You may have noticed recently that there's a trend of folks who leave search engines crafting their own startups. It made me wonder - how are these individuals doing with the SEO on their own sites? Have they engaged a secret formula that the rest of us don't know to achieve incredible results? Or are they languishing in poor accessibility and SEO tactics gone awry?
Since many of these individuals are friends and colleagues of mine, I won't pass judgment (and I certainly won't be comprehensive in my SEO reviews - got lots on my plate), but I do want to share them with you and get your opinions & feedback.
Eytan Seidman & Oyster.com
Eytan worked for Microsoft's Live Search and was often the public face of the search quality team at conferences. You can read an interview with him here re: Live's efforts in the search arena. Clearly, Eytan's a guy who knows search and should get SEO. His new startup, a project with his brother Ellie, focuses on creating the web's most high quality, in-depth hotel reviews.
The Good:
- The reviews themselves are some of the highest quality, most in-depth content I've seen in the travel world, and possibly on the web as a whole. Gorgeous photos with great descriptive text that really takes in the detail - I'm not sure I'd ever even need that much knowledge about a hotel, but I can imagine it's a picky honeymooner's best friend.
- Internal link structure looks solid, and it appears Google's already grabbed tens of thousands of pages - site:oyster.com
- Titles are great - hotel name + geography | Oyster Hotel Reviews. It gets all the relevant keywords in nicely without looking spammy.
- Good breadcrumb navigation that helps with any potential keyword cannibalization (and is helpful for users)
- Excellent internal linking in content (whenever another hotel/region is referenced, Oyster links to it in context)
- Ranking #80 for "Hotel Reviews" and rising fast :-)
- Smart, descriptive meta descriptions
- Good link profile, primarily from press and blog coverage
- Not afraid to link out to external sites in review (and no nofollows on editorial outbounds, either)
- Great architecture means that the right pages rank first for the desired query. For example, even though there are lots of pages about "Fisher Island" in Miami, the destination guide ranks first (fisher island site:oyster.com).
The Bad:
- Exact match domains are so powerful in the rankings, I'm a bit surprised they went with "Oyster.com." That said, there's never been a truly big success in technology that had an associative name - Microsoft, Sun, Oracle, Google, Yahoo!, Facebook, Twitter, etc. Maybe YouTube would be the closest to a billion+ exit that actually had a somewhat descriptive name.
- Oddly, the top button links have no anchor text - they're done with CSS, so there's nothing between the <a> tags,
e.g. <a href="/dominican-republic/hotels/" id="headerDominicanRepublic" class="headerButton"></a> - I'm not 100% sure, but it looks like not all of the gorgeous photos can be indexed based on how they're displayed/linked-to. Google's got a couple hundred in their image search index, so perhaps I'm just missing how they're getting efficiently crawled.
Avichal Garg & PrepMe.com
Avichal worked in search quality and web spam at Google, so he's got a good idea of what makes sites succeed in the engine. Although not publicly facing, as the CEO and founder of PrepMe, a site that aims to provide online tutoring to kids in high school for test prep and class assistance, he's going to be much more in the limelight (particularly as PrepMe has getting exciting levels of traction).
The Good:
- They're already on page 1 of the results for high demand queries like "online sat prep," "online act prep" and "psat prep courses"
- They're employing nofollows on unimportant internal links... (not sure if that's technically good or bad, but I'll bet Avichal knows what he's doing)
- They've got a great set of long tail content in the PrepMe Answers section - Google appears to be doing lots of indexing of these (155K results) and many of them rank well, too.
- Titles, URLs, internal link structure, etc. is all very clean and simple (and well-optimized)
The Bad:
- A lot of page have the same title tags - prepme answers - for example
- It looks like their sitemap file has a limited number of pages listed
- Not that it's a huge deal, but both prepme.com and www.prepme.com resolve (though they do use absolute URL links to ensure that deep pages aren't being linked to on the non-www version)
Jeff Weiner & LinkedIn.com
Jeff started at Yahoo!, where he led acquisitions of search technologies from companies like Inktomi, AltaVista & Fast. After a stint as EVP of Yahoo! Networks, he moved to the VC world before taking the lead role at LinkedIn. Jeff's a very smart, talented guy and LinkedIn is clearly benefiting. They were one of the first to adopt Google's rich snippets throughout the site and have fixed a lot of duplicate content/canonicalization issues related to user profiles.
The Good:
- As I mentioned, their use of rich snippets in Google is awesome (note they grey descriptive text in result #7 for "Rand Fishkin"). I think they could do even more there (photos especially), but Google may want to keep that only for their own profiles.
- Strong basics - titles, meta descriptions, page structure and content
- The use of the member's name throughout the page in headlines, e.g. "Rand Fishkin's Summary, Rand Fishkin's Experience, etc." gets plenty of targeted phrase use without being spammy
- They allow direct links out, without nofollows - a very, very wise choice and an incentive to maintain updated links in your profile
- People directory - example "A" - these are terrific navigation systems for engines and they do a really solid job of minimizing all other links on these pages
- Interestingly, like PrepMe, they continue to employ nofollows - I think this is a positive (doesn't waste crawler bandwidth even if it is "evaporating" juice), although I'd probably recommend link consolidation even more.
