View more presentations from Rob Snell.

Rob: Hope Brett doesn’t mind, I’m going to…Yeah, this is the new stuff I want to make sure I get.

CONVERTING KEYWORDS -- When I’m looking at keywords, converting keywords, these are the different things that I look at, the different metrics in my visits. I look at the number of visits—traffic, I look at the number of orders, I look at the total revenue generated, and then in Yahoo I actually have to make this into a spreadsheet because it doesn’t show me this, but in Google it will show you—revenue per visitor is the quality metric. That tells you how good a keyword is. Not how popular it is, but what the quality is.

If I get one more visitor to my website, on average, what are they going to spend? How much money is that going to be in my pocket?

So in here you see the red; I’ve got a low revenue per visitor. And these aren’t the real keywords; I just kinda found and replaced. But I used real keyword data. So that tells me there’s something wrong with that specific keyword on my site. So that gives me an idea of something to look for.

I also track individual conversion rates by keywords; conversions by people who actually add to cart and by people who place orders. And when I took this keyword sample it was $5 million. I took out all my domain terms, all my brand terms, because I’m going to rank for those pretty much anyway. Those are kind of a gimme.

If you’re taking credit for that…Or if you have an agency doing your SEO and they take credit for traffic coming to your site using your name, you know, shame on them.

$5 million worth of keywords. 21,000 plus keywords in this collection. I can’t remember how long the time period was, but it was a pretty big time period.

I like to prioritize my SEO keywords based upon how much revenue they generate. This graph right here shows you traffic keywords. It’s my top 25 keywords when I sort them by how much traffic’s coming to my website. The height of the bubble on the graph over here is how much traffic. The size of the circle is how much money. So a big circle there is big money; little circle, little money.

If you’ll notice on here, my number one and my number two keywords are pretty much the same. You know, high traffic, high revenue. My number three keyword, however, I got a lot of visitors but very, very low revenue.

If you’re prioritizing your keywords by traffic, you’d be prioritizing the wrong thing. I take the same 20,000 keywords and then I sort them by number of orders, which is conversions or transactions. Same thing. My top three big keywords, they’re showing up. They are high numbers of transactions, high numbers of revenue, high numbers of traffic—those three big bubbles up there.

But if you look, the fifth one has got 650 orders, which is a lot or orders, but it’s only $10,000 in sales. That’s probably like a $3 whistle or a $5 cheap leather dog collar. You need to be concentrating on total revenue, and that’s what this graph is.

And this shows you the same graph; it’s number of orders and the amount of money that’s generated. You can see the circles that are higher up on the graph are probably smaller average size orders. But that’s a lot of $20,000 circles right there.

So when you prioritize by revenue, you’re going to be spending what limited time you have as a retailer on the most important things.

Question: Is your margin percentage roughly the same…

Rob: It really is. And when I say revenue, I really mean profits. Because for me, we’re pretty much, across the board, we have the same margin. A lot of folks have stuff they’ll make 90% on one thing and they’ll make 10% on something else. So that’s a really good question.

I graphed these keywords, and this is like the first 25. And you can see there’s this drop-off. These are non-brand keywords by revenue, probably from a different sample of time. And I said, “Hey, that kinda looks like a dinosaur!” And I like to have a metaphor.

They talk about the “long-tail”. I’m like, “OK, well long-tail. There’s something with a long tail.” But it’s not just your top keywords phrases in the long-tail. I talked about money right? I’ve got $5 million. Let’s think about this in terms of buckets of $1 million. Because in Mississippi $1 million goes a long way. You can get a nice double-wide, you know?


In the first bucket there are only 20 keywords that generated a million dollars in revenue. I obsess over these keywords. There’s one of them, I’m getting my ass kicked right now because they’re cheating. Because Google’s not doing what it should be doing. But instead of losing too much sleep over this, I go, “It’s only one keyword. It’s in one bucket. Think about the other buckets. Concentrate. Do what you gotta do. Don’t cheat. Google will sort it out in another 18 months.”

20 keywords generated 6,500 orders, but this bucket is the same as the other buckets. My girlfriend was like, “Well don’t you need to concentrate on the keywords in the head?” I’m like, “No, I want all the millions,” you know? Those are the head terms, those 20 terms.

Well the next $1 million came from 108 different keywords. And it’s a lot easier to optimize for those and rank for those than it is for the head terms. And that’s 10,000 orders we got from that. And those are the neck keywords terms.

