SearchFest 2010 in review

Tuesday, March 9, 2010

Today I am at The Governor Hotel in Portland, Oregon. I had to leave at the crack of dawn to make it to SearchFest 2010 on time. Which means I was awake at dark o’clock. This also means I got to see the sunrise. That almost never happens.

This is the fourth annual SearchFest, which is Portland’s largest search engine marketing conference. But this is my first time attending the event. Topics include Social Media Marketing, Advanced Search Engine Optimization, Local Search, Analytics, Paid Search Marketing and so on. See SEMpdx.org or follow @SEMpdx for more information.

I am most looking forward to the late morning session on blogging with the very funny Matt Inman (AKA The Oatmeal). I usually live blog about the sessions I attend at WebVisions each year, so I’m going to try to do that here as well. See my notes from WebVisions 2009, 2008 and 2007.

Below is a quick, in-progress overview of the SearchFest sessions I am attending, with links.

Keynote Presentation – Stefan Weitz (Bing)
(8:30 am – 9:20 am)

Bing’s search engine market share is growing. Bing is a “decision engine.” And that isn’t just a cheap marketing term. Except that it is. Lots of shared stats about how Bing is actually usable and growing.

So, what is the future of the Internet and of search? The Web is moving away from text and toward a “web of objects” . . . spatial search . . . semantic modeling . . . emerging modes.

Web of Objects: Real-time Firehouse, Services, Multimedia and Devices. Search can expand what is possible and should “deliver knowledge by computationally understanding user intent.” The goal is to try to disambiguate a very ambiguous query. A search engine result shouldn’t always be a list of links. Attempt to structure, assemble and filter the returned services and objects in an interesting way.

Spatial Search: The Large Hadron Collider as an analogy to search engines. There is a small chance that Bing might destroy the planet. Oh, wait. This might not be the point he’s making. The point is: How do we save Rachael Ray from the wave of Higgs boson particles? Well, by taking all of the data on the Web and mashing it back into the real world . . . by giving it context and reassociating it on interactive maps (physical context). Geolocating tweets and stitching user photos/videos into an interactive canvas.

Semantics: The most overused concept in the history of computing. But this is the year for it. No, seriously. Semantics is just understanding how words and objects relate to each other. It helps use to parse queries and to understand how to ascribe meaning to a web of objects. Perhaps even using natural spoken language: Siri example.

Emerging: Mobile and augmented reality. TED demo. Overlay web video of indoor environments onto existing street-side/map images . . . transition indoors.

Prepare yourself: “Service description, object ranking.”

Competitive Intelligence: Know Thy Competition . . . Know Thyself – Mike Roberts (SpyFu) and Larry Kim (WordStream)
(9:30 am – 10:20 am)

Wisdom of your Frenemies – @mrspy

SpyFu is a competitive intelligence company. How you can take a list of keywords that your competitors buy and create what is the Holy Grail of keyword research: Find out the most profitable keywords of your competitors. SpyFu was the first and has 3 1/2 years of ad history.

When your competitor buys a keyword . . . they care enough to bet some money on it. When they buy the same keyword twice . . . it might be profitable. Or:

  1. They lost money, but it didn’t suck that much, so keep trying.
  2. Inconclusive . . . not enough volume to make a decision.
  3. Asleep at the wheel.
  4. Vanity buying, for branding, client imperative and just because.
  5. Randomness and libel.

So, knowing that a competitor buys a keyword isn’t enough.

There are antidotes for the above list. Antidote: Time. Longer = better, because it eliminates unprofitable reasons. Should be consistent and recent. Antidote: Money. Expensive keywords have to be profitable on a long enough timeline or even lazy or determined people will stop buying. Antidote: Activity. Note changes in ad copy, etc.

Which ARE your competitors’ best keywords? For one: Buy your brand. It will always be profitable. Five out of the top 10 keywords for eBay are its brand. Eight out of 10 for Apple.

Compile all of your competitors’ keywords. Identify the profitable ones. Filter out the ones that aren’t universally profitable. (Make a Venn diagram of profitable ad keywords of competitors and see what overlaps and will also be profitable for you.) Highlight the ones to act on now.

Competitive Intelligence in Search – @larrykim

Four popular competitive intelligence blunders (and how to avoid them):

  1. Assuming your competitive intelligence data is relevant (don’t blindly assume).
  2. Too much analysis, not enough action! (make actions scalable and data-driven).
  3. Mistaking estimated CI data actual campaign data (still give it a try). Ad Rank = Bid * Quality Score. Reflect back intent of the search in the ad text and landing page design to get a better quality score and conversion ratio. Take unfavorable CI data under advisement, but don’t be deterred.
  4. CI data is an important component of keyword research – should be use in addition to other keyword research tools. It should not be used as a substitute. Longer-tail keywords from your site’s web analytics can be effectively integrated to your own site’s ad keywords.

Target sub-niches (long-tail keyword variations). See Wordstream’s Free Keyword Niche Finder. Avoid the “path to madness.” CI is an important component of keyword research for PPC or SEO, but pick keywords that are relevant to YOUR site first. Be bold. Act. Don’t be deterred. WordStream provides keyword management tools for PPC and SEO.

Technical SEO: It’s not just about being attractive . . . you (and your website) also need to be accessible – Todd Nemet (Nine By Blue) and Matthew Brown (The New York Times)
(10:40 am – 11:30 am)

Improving Search Coverage – @nemet

Improving search coverage with increased crawler efficiency and decreased page speed. Page speed will become a factor in ranking soon enough. Ignore page speed at your peril . . . Google delay.

