Skip to main content

Web Site Analytics and SEO Planning Strategies

Proposed by Peter Chen, Mark Branom

Notes

Photos of the notes that were on the board:
http://t.co/vixGrm84
http://t.co/cWrkXVSR
-----------------------------
Peter Chen
Technical Webmaster
Was at Stanford from 1996 - 2006
Jamie Taui - School of Medicine
What are others doing with SEO on their site? Analytics? What kinds of content?
Chuck Gasperi it
Vicky - hr
Phil - Sciences, IT management
Mark - Alumni Association, analytics
J Gingold - Alumni Association / Interactive Services
Manages Web production team
Mark Branom - IT Services Instructor
Linda - IT services central IT web site
Jody Sumrall - Office of technology licensing
thinking bout seo
Eileen bloom - web design & content manager for law school
Bren - Communications Web team
Steve - web Undergrad & financial aid
Carrie - IT services
Susan Watkins - GSB IT analyst
Launched Google Search appliance
* How do you make analytics meaningful?
* What actions do you take to make your site more effective?
* Intranet not Internet per se
Chris - Stanford Univ Press
* Uses google analytics
Jesse - Bioengineering
Kathy Fisher - school of medicine
Todd - lab manager dev biology
---------------------------------------
What does SEO mean?
Phil
Search Engine Optimization, but what to end?
Making things easier to find are not the same as coming up high in a search
Chris
Up against Amazon -- commercial enterprises have advantages, like buy keywords
Susan
make note of traffic trends -- examine analytics to alter site design -- put links where they are needed to direct people commonly used areas
Phil
certain pages attract most of the inward traffic -- seems like all the work in creating a complicated site map was wasted
Chris
That's the point of doing the work -- getting people to what they want to see
J
Use clickpaths to see how people are traveling through your site
speaker
commonly used search terms weren't used on our pages, which made us hard to find (dev bio vs. developmental biology)
Mark
#1 mistake people make is forget who the web site is for
most people design their web site to increase enrollment, or whatever they want to do
that's the wrong reason
make it easy for people to find what they want to find
always look at your site from the perspective of the user
Best way to find out what people are looking for: Look at GA -- what keyword terms are people using to find your site?
Once you know ppls search terms, make sure to incorporate those words into the various HTML tags and elements that google recognizes: -- title h1 h2 h3 --
these tags get the highest weight of page content
title is most important -- also helps people organize their histories, bookmarks
when crafting a title, imagine you are someone searching for your web site
what terms did you use? now, incorporate those terms into the title h1 h2 h3 of your site -- will improve your rankings
google also looks at link text and incoming links -- "link juice" -- powers your site when someone with good juice (like Stanford) links to you
link to other departments, and vice versa
Make sure your highly used search terms appear in the text of your links
the more links that link back to your site, the higher your ranking will be
Linda -- one of our challenges with Drupal is that all h1's go into the navigation -- if the links are too long they clog up the nav and make it difficult to use
Mark - he's heard h2 - h6 get ranked identically; h1 gets higher, title gets highest
metadata is NOT used by google to rank your site. Only meta = description is important, because it shows up as the description in your Google search results.
meta name=keywords is best only for obscure terms. common terms don't improve your ranking here.
Susan -- actual content of page helps too!
Mark
google will punish you if you try to spoof their page rank system
BMW got in trouble for using hidden meta and header tags
Google will block sites that try to spoof or trick their spiders
Peter -- use your robots.txt file to exclude content you don't want indexed
Mark -- robots.txt can exclude directories, entire sites, certain bots while letting others in
meta name = robots can help in a file to exclude just a single page, not a full site or directory
robots.txt does wild card matching
How do you compete with paid search?
If google has a sitemap, any page can generate sub-links
Generate a site map -- site map is an xml document that describes the links and hierarchies within your site
When you upload this to google (Google Webmaster Tools), google will use this information to spider your site
Does anything other than Google matter?
Bing has similar tools, but in reality, Bing pulls results from Google. Google busted Microsoft in this scheme by posting phony data, which got indexed and appeared on Bing
iPhone Siri searches use default search engine -- still probably Google -- but users can change
Yong
Facebook analytics can give useful information -- links, clicks, likes
Facebook generates traffic for you even if you don't have a Facebook page
Has like 1500 followers now
Susan -- used Omniture (?), good service, but expensive. gave up some functionality when switched to google
unless you're really taking action on complex analytics, Google is probably good enough
Mark -- If your site is hosted on Stanford servers, use AWStats (free) to analyze your server logs
server logs don't always match up: Sometimes the same, sometimes vastly different
Google tracks pages by Javascript that loads a cookie to track users. If users shut off either of these (not uncommmon) then Google can't track it
AWStats analyzes raw traffic logs from servers -- more reliable, but it's difficult to tell if the traffic is from actual people, bots
J -- Google can't track non-HTML documents. for PDFS and other documents, you need to use server-based log tool
Mark - tools.stanford.edu techtraining.stanford.edu teaches "10 Common Web page mistakes"
webpagesthatsuck.com
physical impairments, screen readers: If you design your page well for people in these classes, your site will also be optimized for Google and indexing
Avoid drop-down menus -- awful for accessibility. Avoid!!
Speaker
look at your competition -- same or similar departments from other schools, other sites with the same audience as you
Mark
Usability testing -- test your site with people in your target audience -- give them tasks to locate certain items, see how easy or hard it is
Find disinterested third parties without an affiliation -- need a fresh look -- should be as close to your target audience as possible.
Peter -- UX group forming on campus -- meeting twice a month: http://ux.stanford.edu/
J -- you need to have a solid information architecture, as well as be SEO optimized
can people find other information on the site once you've pulled them in?
Susan -- her students say they don't navigate, they only search
question now: Based on analytics, do we need to worry about the architecture?
J -- probably yes, but start with some basic usability testing
Photos from whiteboard notes in Mark Branom's Twitter feed:
https://twitter.com/#!/search/realtime/markbranom
Twitter hash tag #itunconf #seo