Well the next bucket, the next million, came from 800 different keywords. And a lots of these are the same keywords as before but they’ve got modifiers on them. So it’s just kinda like a different flavor of those keywords. It’s a lot easier to rank for those. That’s 9,000 orders I got, same million dollar bucket; that came from the back.

Well look at this. It’s even better. 2,500 keywords generating $1 million. I’m not losing sleep over any of these. I optimized my sites. It’s baked in. It’s built into the templates. It’s automated. I’m not freaking out about this kind of stuff. Same million dollars. I call that the rump. Don’t neglect the rump. There’s $1 million in the rump!

All right, long-tail. And for me, long-tail is the last bucket. A million dollar bucket, 16,500 keywords. 16,500 orders. One unique keywords, one order. This is why you take all this cool stuff that I’m about to show you and you sprinkle it over your site. Bake the stuff in your templates, you take the keywords, you put modifiers all over it. That’s the long-tail. Another million dollars. There’s lots of meat in that dinosaur.

And so your homework is, what are your top 100 keywords by revenue? I really want to know what all your keywords are by revenue. If you only got one conversion, it’s a fluke; don’t worry about it. If you got two or more conversions for a keyword phrase, you need to be considering it. And it starts at the rump.

So now you’ve got 20,000 keywords. What the hell do you do with it? You want to organize buckets of keywords. So what I do is I take my converting keywords and I chunk them into and make a tag cloud. This gives me ideas of keyword phrases that appear. The larger the keyword is, the more frequency it appears in my keywords list. So you can get an idea of what I’m selling online just based on the number of different times it appears in all my different keywords.

I organize keywords by both manufacturer and generic terms. Those are two different ways to organize buckets of keywords. I don’t work it one keyword at a time; I work it tons of keywords.

Here’s an example of a generic term: dog boots. I take off the root phrase, because if I didn’t it would be 17 feet tall and you couldn’t see anything else. This is what’s left after I make a tag cloud. You can see that it’s different types of dog boots. You can see it’s where people use dog boots—in the snow. There are buy words in there like coupon, reviews. You can see people talking problems, like pad protection, dog, foot. You can see manufacturer brands in there pretty big—Lewis and a couple others.

I want you to add buy modifiers to the texts on your pages. And here is why. This will pay for your trip to PubCon. I have clients who are telling me to shut the hell up and not tell other people about this. But I love my peeps.

We generated $360,000 last year that I can attribute solely to the modifiers in my modifier list. And I’m not going to give you the keys to the kingdom. I have 250 different buying modifiers that apply to any product on my website. Some of the better ones are “buy”, “cheap”, “best”, “reviews”. Buy me a Honeybun I might share a view more.

Here’s an example of rankings right now. “Discount” is the modifier. “Hunting dog vest” is the keywords phrase. Number one and two. And this is old. It’s three months later. Who knows what’s happened? “Buy hunting dog vests”—one and two. “Hunting dog vests online”. All of these are converting keywords phrases. We’re number three and four. One of my competitors has the word “online” in his name. There’s no way I’m going to beat him for that. I’m not even worried about that. Lots of meat in that dinosaur.

So your homework is collect buying modifiers from your analytics. And some modifiers apply to specific buckets of products, not everything. You want to apply these buying modifiers where it makes sense on your site.

One thing I talked about in Vegas was you can actually put the word “buy” and “cheap” in the Alt text of your thumbnail images. That’s a really good one. That will pay for PubCon for somebody. That way your pages will rank for the words “buy, your keyword here”. “Cheap, your keyword here”.

Well now what? Pages. This is a big change that we had when the Mayday update came out last year. I had a friend of mine who’s got…he does 10 times the volume that we do, and he’s a really good SEO, and he’s twenty-something. I mean I’m going to be 43 years old. I’m gettin’ old. He couldn’t figure out why his traffic was off, because his rankings for his top 5,000 keywords phrases were untouched.

I said, “Well, are you prioritizing your SEO pages based on revenue?” That was kind of a new idea for him. Basically, I’ve got 20,000 pages on Gun Dog Supply. 5,000 of those are PPC landing pages. 5,000 of those are what we call “mo’ pics”—additional photography, pages that have to have their own page because it’s a Yahoo store. So there’s no real content in that.

