Eric Enge and Stephan Spencer Talk about SEO
Podcast Date: April 16, 2008
The following is a written transcript of the April 16, 2008 podcast between Stephan Spencer and Eric Enge:
Eric Enge: Hi, my name is Eric Enge, I am the founder and CEO of Stone Temple Consulting, and you can find our website at www.stonetemple.com. I am here today with Stephan Spencer, the founder and president of Netconcepts, you can see their site at www.netconcepts.com. How are we doing today, Stephan?
Stephan Spencer: I am doing great, thanks.
Eric Enge: Excellent. So, one of the things that I find interesting when I go to your blog is that the title talks about a scientist become a web marketing virtuoso. Why do you call yourself a scientist?
Stephan Spencer: Itís interesting, itís actually one of my designers who designed my blog came up with that tagline. The reason why she came up with that was because I have masterís degree in Biochemistry. I was actually on my way to achieving a PhD in Biochemistry when I quit school at the University Wisconsin-Madison to start Netconcepts. As a result, I have a very scientific approach to SEO.
Eric Enge: Right. Now, I can tell you that my experience is that, itís one thing to have gotten a degree in something, itís another to actually think that way, they are slightly different things. I have seen the scientific approach in what you write, and the things that you have done with Netconcepts.
The reason why it intrigued me is that, my father taught physics in MIT for 30 years, and I learned my way of thinking from him. So, I always described myself as being a scientist. People ask me what I am and thatís the answer; what they like to do, well thatís a completely different thing. Sort of a mindset in how you approach things. How do you think it affects your approach to doing SEO?
Stephan Spencer: Well, it really permeates everything that I do, and in fact the way that Netconcepts operates. We take a very experimental approach to SEO and apply the scientific method where you come up with hypotheses and you test your assumptions. Just like with email marketing, you test everything, you test the subject line, you test the from and the to and everything.
So, why canít you apply that same sort of methodology to your SEO initiatives, testing all your assumptions in regards to the right keywords, to the right page titles, to the right HTML coding, use of no follows, and sculpting PageRank; and you name it, everything can be tested.
Eric Enge: So, there is lot of opportunity to test and do different things with your SEO campaign. How do you think that differs from what you have seen other people do out there? And, I donít mean necessarily be critical of other people, but just you know what are the things that are a bit different?
Stephan Spencer: Alright. So, some folks are just kind of on their high horse saying this is how SEO should be done, and thatís their line and they are sticking to it. Now, if thatís unsubstantiated, there is a problem there because you canít just take somebodyís opinions on SEO whether there are a prominent SEO blogger, or in the forums, or communists, or magazine or whatever; you canít just take their assumptions that they sell you and say okay this is the way that this particular factor works with regards to Google or another engine.
You really need to knock people off their pedestal to some degree and just say well, show me the evidence or at least convince me that youíve tested these hypotheses that you are making.
Eric Enge: Right. Well, one of the things I picked out of what you said there is this notion that for some people one size fits all and they want to use the same formula on every website. Certainly, there are certain aspects of tactical SEO where that makes sense, but certainly when you turn around to the link-building side of things, one size clearly does not fit all.
There are a lot of different ways to slice that pie and you can have people doing Digg campaigns and hey that works great for them, but you will have somebody else who doesnít want to do that. And, they might be the ones who are out there pushing their stuff out through a widget because they have something that they can syndicate in that fashion. So, you have to bring a certain mindset to uncovering the uniqueness of each business, their website, their markets, and then build your plan to fit in that, does that make sense?
Stephan Spencer: Absolutely. Different clients have different levels of tolerance for Digg baiting, link baiting, and various types of link building tactics. One client may just be totally not interested in social media campaigns and another who you might not think is well-suited to doing social media marketing is all into it.
So, I know a great example, not one of our clients at the moment, a company called LifeInsure.com, and they are in the top ten in Google for life insurance. These guys have done some really fantastic social media marketing, they have put out some great Digg bait campaigns. One of them was called nineteen things you didnít know about death, it was pretty creepy actually.
Eric Enge: Yes, indeed.