- Brilliant use of individual name directories, e.g. John Bennett to help searchers find the right person
The Bad:
- They're disallowing a lot of stuff in the robots.txt file that might be earning links (like /search*). As we've noted in the past, blocking pages rather than simply noindexing them, prevents juice from passing.
There are lots more that I won't dive into detail on, but could be interesting to review as well.
Barnaby Dorfman & Foodista.com
Barnaby headed up search at Amazon's A9 (back when they were building their own engine to compete with Google/Yahoo!/etc). He's now the founder and CEO at the Wikipedia of Food - Foodista.
Patrick Li & Raptr.com
Patrick previously worked on Google's datacenter infrastructure, then on the launch team, where he audited new products/features before they went live. Raptr's online games platform is fairly addicting and they've got some impressive rankings.
Bret, Jim, Paul, Sanjeev, Ana, Tudor & Gary at Friendfeed.com
A large number of Googlers at this social, real-time startup. Friendfeed clearly wants to capture people search the same way LinkedIn, Twitter & Facebook have, and they're doing a decent job of it.
Mark Lucovsky & VMWare.com
Former Microsoftee and Googler (where he was an engineering director), Mark's now with VMWare, who still has a way to go before they're in the top 100 for "cloud computing" although that's unlikely to be Mark's sphere of influence.
Vanessa Fox & JaneandRobot.com
Vanessa was the creator of Google's Webmaster Tools platform and her new company, focused on making SEO accessible to developers, has held a number of industry events. Obviously she's kicking butt. :-)
Tim Cadogan & OpenX.com
Tim was a search executive at Yahoo! before moving on to lead OpenX, an ad-serving platform. There may not be a great need for SEO here, but Tim probably knows some folks who can help if they do want to leverage organic search to drag in more targeted leads.
There are many other examples, but I think this is a decent start. If you've got others to share, we'd love to hear about them in the comments.
I always tell people "do as I say and not as I do!" :) As optimizing my own sites is sadly, one of my lower priorities right now. I could write an entire post about the things I'm doing wrong with my sites and just haven't had time to fix (starting with a lack of updated content!!).
I know dude - same here. We had a full audit of problems on SEOmoz recently and there's a long way to go :-)
Let us all know when you are ready to publish that list Rand! How about a competition to see which leading SEO companies make the biggest errors on their own site? Then again, they say a good plumbers tap always drips...because he is focused on doing good stuff for his clients ;)
As a pro member I am a little concerned about this comment :o
Well, actually, it's because of PRO that we don't do everything on the SEO side right. We're really focused on building tools and software for PRO members, so other dev tasks like SEO friendliness stuff on the front-end slides.
My guess is that apart from a handful of specifics that fell within their own direct perview, most search engine employees, even (or especially) senior ones, have about as much idea who to game the algorithm as we do.
There are so many different components to the algorithms now, and individually and as a whole they're so complex, no one can hope to know all of it.
i don't think the algorithm is too complex to make use of it. that would make it unmaintainable.
imho "the algorithm" is somehow parted into different aspects which build abstract scores on their own and then merged, the problem seems to be that the scoring algorithms are constantly changed and what gives you an advantage today will probably be a disadvantage tomorrow. especially if you are not employing best practices,
so maybe this is what we really should learn from this case study: it is a good idea to do what other white hats do. no matter what that is. the engines will acclimate
That's kind of what I mean. PageRank and all the other ranking factors might themselves be well understood by the engineers who work on them (I would hope it is!), but the total system is probably not understood by anyone.
In the same way, a computer's CPU is made of a lot of well understood parts, but the sum total of all the billions on parts on the chip is not understood as a whole at all. It can't be, it's too complex.
Great post Rand, thank you.
If I have decent amount of money and skills, I will go for branding, it's the only way that people can keep your company in their mind for a long time. Here are some examples:
If I go travel, I search expedia rather than something like "xxxtravel.com"
If I go shopping online, I search ebay rather than something like "onlineshop.com"
Sure: of course you search for "ebay" or "expedia", as opposed to "auction house" or "travel website" - but thats only because you have brand awareness in these markets.
 If for example you were looking to buy a used car after just moving to the UK, how would you know that the brand leader was auto trader for example, as you WOULDNT have brand awareness at that point.
that is why natural seo is important, and a strong element of that is a keyword rich domain name.
heck I wish the company I SEO for was called "keyword1keyword2.com" as it would make my life easier! :)
MOG
You are right, seo is definitely very important. I guess my examples were not clear explained, my bad :D
What I was trying to say is, "brand" domain can beat "keyword" domain, so why don't go for brand? Cos I can remember it and I dont really want to click something like "cheapflights.co.uk" even if it's at 1st position. :D
While I agree to a certain extent, branding and SEO go hand in hand. From my past experiences, often times, putting in efforts to brand your website along with great on-site optimization and some of the most basic SEO tactics do much better than websites that I've worked on where I implement all the tricks in my book.