So I’ve got about 10,000 good pages. If you go to Google, you can see, with a, what you think Google knows about the number of pages on your site. And I don’t trust this number because it changes a lot. But it’s a good metric to see, “OK, well what does Google say it knows about my website, the number of pages on my site?”

So I’ve got 20,000 pages. Google knows about 18,000 of them. Well in the past year, 6,500 of those pages have actually gotten me a visitor from Google; looking at analytics, seeing how many pages actually have an entry from Google organic.

But like I said, I don’t care about traffic. We’ve got a shorter buying cycle. If you’ve got a longer buying cycle, you want to look at time on site as a good analog to revenue. It’s really hard sometimes, if it takes 20 visits to your website for somebody to buy, for you to tie the keywords phrases that originally got them there to where you got money. But in our case, it’s usually a day or two. With most retailers, it’s a day or two.

Ovanash [sp] told me, it’s like, “If you got 80% of your conversions in less three days or less than three visits, then don’t worry about coming up with an attribution model.”

2,000 of the pages on my website, 10% of the pages, are actually driving conversions, revenue, with our store. I’m going to worry about those 2,000, not those 20,000. This is what kicked me in the head. My top 100 entry pages from Google, 75% of my revenue came from those top 100 pages.

I asked my brother where he wanted to start when we were writing content. He’s like, “We’ll start with the A’s!”


And I love my brother, but, you know, sometimes he’s not the brightest guy on the planet. I was like, “Why don’t we start with the $100 bills?”


Or the million dollar bills. So your homework is figure out what your top 100 Google SEO entry pages are. This will inform what pages you link to on your website, what content you build, what gets a homepage link, what rankings you need to check.

I’ve got 10 minutes. This is the good stuff, so I’m going to kinda try to go a little slower through this. Every single keywords has a most relevant page on your website. Take “leather dog collars”. For me, that would be my subcategory page. There is one page on your site that is the most relevant page for that keyword phrase. I know that sounds kinda basic, but it’s a pretty important concept. And if you don’t know which page is the most relevant page, you can do that on Google and it will show you, in descending order, what Google thinks are the most relevant pages.

But I’m kinda greedy. I don’t want one page. I want two pages. Clustered results. You know how sometimes you’ll see two or three results from the same website on Google? I told my momma back in 1997 that it would never get better than that. We were the first 30 listings on InfoSeek for all our keyword phrases.


I was a little greedy. Nowhere to go but down. So there’s clustered results. So your homework is to find the two most relevant pages for those top 100 keyword phrases that you got just a minute ago. I like to do that by hand. I see a lot when I’m doing that research. It’s only 100 queries.

All right, site structure. So you’ve got 20,000 pages all laying out in the floor. How are you going to organize them? Well, how you organize them affects your search engine optimization. I want one URL per page. I don’t want the WWW pointing to it, and without the WWW, I don’t want all these attributes, I don’t want all these ad tracking clicks. Using 301’s, and canonicals and whatever your propeller-head smart guy’s gotta do, I want one URL to get indexed per page.

On my homepage I like to link out to my 40 best pages using my 40 best keywords. Look at your homepage. Here’s your homework. Look at the anchors and the URLs that are getting linked to from the homepage. I’m making money for somebody right now. This is the most important thing you can do. View source and see are you linking to the top…are the pages that you are linking to in your top 100. And if now, why the hell are you linking to it? Are you using the right keywords, the best keywords, and linking to the most relevant pages? And are you linking to the same page with multiple links with different anchors to the same page? That doesn’t count. First link counts. One link.

Site pyramid. You start at the homepage. Any page that you link to from the homepage is a second level page according to Google. You may only have 10 master categories and you may think it’s categories like homepage, 10 master categories, subcategory, product pages. Well, it’s not. If you’re linking to a product page, it is a second level on the website. It gets the same priority that one of those categories gets.

One thing I get a lot of folks to do is you put your top 100 products on a page called “Best Sellers” and link to it from the run of site navigation and from the homepage. That immediately gets the spider to your product pages, which your product page links are one click of the homepage. It’s great.

I do revenue-based navigation. I’m all about the money. I don’t have A-Z on my run of site navigation. I have my best selling brands. I have my best selling products. I have my best selling categories.

