Data Feed SEO

A4uexpo London, October 2010

Will Critchlow

Data Feeds Are Not Unique

The “Affiliate” Penalty

Unique Content Matrix

Unique Content, Low Trust

Strong site, unique content

Non-unique Content, low trust

Strong site, non-unique content

Un

iqen

ess

Site Strength

Case Study“Welcome visitor, please find out selection of

[insert product] below, we have [number of products] items. We think you’ll like them!”

User Generated Content

“User” Generated Content

User “Generated” Content

Mozenda

Building quick & dirty SEO ToolsA Cheat Sheet & Inspiration

By Will Critchlow, www.distilled.co.uk. First published: www.seomoz.org

APIs (more on programmable web)AdWords – KeywordsAlchemy – Structured data & textBing – Search, news, spellingEvri – Sentiment and popularityFace.com – Face detectionFacebook – Social graphGoogle Analytics – Visitor dataHostip – Geo dataLinkedIn – Professional dataPingdom – Website uptimePostrank (1, 2, 3) – real-time & influenceRapleaf – Social media profilesTwitter – Real time and social... And of course:Linkscape – Links

YQL – Yahoo! Query Language

select * from html where url=“<url>" and xpath=“<xpath>“

select * from html where url=“<url>"

select * from feed where url=“<url>”

select * from search.web where query = “<query>"

xpath (more examples)/foo – the element ‘foo’//bar – all elements ‘bar’foo/bar – all bar elements children of foofoo//bar – bar arbitrary levels below foofoo/*/bar – bar grandchildren of foofoo/* - all children elements of foofoo/@bar – bar attribute on foofoo/[@bar] – foo with bar attributesfoo/[@bar=baz] – where attribute=baz

PythonSince Python is the language of Google App Engine, here is how you can use YQL easily within Python:Download source – extract to yql folder within your application

import yqly = yql.Public()result = y.execute(“<yql query>”)

Crawlers / ScrapersMozenda80legsGoogle App EngineAmazon Web Services

Human TouchAmazon Mechanical TurkSmartsheet (interface to Mechanical Turk)oDesk

Sources Magic Horsepower

Data (more on infochimps)Data.gov – US government dataData.gov.uk – UK government dataDelicious list – from Peter SkomorochGoogle Public Data - DirectoryGuardian – content and dataWorld Bank – finance, health, etc.80legs – prepackaged crawl data

User Generated “Content”

• External search queries

• Internal search queries

• Tags• Testimonials• FAQs/Support

emails

Tracking # of Reviews_gaq.push(['_setCustomVar',

      1,                   // This custom var is set to slot #1.       ‘Number of Reviews',       // The top level name for the variable      ‘1',   // The Number of Reviews      3                     // Page level variable

   ]);

Context Is KeyGoogle News: Google likes alternative factsLyrics: Never considered duplicate content

Context is key

Look to stand out from your competitors “Use a source of content that’s not

unique, but that no-one else in your space is using”

Manipulate & Clean Your Data“Kingston

DataTraveler 101 USB flash drive - 4 GB – Cyan”

“Kingston USB memory stick 4gb”

vs

Of Course, Links Always Win

http://www.seobook.com/black-hat-seo-case-study

Manual Reviews – aka “Hand Jobs”

Check out the quality rater guidelines

“Add value to users”

“Relevant”These are

subjective!!

Resources

• http://www.seomoz.org/blog/whiteboard-friday-flat-site-architecture

• http://seogadget.co.uk/solving-site-architecture-issues/• http://www.seomoz.org/blog/api-and-dataset-cheatsheet-building-

quick-dirty-tools• http://www.mozenda.com• http://www.seomoz.org/blog/leveraging-mechanical-turk-odesk-el

ance-craigslist-for-seo• http://www.seochat.com/c/a/Google-Optimization-Help/Googles-Q

uality-Rater-Guidelines-Leaked/• http://www.flickr.com/photos/rosaydani/77371897/

Thanks!

Director> will.critchlow@distilled.co.uk> twitter.com/WillCritchlowDistilled

> www.distilled.co.uk

Will Critchlow