're running experiments there, you can simply just change it and then look for the change posts.

[01:05:36] Ethan Smith: You can do that. You can also just look at traffic. So if you're on page one and you, we, we did this test with Thbtack. So we had plumbers in Santa Monica and then we tried 10 best 20, best 30 best 50 best, best top, the best, the best of 2023. And , you can, you can look at click through rate and search console, and then you can also just look at traffic.

[01:06:01] Ethan Smith: So if the visits go up, then it's working.

[01:06:04] Nima Gardideh: , can you share which one won? And that group of tests, or is that trade secret?

[01:06:10] Ethan Smith: I can, and I mean, if you, if you look for plumbers in Santa Monica, you'll see a lot of top 10 lists, and that's because the top 10 lists do do the best. So what we found is that 40 did better than 50 and 30 did better than 40, and 20 did better than 30, and 10 did better than 20. , and 10 and five did equal.

[01:06:29] Ethan Smith: So 10. Basically fewer is better because you don't, as the user, you don't want to go through 50. , 10 is a good number and five is a good number. That that's enough, that that's not too many to go through, but it's enough so that you have options. , so five to 10 usually does the best

[01:06:47] Ethan Smith: and best does better than top,

[01:06:50] Nima Gardideh: Yeah.

[01:06:50] Ethan Smith: slightly better than top.

[01:06:53] Nima Gardideh: , cool. I had, so as you were saying this, there was like a part I of maybe we should, we should address before we wrap it up, which is sort of what not to do. , and I think there's like plenty of maybe information out there on how to be good at seo and some of them are very bad and, and can penalize your your success.

[01:07:15] Nima Gardideh: So what are the, maybe things that are commonly spoken about in the industry, but you just believe it should not be done anymore? Or it's just an old version of how Google used to rank sites and, and they should be ignored.

[01:07:30] Ethan Smith: I would say that I, I wouldn't say that there's a lot of things that people say that are harmful. I would say that there's a lot of things that people say that are not, that are true and not impactful. Most SEO work is not impactful. And what I've observed is that a lot of consultants and agencies and SEO people will optimize for amount of work done versus for impact.

[01:07:56] Ethan Smith: Like how many slides, how many little tasks how many data analyses can I have, and I just give you as much data and as many tasks as I possibly can. And the client says, wow, this is a lot of work that you did a lot of stuff

[01:08:11] Nima Gardideh: Mm-hmm. 

[01:08:12] Ethan Smith: and I'm optimizing for stuff, amount of stuff. But really what you want is impact.

[01:08:16] Ethan Smith: And so then what happens is you, you work on a bunch of stuff and you get no results. And all of it's true. Like you fixed this alt tag that was empty, which is, which is a true thing. It's, it's better to have an alt tag, but it doesn't drive any impact. And so I would say that it's mostly around prioritization and there the distribution of the stuff that's impactful versus not.

[01:08:38] Ethan Smith: It's probably 95 5. It's not, it's not even 80 20. There's a small set of few things that are really impactful. Those things are making sure that your cross-linking is set up properly, so you're linking across other pages, making sure that Google can even read what's on your page. If your page is empty, it's not gonna work that well.

[01:08:55] Ethan Smith: That's on the technical side. On the editorial side, write articles that are adjacent to where my topical authority is. So write stuff about LinkedIn pixel tracking. Don't write stuff about Instagram cause we don't have any authority there. It's like prioritization of the pieces of content that you have.

[01:09:10] Ethan Smith: Making sure the, that the outline is comprehensive, making sure that our conversion optimization is set up and people aren't visiting and leaving on the programmatic side, category pages and product pages. So making sure that those are set up properly. Google can read them. We've launched the ones that , are impactful.

[01:09:26] Ethan Smith: If we're Thbtack, we've launched all the different plumbers plus city combinations. Those are the and title tags. That's what's impactful. The giant list of a hundred different things of alt tags missing in a, in a one URL that's has spam on it. Like that's not impactful. So it's mostly around prioritization.

[01:09:48] Ethan Smith: ,One common thing, I just had a LinkedIn thread about this is, Page speed. So page speed is a factor, and it's a teeny tiny factor, quote unquote. , but it's some, it's one of the few things that you can measure. You can actually measure page speed and  that it's, IM impactful because Google said it is.

[01:10:09] Ethan Smith: And so people will latch onto page speed. And page speed is useful because a faster page means better engagement, but it's more so around if your page takes 20 seconds versus one second, your, your engagement will be bad. That's why it matters as a ranking signal. It's a teeny tiny factor. And Google is specifically said, this is a teeny tiny fa, it is a factor, and it's teeny tiny.