Ten directives:

  1. Avoid duplicate copies of your site (development versions, subdomains, https/http, www vs. non-www). Ues Robtex Tools (www.robtex.com/dns) and Google Webmaster Tools (define www vs. non). Stay out of duplicate content filters.
  2. Make your site structure obvious.
  3. Make redirects permanent and short. Use 301, not 302 or meta or JavaScript. Detect: Try LiveHTTPHeaders add-on in Firefox.
  4. Don’t send crawlers on a good chase. Move paramters out of the URL. Use rel=nofollow for links.
  5. Gather similar content under one URL. Use: <link rel=canonical href=”…”>.
  6. Discourage/disallow crawling of unimportant pages. Robot out site search result pages. Require authentication for development/admin pages. Return 404s (avoid “soft 404s”).
  7. Work with browsers, not against them. Network Round Trip Time (RTT). RTT time matters when you’re doing a DNS lookup, TCP handshake. Minimize the number of DNS lookups and file transfers. Combine .js and .css files. Use CSS sprites (one image that gets shifted) instead of a lot of icons. Order of .css, .js, inline <script> matters. Always put .css first for maximum page speed.
  8. Served optimized files, including images. Accept-Encoding: gzip. “Minify” JS and CSS. Use Google’s Page Speed plugin.
  9. Avoid serving files at all, if possible. Return 304s and check modified/cached. IMS preferred. Use cache-control header: max-age, with expiration.
  10. Take advantage of available tools: Google Webmaster Tools. Google Page Speed, Yahoo! YSlow, Firefox Firebug, Web Inspector (Safari, Chrome), Firefox LiveHTTPHeaders, Redbot, Robtex DNS Tools.

Already in effect in Google AdWords: Your bid goes up if your page is slow (above four seconds of load time).

How to Make Awesome Technical SEO Mistakes – @matthewjbrown

Build the SEO into the work foundation of publishers. Technical/editorial SEO work together. Examples from nytimes.com: SEO horrors. SEO automation challenges: auto-gen content, automated tags, default CMS settings, etc. Avoid creating redirect loops and 302 and JavaScript redirects. Be careful with CMS autopopulation of title tages, URLs and meta tags. Take advantage of Image Search by using alt and title tags.

Technical Tools: Server header checker (URI Valet), SEOMoz.org, SeoBook.com, Xenu’s Link Sleuth, Chrome Toolbar, Google Insights. SearchCLU site analysis: Google Webmaster Tools on steroids. Notice and correct domain-wide errors and page/template errors. Make sure site architecture lines up with marketing goals.

The Blogosphere: Boosting Performance Through Connection – Matt Inman (The Oatmeal) and Rebecca Kelley (10e20)
(11:40 am – 12:30 am)

The Basics of Blogging – @rebeccakelley

Blogging is an easy way to create content, bring traffic to site, attract links, get your site ranked, manage reputation, engage users and build community. Blog blah blah. I am charging my laptop. Sorry, Rebecca, but I am mainly waiting to hear Matt talk about velociraptors and punching dolphins in the mouth.

The Oatmeal – @Oatmeal

(Since Wi-Fi was so poor at The Governor Hotel, I lost some of the notes I originally typed for this session when a WordPress save timed out, but this is what I remember.)

Matt started The Oatmeal humor blog about nine months ago. Before that, a couple of years prior, he founded Mingle2 while working a day job. Mingle2 is an online dating site that became hugely popular through Matt’s illustrated linkbait, fun quizzes and viral marketing.

After selling Mingle2, he shifted his focus to The Oatmeal, which is essentially linkbait and viral marketing for its own sake. Now Matt makes $11,000 a week and has recent book and TV deals, just for making funny Internet crap. He says the secret to his success is focusing on humorous, annoying topics that everyone can relate to. He mentioned the hilarious Zero Punctuation and used it as an example of effective video-based humor as well.

If businesses are looking to “go viral,” Matt says the best thing you can do is attach weird or geeky things to your commercial properties. He says that Digg, Reddit and StumbleUpon are hugely important to his success. Here are his recommendations for having submission success with each:

  • Digg: Just make good content and become an active member of the community.
  • Reddit (social news site): Submit your work yourself and write it using a first-person perspective. Redditors want to promote individuals creating work as part of their community.
  • StumbleUpon (install the toolbar): Make sure to have a giant graphical header on your pages. “The Vietnam of viral marketing.”

The main takeaway for in-house designers and marketers is that being creative, experimental and social in your efforts may serve you well. So, basically, you all need to go back to your offices and incorporate illustrations of vicious dinosaurs and bears into your marketing campaigns for maximum humor.

(For these last three sessions, I will paste in my notes from Word soon, since lack of Internet connection prevented me from live blogging.)

SEO Tools: Learn how to leverage advanced SEO tools to improve the performance of your site – Rand Fishkin (SEOmoz) and Richard Zwicky (Enquisite)
(1:30 pm – 2:20 pm)

10 SEO Problems & the Tools to Solve Them – @randfish

See complete presentation: http://www.seomoz.org/dp/10-seo-tools

Stuff You Possibly Might Not Already Know and Tools – @rzwicky

Notes coming soon.

Measuring Online Success: Top Down and Bottom Up: We’re not being anal-retentive . . . Analytics matter – Eric Peterson (Web Analytics Demystified) and Aaron Gray (Greater Returns)
(2:30 pm – 3:20 pm)

Notes coming soon.

Local Search: Up-to-date strategies for achieving prominence within Local Search listings – Mary Bowling (SEOverflow), Matt McGee (Small Business SEM) and Chris “Silver” Smith (KeyRelevance)
(3:30 pm – 4:20 pm)

Notes coming soon.

Similar posts that may be of interest:
    None Found