←

→

Some thoughts on SEO

December 2nd, 2009 | online advertising, product design, startups

“SEO” (==”Search Engine Optimization”) is a term widely used to mean “getting users to your site via organic search traffic.”  I don’t like the term at all.  For one thing, it’s been frequently associated with illicit techniques like link trading and search engine spamming.  It is also associated with consultants who don’t do much beyond very basic stuff your own developers should be able to do.   But the most pernicious aspect to the phrase is that the word “optimization” suggests that SEO is a finishing touch, something you bolt on, instead of central to the design and development of your site. Unfortunately, I think the term is so widespread that we are stuck with it.

SEO is extremely important because normal users – those who don’t live and breath technology – only type a few of their favorite websites directly into the URL bar and for everything else go to search engines, most likely Google*.  In the 90s, people talked a lot about “home pages” and “site flow.” This matters if you are getting most of your traffic from people typing in your URL directly.  For most startups, however, this isn’t the case, at least for the first few years. Instead, the flow you should be thinking about is users going to Google, typing in a keyphrase and landing on one of your internal pages.

The biggest choice you have to make when approaching SEO is whether you want to be a Google optimist or a Google pessimist**. Being an optimist means trusting that the smart people in the core algorithm team in Mountain View are doing their job well – that, in general, good content rises to the top.

The best way to be a Google optimist is to think of search engines as information marketplaces – matchmakers between users “demanding” information and websites “supplying” it. This means thinking hard about what users are looking for today, what they will be looking for in the future, how they express those intentions through keyphrases, where there are gaps in the supply of that information, and how you can create content and an experience to fill those gaps.

All this said, there does remain a technical, “optimization” side to SEO. Internal URL structure, text on your landing pages, and all those other things discussed by SEO consultants do matter.  Luckily, most good SEO practices are also good UI/UX practices.  Personally I like to do all of these things in house by asking our programmers and designers to include search sites like SEOMoz, Search Engine Land, and Matt Cutts in their daily reading list

* I’m just going to drop the illusion here that most people optimize for anything besides Google.  ComScore says Google has ~70% market share but everyone I know gets >90% of their search traffic from Google.  At any rate, in my experience, if you optimize for Google, Bing/Yahoo will give you SEO love about a 1-6 months later.

** Even if you choose to be a pessimist, I strongly recommend you stay far away from so-called black hat techniques, especially schemes like link trading and paid text ads that are meant to trick crawlers.  Among other things, this can get your site banned for life from Google.

Share:
  • Digg
  • del.icio.us
  • Facebook
  • Google Bookmarks
  • Reddit
  • Slashdot
  • Suggest to Techmeme via Twitter
  • Tumblr
  • Twitter

Related posts:

  1. Some thoughts on the “geo stack”
  2. Anatomy of a bad search result
  • white hat seo is the way to go it will help you in the long run. and i tottaly agree with this "I think since other search engines try to emulate Google, optimizing for Google tends to optimize for others." this would definitely make the task of webmasters much easier, just focus on optimizing for one serp and others would come along.
  • Luckily, most good SEO practices are also good UI/UX practices - I wish it always be so. But if it's not the case, I'm more for UI.
    Said by someone working from year 2002 in SEO :)
  • I'm glad you added the note about black hat techniques. There are a lot of companies that still continue the practices of spamming and as the search engines (especially Google) continue to evolve, these techniques will be recognized even faster and penalized more heavily.

    I'm glad to see a lot more SEO firms like my own are taking purely white hat approaches, which are not only more ethical but more reliable for the long term rankings.
  • Call it content, or call it making your posts linkable (link baits?) - the bottom line is the same: search engines are trying to reflect surfer preferences and behavior in an effort to provide a better service for other surfers. If you manage to get genuine targeted traffic by building real good quality links, then the search engines will follow.