Stephan Spencer: The first item on list was that you are still conscious for 15 seconds to 20 seconds after being decapitated, and that was a real treat reading that. But, itís actually something that has been really valuable for them, they were willing to take that risk of putting something quite controversial and appealing to the alpha geeks on Digg.com. They didnít link to it from anywhere on their sites, so their loyal customers and visitors would not be stumbling across this particular article when they browsed around.
But, it made it to the front page of Digg as you would have expected, and got them quite a lot of links. So, it really depends on that tolerance level for being really edgy and also what content you can leverage and come up with. So, some clients have more interesting content than others.
Eric Enge: Right, absolutely. And, all this thinking fits into how you put together and plan an SEO engagement. And, youíve blogged about this recently in your blog, now tell me what your thoughts are in planning an SEO engagement.
Stephan Spencer: Well, first of all not every client or not every prospect that comes to us would make a good client. The first thing we need to figure out is are they innovative in their thinking, and are they willing to invest in SEO? And actually, an article that you will see on Search Engine Land coming out tomorrow that I wrote is on SEO is not free, you canít just think of natural search traffic as free traffic because you have to invest in your SEO and it is a continual investment, you canít just do it as a one-off project.
So, first thing we need to figure out is, are these prospects right for Netconcepts. And, can we actually do a really good job for them? So, there are cases where it is just not the right fit, and we need to refer them on to others, and I have actually blogged that recently too, that I have put out there a few other SEO consultants that I recommend to folks who are interested in doing SEO but arenít the right fit for Netconcepts.
So, the first thing is making sure that the fit is right and that this is going to be successful engagement. And then, we need to map it out, and we very much take a project management approach to SEO, everything needs to be managed. So, if we can show the roadmap and get the client onboard with that, itís going to be much more successful than if they are not fully committed or donít fully believe in the journey we are taking them on.
There are different stages to this journey, so on that blog post that you alluded to, we talk about getting indexed and having crawler-friendly sites, looking at things like site structure, and best practices, indexation levels, navigation, templates that sort of stuff. Step 2 is getting found for the right keywords, and that involves the keywords research, and the content optimization, and programmatic optimization.
Then step 3 would be increasing visibility through natural link building and that link building really is an art as much as it is the science, and getting the right sorts of links and high value links, itís hard to do, but definitely doable. So, the spectrum kind of goes from the real easy, not very valuable stuff like directory listings to really highly sort after, high value links, very high trust, high authority, aged sites linking to from high PR pages.
Itís a kind of a process that weíve come up with, and then there are other pieces that kind of fit into that latter step 2 that as much online marketing as it is SEO, and that could be like developing widgets or making existing widgets more search engine optimal so that they are passing page ranks when bloggers and so forth put those widgets on their site to blogging.
And, being really effective and integrating yourself into the bloggers here, and not just putting out some corporate shell of a blog and expecting people, bloggers, and journalistsÖ
Eric Enge: To embrace such a thing, right.
Stephan Spencer: Yes.
Eric Enge: So, itís sort of leads into, excuse me, whenever I talk about link-building, I always, in the same breath almost, talk about content; because the questions that people need to answer with their website is why would someone link to it. And, my favorite way of illustrating that point to people is, they donít link to you to help you make money.
Stephan Spencer: Yes, thatís right.
Eric Enge: Right. So, they link to you because you are doing something thatís exceptional in someway and that gets back to the content or you know there can be outbound programs such as well-structured widget, which is a content syndication thing; but again, you are still using content. Thatís a huge part of the picture I think as well.
Stephan Spencer: It doesnít actually necessarily have to be content, it could be functionality, just has to be something thatís of valueÖ
Eric Enge: Yeah, it could be the tool, right.
Stephan Spencer: Exactly. And, one of the most successful link baits weíve ever done in Netconcepts is to create a WordPress plugin for SEO and that is called SEO title tag. That plugin page gets more links and more traffic than our own homepage does, which is pretty unusual.
That was the a successful link bait campaign, and it really, it was a little bit of a hard sell within Netconcepts with our management to invest the time into really building out this plugin.