Overall, my philosophy on SEO is that great SEO should be always geared to brand your website.
Very, very interesting article!
I am encouraged that I do a lot of "The Good" things you mentioned.
Thank you.
Interesting stuff Rand - particularly to find out where these search boffins have ended up!
How come you didn't analyse whether any linkbuilding / linkbait efforts have been made? For me that would be the most interesting analysis, to see how search engineers use their knowledge of the link algo to get ahead in search. Also as you recently pointed out from the new ranking factors data, 70% of the algo is links.
Yes the link profiles would make a great post in itself. I've just ran prepme.com through linkscape as i'm after education links. They have 8000+ links, but only from 275 domains! I thought sitewide links were frowned upon, what's that ex-google worker up to!
I like how most of the folks that leave end up trying to bring search and algorithm principles to other companies rather than trying to crack the algorithm of the search companies they leave.
Does this mean that a) there are rigid no-competes involved or b) the SEO industry has done enough to crack the algorithms already?
Interesting.
does anybody know about non-disclosure agreements for search engine employers?
if i was a leading search engine, i would not just make sure to prohibit publishing ot certain info, but i would also forbid usage of it.
Great question and...maybe the hired temps/non-employees known as "Quality Raters" could be included in that inquiry.
But, how could this possibly be enforced/monitored?
Rand,
Appreciate the shoutout for PrepMe!
I've put the things you pointed out on my todo list. Thanks for the free SEO audit ;-)
Avichal
Hi Prepme, sorry for sniffing around your backlinks(my above comment)and posting here!!
No need to be sorry. It's good to know we're getting enough traction for people to care.
 Btw our we haven't done any active link building in the past so our backlinks have been organic and have been aquired over a long period of time. I first registered the domain in 2001 so it's been slowly building link history for a long time.
Great post.
Thanks for bringing some insight on websites that I have found to be very admirable.Â
Great post as always.
I think the point is that no matter where you "worked" the job of optimizing sites is never really done and in many cases the first priority may be to optimize the business plan and baking the oprtimizing into the platform a secondary concern. In other cases, their personal sites don't get the love that their clients sites do as a reality of the time constraints that we are all under.
I am my own worst critic which makes it hard to spend any really time on my site because I am never satisfied.Â
Thanks for bringing some insight on websites that I have found to be very admirable. These are websites that I frequently use or have come upon several times in the past and its good to know and hear where they're from!
It's really strange to see, that they really doen't know much more than we do. Dissapointing.
Surely it's the opposite of disappointing? If these guys don't know much more than we do, that means that the search engine companies are actually sharing a lot more than previously thought.
I really appreciate the explanation of the The Good and The Bad. Useful and understandable. Thanks.
Deb Ward
intersting post. it would have been really funny to me if you had found something from them that seos havent seen before
Great post, Rand. :) I think that anyone who creates a site and was formerly employed by a large search engine co. has to have some sort of an SEO advantage over their competition. Perhaps, they've seen complex data that others have not been exposed to and know how to soak in the glory of getting first page serps for very highly competitive terms. Speculation, of course. I think the possibility exists that many insiders might consider leaving a good paying job for one that pays an excellent sum, if one has an understanding of how things (SEO) work in their companies' search engine.Â
Based on the good outweighing the bad, in the short reviews you performed, I'm leaning towards, they've got the SEO nailed down for the most part.
(for the "linksters" above in this thread: a thumbs down was sent your way)Â
I think this is what happens to a person who have the work to analyze something... you see what is missing and try to create something to complete this "hole".
People that know their own market area inside out make great SEOs. Industry experts with years of experience will know all kinds of info which is priceless for SEO, like technical terms or industry jargon are an excellent starting point for keywords. They will also know buying patterns and the seasonality of the industry which helps you to upsell and convert users better.Not seen many of these types of people working in SEO though... If you are an in house SEO, it's always good to find these knowledgeable people and absorb as much info from them as possible!That hotel review site looks incredible... I'm off to take a look!
Hotel Reviews is actually ranking 15 now.
A great post on many levels.
I love picking things up by reading case studies. In this case written by an SEO guru on sites belonging to search industry veterans. Very sweet.
I'm aware of breadcrumb navigation enhancing site usability as in the case of oyster.com, but I've missed something important as to how it can help with preventing possible keyword cannibalization.
A great list of sites to see peek over Rand's shoulder for the initial SEO review - and a couple of new names to follow on twiter too.
That wasn't supposed to rhyme.
I liked this post. Would like to see more posts like this (SEO case studies)as they can be very educational, even for the top SEOs here. Also great having community feedback in the mix.
Great article and very informative. Creating a domain with at least one keyword in and combined with good SEO will help a site rank high in Google.
You're right.. Exact match domains are extrememly powerful for SEO.. so the only thing that I can think of for Oyster.com is branding...
it's alot harder to work on branding when your domain is somehting like Best-Hotel-Reviews.com in my opinion..
 ... maybe im wrong, & they DO have a secret SEO formula that no one knows about.. :) ... but highly doubful..Â