    Good post btw - I learned so much from the comments too.
  • clubs_com
    Would anyone know of a good, reputable SEO company to go with? Would you trust outsourcing to India, for example, those seen on DP forum?
  • Great post Chris, my background is in Search so I'm glad to see some rational discussion on the execution and potential benefits of SEO. It's definitely true there are a lot of hucksters in the industry, but there are also very educated, respectable professionals as well (Rand being one of them).

    The goal for these professionals is to present a site in a way that is most compatible with Google's interpretation of their site content. I say interpretation, because Google is neither omniscient or all-powerful, and they're still decades from being able to view and understand a page of content as a human user would. Instead they use a series of interpretations to reduce a page to something that is mathematically understandable. Given this interpretation and the broad set of standards people use to publish content on the web, there's a lot of room for misinterpretation and resulting reduced ranking if you happen to choose the wrong IA, CMS, etc.

    I'd argue that good, white-hat SEO is the process of minimizing these potential mismatches between Google's relevancy interpretation and the content on a site. Until engines are able to view and assess a page 100% the same as a human user would, they'll remain imperfect tools. Adapting to these imperfections is what good SEO focuses on. Implying that this adaptation is somehow unethical, or that Google's algorithms are sacrosanct is something I've never understood.
  • Good points and like the way you express it. I've never understood the objections to white-hat SEO. I think it's mainly because people confuse it with black-hat SEO.
  • GeekMBA360
    I actually think that it's risky for a start-up to put too much emphasis on search engine traffic.

    I would argue a 50/25/25 initial goal for traffic generation for the first year -- i.e. 50% search engine traffic, 50% referral traffic, and 25% direct traffic. It might be hard to achieve, but I think it's a worthy goal. :-)

    With the increasing numbers of "spammy", tail content aggregator sites such as Demand Media, search engines are getting more and more noisy. A more diversified traffic generation strategy will help protect the downside, and reduce reliance on Google.

    I guess you can call me the Google Pessimist. :-)
  • randfish
    Chris - I think the biggest thing you've forgotten to mention is that 70%+ of the weighting/ranking used by all of the engines depends on links. If you're not thinking about how your content and pages will incent users/bloggers/writers/media/other sites to link to your work, you'll lose out to someone who does.

    A while back I got riled up about the lack of SEO in startup marketing and wrote about it - http://j.mp/4q9zkh - might be relevant/useful, though I did write with a bit more anger than was likely deserved.
  • "I think the biggest thing you've forgotten to mention is that 70%+ of the weighting/ranking used by all of the engines depends on links."

    That is absolute BUNK, Rand. Links are more important to the SEO community, not to the search engines.

    Unless you can show where the search engineers establish the 70% weighting for links in their algorithms, you're once again confusing the importance you place on links with what is actually happening in very complex algorithms.

    Several Googlers -- including Sergey Brin -- are on record as saying the link part of their algorithm is about 1%.
  • CF
    I think there's different issues at play here. I can agree with you that links are not the most important factor for being a result in Google, that would be having content relevant to the search query.

    The problem with relevant content is that lots of people have it. Sure, getting 100,000 links to a blank page won't make it rank for the search "credit cards", but neither will writing the best analysis of credit cards ever written. You need relevant content to be one of the millions of results returned for most queries, you need links to be at the top. Links convey authority, they convey trust. If 100,000 people are linking to an article on credit cards, then that is a good article about credit cards in a search engines mind.

    In that respect, I don't think its unfair at all to say links are 70% of "ranking" because the things that are ranked well, have the links to prove it.
  • It is fair to say that links are 70% of what many people in the industry do because they don't know what else can be done.

    It is NOT fair to say that links account for 70% of how search rankings are determined.

    Nor is it fair to say that links account for 70% of the work involved in search engine optimization.
  • "Several Googlers -- including Sergey Brin -- are on record as saying the link part of their algorithm is about 1%."

    @Michael - Two points here. It's rare for Google to ever disclose parts of it's ranking algorithm simply to stop potential spamming.