Because, the target market for the plugin is not our companyís target market; our target market in Netconcepts is really focused around larger brands, and online retailers, media properties; and those arenít the folks who are going to be using the WordPress Plugin. It is the small bloggers, just individuals out there that would be using it, but the value is of course in the links. And so, they picked up on that and got the religion.
Eric Enge: Right.
Stephan Spencer: So yeah, itís been a fabulous link bait campaign for us. Now, there are great link bait things you could do that have absolutely nothing to do with your business, and itís just still great content. So there is, I forget who it was, but somebody put out an article and got it submitted to Digg, and it was really funny or interesting urinals. And all they did was they combed through a bunch of a Flickr photos looking for really bizarre urinal photos.
They found some really interesting one, they compiled that into an article with a bunch of photos from these Flickr pages and submitted that to Digg and got a ton of traffic from that and linksÖ
Eric Enge: More than 10,000 diggers, as I recall.
Stephan Spencer: Yes. But, unless you are selling urinals, probably have nothing to do with your business, but the links are still valuable.
Eric Enge: You have to be careful to manage that, with respect to what your site is about and what you are trying to do with it, and how it conveys the reputation of your business out there.
Stephan Spencer: Yes. You donít want to destroy your brand in the process, and thatís where doing it on a micro site or on a blog can be less risky.
Eric Enge: Right. So, you mentioned that you focus on big brands and large retail sites; what are some of the challenges that large retail sites face from an SEO perspective?
Stephan Spencer: They are an interesting sort of animal, in that most of these large retail sites donít actually know how many pages they even have, for one thing.
They believe they have x thousand pages, because they have x thousand numbers of skews and therefore product pages and then extrapolate to add informational pages and so forth. But the thing is, this is all database driven and they donít really have an accurate count. Usually what they use is the estimated number of pages that Google reports as indexed, which is really inaccurate.
It is just a very rough guestimate, that estimated number of pages; and so itís not really an effective number to be basing various metrics on. So, there are other things they donít really have a good handle on, and when you are talking about the kind of scale data that an online retailer could have in terms of number of pages and number of keywords they are managing in their keyword portfolio, itís really hard to scale across that, right.
So, when you have hundreds and thousands of keywords, hundreds and thousands of pages; how do you optimize them all? Itís really tough, you canít go page-by-page, and so there are different tactics that weíve come up with and which I can talk to you about if you like. But, that is another challenge is rolling out optimization across lots and lots of pages. And then, the organization constraints are usually overwhelming.
They donít have a lot of IT resource to dedicate to SEO, or not enough; and the IT departments and the marketing departments at these various companies often times at loggerheads, at odds with each other. They donít all see at eye to eye and have conflicting priorities. So, thatís a real challenge, itís hard for marketing, to get stuff pushed through thatís important to them, when yet another IT project is getting in the way from the IT folks going home at 5 oí clock at night.
Eric Enge: Yeah, it can be painful. You know you have made this suggestion, itís a great suggestion. Marketing is onboard and then you can wait months before it gets implemented and you are grinding your teeth the whole time because you know what the impact is going to be, or at lest you have some idea of what the impact is going to be, it can absolutely drive you really nuts.
Stephan Spencer: Yes, alright.
Eric Enge: But, yet another thing that strikes me for these kinds of environments is, a lot of time large retailers, the various pages of their catalog donít have a lot of distinction in the level of content. And, they may even be using nothing but third-party manufacturersí supplied text. So, they may or may not have a lot of text, but itís all duplicated from somewhere else.
Stephan Spencer: Yeah, duplicated content is the big issue for retailers, not only between the multiple retailers using the same manufacturerís supplied content for the product descriptions, but also a lot of pages are pretty much exactly the same or considered to be quite similar to each other by the search engines due to the way that they have built out the site and the technologies they have used.
Letís say they are using Endecaís Guided Navigation, and they have not implemented it in a search engine optimal sort of way. So, you have all these different permutations that look like a lot of the other permutations, and this creates many thousands of pages that have the same content on them, same products, price range might be slightly different, maybe one product is not listed on this page that was on another page. But for all intents and purposes, pretty much exactly the same pages.