[01:10:31] Ethan Smith: And, and it's very high scope frequently, so if you want to get from 20 seconds down to five seconds, down to one second, that's a lot of work. And it's, and, and, and I would only prioritize it based on engagement. But again, it's, it's something that is true, but low impact typically. And there's a huge list of other stuff in SEO that's, that is true and not impactful.

[01:10:51] Ethan Smith: The stuff that's impactful are, are some of the things that I listed. Most other things are not impactful.

[01:10:56] Nima Gardideh: The amount of work it takes from 20, 20 seconds to five might be not that crazy, but five to one gets all of a sudden way harder and probably the has very diminishing returns. , and so like having the engagement as the metric or as the way you want to be judging that instead of just these raw numbers of how many milliseconds it takes to load this thing is probably makes a lot more sense.

[01:11:18] Nima Gardideh: , wow. So the, this is, this is cool. Like the, and then the, the reason for the having a lot of sort of back links or not back links like crosslinking is because Google is not actually that good at. Indexing all your pages, so you want to give it more opportunities to keep following through and finding your pages easier.

[01:11:36] Nima Gardideh: Is that sort of the, the thinking there?

[01:11:38] Ethan Smith: It's slightly different. So, and this is one of the most one, one of the least understood areas. , let's say that you have a site with a million pages and you have, you, you have a million articles and you have no crosslinking. Google will probably find all of your pages or most of them, and they'll index them over time.

[01:12:00] Ethan Smith: So it's not that they can't find them, it's that they're not crawling them. So as soon as you add crosslinking across your articles, They, they already found, they already have the articles indexed, but now they're crawling them more. And when they crawl them more, now you get more traffic. So a common thing is a company launches a thousand new pages and it takes six months for all them to get indexed, but they eventually do.

[01:12:23] Ethan Smith: Why did that happen? It's because you didn't have cross syncing. You had your site map and you had, you had your XML site map and your HTML site map, and then you submitted the URL with the a p i, but they're still not finding it. And then when they find it and index it, they're not getting traffic. Why are they not getting traffic?

[01:12:37] Ethan Smith: Because you didn't have enough crosslinking. So crosslinking increases the crawl rate and the crawl rate increases the, the, the rank and the traffic.

[01:12:46] Nima Gardideh: Yeah, that makes a lot of sense that that's funny enough that this exact thing has happened to us this past couple of months. We just launched a series of what you called category pages for. A bunch of analysis that we've been doing for years for ads that, that other companies have produced. They were all in the newsletter, and now we have like these category pages for every single ad and like an analysis on the ad.

[01:13:06] Nima Gardideh: , and I went into the search console before our call and noticed that like a massive majority of them have not even been indexed yet. , and it was just such an interesting thing. So the, the next, literally the next thing I'm gonna do is go and add a lot of cross-linking to all these pages cuz it, it clearly is going to help that process become faster.

[01:13:25] Nima Gardideh: , but Ethan, thank you so much for this. This was, it feels like you're like a wealth of knowledge of seo and I've probably have just touched the surface. We could have spent two hours on, on understanding how, how you do your work at Graphite and, and how you've been thinking about this, this channel as gr as a growth channel for, for companies for the past really all, most of your career.

[01:13:47] Nima Gardideh: So I really appreciate your time here and I'm excited to apply some of these learnings already to our own, own work. And I hope to have you again sometime soon and have you talk more about all the successes that you've been behind.

[01:14:01] Ethan Smith: Absolutely. Thank you for having me.

[01:14:02] Nima Gardideh: All right. And that's a wrap with Ethan. I'm so grateful for having spent the time with him understanding SEO a little bit better. I'm getting to understand his process. It was quite interesting to think about. pinpointing or putting against SEO budgets, with other programs like paid is an interesting problem that you have to deal with as a, as a cm o or founder, in terms of how to allocate your resources and budgets and efforts towards these different programs that have very different return timelines.

[01:14:36] Nima Gardideh: an interesting problem to think about, if you're trying to make some of these decisions. And hopefully that was a bit helpful, for you to think through , as always, if you're enjoying these shows or have feedback, please email me@nemaat.com or subscribe and follow us on the different channels that you're listening to.

[01:14:57] Nima Gardideh: Stu I really appreciate it. The next person on the podcast is gonna, very different. We got a growth engineer coming on. gonna be a little bit more technical, but I think it's gonna be interesting. And we also are telling a story of StockX with the next one as well. So there's a couple really interesting guests coming on and I'm excited for you to listen to them and hear how people enjoy this.

[01:15:20] Nima Gardideh: So appreciate it. Have a good one.