    Second, the algorithm constantly changes. The internet is constantly changing and Sergey Brin (although founded Google) is unlikely to be kept up to date with every single algorithmic change.

    NB: SEOmoz ran a big survey on possible search algorithmic factors and came up with this: http://www.seomoz.org/blog/ranking-factors-vers... - anchor text, 'trust and domain authority' and link popularity all imply (some) root in link building.

    Is it possible to send a link with Sergey's comments? How long ago was it?
  • I am familiar with SEOmoz's opinion poll on ranking factors. It carries absolutely no weight or authority with respect to what search engines do and do not value.

    SEOs mistake their own opinions for what search engines are actually doing quite often -- but the outrageous claims that Rand makes about the importance of links to ranking algorithms are completely without credible support.

    Here is a link to a Search Engine Land report where Sergey Brin said that links are only 1%: http://bit.ly/eeqwI

    Keep in mind that for every monetized query where SEOs resorted to inflating the value of sites through links, there are thousands more where no links have been used. The search algorithms get along just fine without all the spammy links.
  • Rand - totally agree re links. But isn't getting links primarily about creating great content?

    Read the article you link to btw and am in complete agreement.
  • randfish
    Tragically, at least in my experience, the answer is a resounding no. Great content is easily missed by the web's link-heavy audience, while some pretty crummy content that's been marketed well (or made the right connections or comes from the right sources) will tend to overperform.

    The web's link graph isn't a meritocracy - like everything else in life, it's a popularity contest. Those who find the best ways to distribute, promote and market their works to the audience most likely to link to it are going to succeed much more so than just the "great content" producers.

    Just think of it like politics. The best, most rational, reasoned, intelligent arguments are the exception, not the rule. Instead, the conversation and media attention (and thus, public awareness) is focused on concepts that are easy to grasp, virally distributable (which often puts rumor and innuendo above fact) and fit a compelling narrative (rather than add complexity).

    A post on this topic - http://j.mp/4tYThK
  • Interesting - really like "linkerati" concept (in fact may deserve a full blog post to discuss further...).

    In working on natural search for Hunch, while things have generally gone well (hundreds of thousands of inbound organic clicks per month a few months after launch), some things that has surprised me are
    1) pages with really low bounce rates aren't getting rewarded the way I would have expected. If I were building a search engine I would make this a strong signal.
    2) pages with lots of text seem to be disproportionately rewarded, even when all other things (inbound links etc) seem to be equal
    3) in general we seem to get rewarded for sheer volume of pages indexed, creating perverse incentives to simply generate more pages even if the quality is lower
  • randfish
    None of those surprise me at all (and are very similar to the experiences we see of a lot of our members/clients).

    #1 - Bounce rate is an exceptionally noisy and easily manipulated signal (think Mechanical Turk with lots of Google account logged-in clickers). Spammers watch these "usage data" signals all the time, and the minute Google makes them a big part of the algo, they'll pounce.

    #2 - Pages with lots of text have the opportunity to earn dramatically more search volume via the diverse range of queries users perform (the long tail principle holds up almost perfectly in search/SEO). It's also the case that you may be conflating some metrics as I'd guess that pages with more text are also on more interesting/valuable subjects and have more contributions/content/links/attention than their text-poor colleagues.

    #3 - Yes; this has been written about quite a bit lately in the SEO sphere, and it's one of the reasons Google's been a bit more diligent of late about tossing pages out of its index form large sites that it deems to be providing lower value.

    I think I might try to write a post about all this myself, as it's always interesting to see how folks outside of SEO see and perceive our field (particularly those who bring strong backgrounds and an open mind - a tragically rare thing).
  • Chris, maybe a little off-topic, But since you are talking about importance of SEO I thought I'll ask this. IMO having the title of the post in the link goes a long way in improving your rank in search engine (maybe not a right technique, but I guess most of the search engine honour that).