Eric Enge: Right. Now, you mentioned earlier in some of the, or you alluded to earlier some of the things that you do to deal with the scaling problems, and you suggested you might be able to go a little deeper, I think one of the things was, you can individually pick title tags and things like that. So, what are some other things that you do to work around those kinds of issues?
Stephan Spencer: Right. In fact I wrote an article for Search Engine Land on this topic of some scalable approaches to optimizing very large websites. Iíll just mention a couple. So one of them we call thin-slicing, and thatís after the term that Malcolm Gladwell in the book Blink, where he talks about just kind of making split second decisions not over thinking something, assuming you have an expert pedigree or opinion on this and this, and this is just not an uninformed amateur making this assumption.
You have to really kind of step back and not over think your SEO, and if you take a thin-slicing approach to letís say your title tags, you donít have to spend a lot of time doing keyword research and rejigging things within the page copy and doing the title tag at the same time.
What if you just took a very cursory thin-slicing view and cranked through thousands upon thousands of title tags very quickly, maybe spending only 10 seconds or 15 seconds on each title tag, where you are working on a synonym in changing a singular to a plural and then moving on to the next one, maybe moving a few words around and the another one, just that sort of thing. And, not tying in keyword research and so forth, just kind of using your best guess.
That can work really well, especially when you are talking about a huge website with hundreds and thousands of pages, if you can crank through thousands of titles in the day instead of hundred, you are going to get through all the important pages, the ones that are higher up in the site tree that have more importance in the eyes of the search engines a lot quicker.
Eric Enge: Right. You are subdividing the task and prioritizing, and allowing things to happen a little more on a demand basis almost.
Stephan Spencer: Right. Now, there is another tactic that you can use and we have actually developed the whole product around it. The tactic is actually using a proxy server to make changes. And then, you optimize the proxy version of the website and see how that performs. So, we had at one time, a client who was insistent that kitchen electrics was a category they were going to stick with even though we were trying really hard to convince them that it was not a good category name, because nobody is going to type that into Google.
And so, this is a category that refers to food processors, and blenders and various other small appliances for the kitchen. Kitchen electrics, thatís the term for the industry, they would be laughed at by their peers if they changed it to something else.
Eric Enge: Right.
Stephan Spencer: They insisted that, no you are going to have to actually prove to us that this is worth seven figures in additional revenue if we changed that, otherwise we are not going to change it. This is like the bane of SEOís existence, right, is when they get a client who makes you prove everything in advance. Well, what if you had a proxy platform or a proxy server based platform where you could change that, change kitchen electrics to kitchen small appliances and see what the impact is? Now, see what the additional trafficÖ
Eric Enge: Like an AB test basically.
Stephan Spencer: In a way, yeah. But, itís not an AB test where you can test both in parallel; you can only test them sequentially. So, you have the baseline and then you conduct the test, you see what the results are in terms of once the page gets indexed and the rankings shift, and then the traffic then starts to flow in, you collect a reasonable sized samples so it is statistically significant and then you can conduct another test or switch back to the baseline.
Imagine being able to do that in a very large scale, conducting all sorts of different kinds of tests through a proxy-based platform. You could make page specific changes, you could make site-wide changes and do this very quickly and easily where something like a category name change could take months to get pushed through the IT departments, to implement on the native website, you do this on the proxy based website.
You can conduct this test and implement the change in minutes. So, thatís the idea behind our product, Gravity Stream, that actually I came up with that based on the frustration I had with this particular retail client back in 2003, that was very frustrating and I was like, if only I could just show them what needs to be done.
Eric Enge: It is a lot easier to talk them into doing a limited scope test and giving them real data basically. So, and it sounds very cool and definitely will look forward to digging into Gravity Stream in an upcoming conversation. What do you thought on SEO reporting?
Stephan Spencer: Right. So, this is an interesting one, most of the SEOs out there really heavily focused on ranking reports. And so then, there are also the indexation reports and the backlink reports and things like that. ButÖ
Eric Enge: Those are just really tools of the SEO trade, right?