De-templify your story. I think there’s a diminishing return on run of site links. I’ve got 20,000 pages. If I’ve got 20,000 links that say “hunting dog supplies”, I think at a certain point Google goes, “OK, Rob, I got it. That page is about hunting dog supplies.”


So what I try to do, I have three separate anchors for every single phrase. I’ve got the name, I’ve got the short name, and I’ve got the Alt name. And I mix it up based on my templates. I’ve got a different template on my homepage, a different template on my category pages. I don’t even have top-level navigation on my product pages, because once you get down there I want you to buy. I don’t want you to think about what you’re going to get. I want you to give me some money.

And then on my reviews pages I have a separate navigation as well. I want you to cross-link your product pages. This is another good one. I had a guy go, “Why don’t I rank for this magic trick?” I’m like, “Well you only have one link. It’s from your subcategory page called Magic Tricks. And you have 26 products in this subcategory.”

In the old days I would like on every single product page to the previous product and the next product at that level. So you could just kinda skip across. You wouldn’t have to pogo stick back and forth to a category page. You could actually get to the product, and like a fashion show, you could go through all the different dog collars.

Well nowadays, I’ve learned that it’s even better for us to actually have links to all your brother and sister pages on every single product page. So all those 26 products, I have 25 links coming from those other pages in the same subcategory. It’s 25 more links and it really helps.

All right, links. How much time?

Woman: Five minutes. Talk faster!

Rob: I will! I’m from Mississippi. I’m sorry I talk so slow.


Rob: If you control a site that has 10,000 pages, if every page has 100 links on it, you control a million links. Is that right? I went to Mississippi State, so I’m not sure. Is the math right? 100x10,000, isn’t that right?

But anyway, you control a lot of links. You need to make sure that you’re using the right keywords in the anchor text and that you’re linking to the right URLs. Redundant links are bad; I said that. Bad link text like “Click here” is bad. Use your run of site anchor text lightly. Those three different navigations, we talked about that.

Embed links in caption fields. This is the best thing ever. I thought this was dead thanks to Bing. I hated Bing taking over Yahoo because I was losing all my cool tools that Yahoo used to let me do with these advanced site operators. This is the big takeaway. Get your pencil out. This is going to make somebody some money.

I like to link inside my product body text. We call it Caption on the Yahoo store. I like to link the two whatever I’m talking about. If you talk about a keyword phrase on a page, link to it from the text like Wikipedia does.

Well, this is a killer trick. We though it went away, but a buddy of mine found it. If you go to, they don’t use Bing in the UK. Thumbs up. Do a search for That shows you all the pages that are in that index, right? Add your keyword phrase. That then shows you all the pages on your site that have that keyword phrase.

So what? Google can do that. Google can’t do this. And link: and the URL that you want to rank for that keyword phrase. That will show you all the pages on your site that have that keyword phrase that link to the page that you want to get links for.

Well that’s great, but those pages already have links. Who cares about that? I’m giving away one of my best secrets here. Put a minus in front of link. This will tell you all the pages on your site that have the keyword phrases you want to rank for that do not link to the page you want to rank. That’s easy links. Take your top 100 converting keyword phrases, run it through that. You’ll find dozens of pages on your site that you can link to those pages with and increase your rankings.

All right, text. You need text. I’m going to blast through that. This is all review, blah, blah, blah. Make a lot of content. I want you to write one piece of compelling content. For us, the first one was buyer’s guides. Steve wrote a comprehensive buyer’s guide for one of the categories that we sell, and it’s like everybody in the industry is copying this. I know it works because everybody is copying this.

So now we do individual product reviews. If you can do one thing, write compelling content. Put text on your homepage. I like 500 words. Nobody likes texts on category pages anymore. Put three sentences on category pages. If you need to beef it up with some more text, stick it on the footer.

Hand-tweak your top 100 pages. Sorry I’m blasting through these. These will be available on And if you email me,, I’ll link to it.

We automate the other 19,000 pages.

These are keyword modifiers, blah, blah, blah. That’s too much money. I can’t tell you that. That’s it!



Rob: And next up is Krista to talk about…what are you talking about?

Krista: Social media!

Rob: Social media!

Krista: Do you want to answer a question while I’m getting ready?

Rob: Yeah, yeah. Anyone got a question? Yes, ma’am?

Question: So when you’re dealing with clients who have very large brand parameters, you’re not always going to win the battle for your assumptions you…what you want to do as an SEO person…