    Your's perhaps is the only blog which keeps the cryptic link name and not the title of the post as its permalink. You said it helps you use the complete link in twitter without having to shorten it. But aren't you losing out on some SEO points?
  • You are right - this blog is just a fun side project for me. If I were a better sysadmin/had more time I would definitely do what you say.
  • Nice job breaking down why “Search Engine Optimization” is not an optimal name (reminds me of how people misinterpret the meaning of “Net Neutrality”). Countless people mislabel SEO as being the same as blackhat SEO.

    Here’s an example of Danny Sullivan rebuking fox on the subject http://searchengineland.com/dear-fox-news-seo-i...

    and Rand Fishkin referring to a thread on Metafilter http://www.seomoz.org/blog/sigh-how-can-we-over...
  • As Fox News would say: "Some say SEO is a dark, illegal activity. Others say it isn't so bad..." :)
  • My early marketing/SEO experience revolves around social media and referrals. Google may like my stuff in a year or two, but I've got ideas that need testing against sharp intellects now.

    That means building a ubiquitous web presence, commenting at each opportunity to find like minds, and creating virtual landing pages/signs to make it easy for browsers to find me.

    Crowd source sites like Hacker News and reddit quickly and ruthlessly filter subpar content thanks to the attention of many thousands of curious readers.

    I found your blog from Fred Wilson's virtual bar/coffee shop, AVC. I found AVC from Hacker News. I found Hacker News from reading Paul Graham, and I found Paul from googling startups, then hacker on another occassion.

    I seek out the best communities of shared knowledge. There's a vibe of optimism, of change, and altruism. Those in a position of flow are helping out the folks that are just getting started or changing career paths. Startups have become a Zen/compassion creative space. It's amazing to see the juxtaposition of such critical judgement combined with a super charged sharing network of rich information.
  • Chris,

    Your doing the 'SEO thing' in-house makes perfect common sense. Derek Powazek (ex Blogger & Technorati) had an insightful post on SEO, which opens explosively as:
    ---
    Search Engine Optimization is not a legitimate form of marketing. It should not be undertaken by people with brains or souls. If someone charges you for SEO, you have been conned.
    ---
    Here is the link to his post: http://powazek.com/posts/2090.

    And my own take on SEO scamsters: http://vsagarv.posterous.com/against-the-great-...

    Not sure what your policy is w.r.t posting URLs as part of comments. Delete them if it is an issue.
  • Yeah, I read that. I think I am interpreting the word "SEO" in a broader sense as including creating good content that leads to organic search traffic. That said, I agree for the most part with Powazek's assertion that: "The good advice is obvious, the rest doesn’t work." Although I would say what is obvious to experienced web people might not be obvious to some offline business person first setting up a website, in which case a consultant helping them with the "obvious" stuff can make sense.
  • GeekMBA360
    It makes sense to do most of the "technical aspect" of SEO in house.

    But, I do think there is one exception. Let's say that you're starting a niche site in dating. A "SEO expert" (or whatever you call them) that have extensive experience in that niche could save you a lot of time (e.g. they know what keywords convert better in that space). Your in-house staff could learn this by trial and error, but external consultant will save you a lot of time.
  • Adrian Bye
    Interestingly, when i interviewed the guy who runs the Yelp.com of China, I asked him all about how they do their SEO for Baidu. He said they optimize instead just for Google and it ends up working for Baidu (and he's 100% Chinese, living in China):
    http://meetinnovators.com/2009/07/02/tao-zhang-...
  • Yeah, I think since other search engines try to emulate Google, optimizing for Google tends to optimize for others.
  • AndreaF
    I agree that good SEO practices go hand in hand with good design practices so not everything is bad, but I hate the idea that we have to be careful on what we do on our sites and how we do it just in case we upset Google. Don't get me wrong, many SEO related requirements are important for the users anyway, but there must be a better way than just praying that Google likes your site. I suspect things will need to change sooner or later and there will be some sort of universal index that supersides Google's. At least I hope so.
  • Yeah, I agree. Google is so powerful now that it's a bit scary.
blog comments powered by Disqus