Stephan Spencer: Yeah, they are kind of like necessary blocking and tackling sort of things, but it really doesnít give you whole lot of insight.
Of course you are going to do that on your clientís competitors as well and look at their backlinks and indexation levels and so forth, and benchmark against your competitors and look for opportunities. But, that is really basic stuff. If you could go into some KPIs there are metrics that are really unusual, but give real insight into SEO, that would be pretty cool. And so, we actually came up with some KPIs, that actually my colleagues here at Netconcepts, he came up with seven of them.
It was Brian Klais our executive VP of Search, and one of them is the brand to non-brand mix because when we are dealing with so many online retailers, they are so heavily focused on branded keywords that they miss the non-brand opportunity. If you look at the long tail of natural search, most of the retailers are comprised of non-brand keywords.
Eric Enge: Right.
Stephan Spencer: They are so focused on the branded keywords that they lose sight of the long tail and end up losing the non-brand potential. So, if you establish some metrics around this brand to non-brand mix that would give you some insight into what you are leaving on the table. Another KPI is unique pages, and this alludes to something that we talked about a little bit ago and that most retailers donít actually know how many pages they have.
So, once you establish a KPI around unique pages and you donít assume that the estimated number of pages that Google is reporting from a site: search is your actual number of unique pages, base it on database queries to your database or base on unique pages that have been crawled, you are going to have a much more useful metric to use in other calculations.
Another KPI is pages yielding traffic, or what we call page yield. Now that one is, if you imagine that you have say ten thousand pages to your website and you actually you can measure which pages are driving traffic from the search engines. Well, you would be surprised how few pages actually drive traffic from search engines. So, maybe out of a 10,000-page website, you have 1,000 pages over a given month are actually bringing in search visitors.
So, once you start measuring, that you can focus on that 9,000 other pages that are just sitting there on the bench, the free loaders, they are not doing anything for you there, they are partly a virtual sales forces thatís collecting a paycheck but not actually doing anything to help the business. So, measuring that metric will help you to increase your long tail potential and get more pages delivering traffic.
Another one is keywords per page, so you can measure the keyword yield. In other words, how many keywords per page are coming in from the search engines. So, if you have letís say on an average, two keywords per page over a given month, thatís how terribly healthy, whereas if you had letís say five or ten keywords per page then you have a much broader, these pages have much broader appeal.
Eric Enge: Yes, some of those are long tail terms.
Stephan Spencer: Exactly, again this is the start of long tail health metrics.
Eric Enge: Right.
Stephan Spencer: Another related KPI would be visitors per keyword, also if you letís say have a average merchants attracting letís say 1.9 visitors per keyword while we actually did a study and wrote up a research report on the study and found that that was indeed the case, that 1.9 visitors per keyword was an average yield for the average merchant in our study.
Then, there are a couple of others that I will mention real briefly: index to crawl ratio, thatís taking the number of pages indexed versus the number of pages crawled so you can see if you have a lot of pages crawled that arenít being indexed.
Eric Enge: Right, potentially a bad sign.
Stephan Spencer: Yes. Then, whatís each search engine delivering in terms of traffic, because each engine has different audience sizes of course, but also different demographics and if the people would be interested in your products and services and might actually buy from you.
Eric Enge: Absolutely. The other thing that Iíll add which is that when I am on the phone with a prospective client and they are asking about their search engine rankings and how to improve or how many links they are going to get per day, or whatever the questions are, those are useful things to think about perhaps. But, I always try to get them focused on, well how much business are you closing on a daily bases through your website.
How do you measure that in terms of revenue impact or cash value depending on what you are doing? And, at a goal level, which is little beyond the reporting question I asked you, is the goal to double, triple, or quadruple the revenue. The other kinds of metrics they typically ask about are just part of the process of accomplishing that. Anyway, thanks for taking the time to join me today, and Iíll look forward talking to you again soon.
Stephan Spencer:Yes, thanks, I really had fun talking with you.
About the Author
Eric Enge is the Founder and President of Stone Temple Consulting (STC). STC offers Internet marketing optimization services, including SEO, Social Media and PPC optimization, and its web site can be found at: http://www.stonetemple.com.