Search Engines

History

FTP

Annonymus ftp file listings weresearched (i.e. Archie)

Gopher (predecessor to www):links (mulitmedi,a text etc. couldnot be displayed, just downloaded)i.e. jughead

gopher index system were searched / indexed

www

First full text crawler based vs. directory or indexed based search.technology first indroduced"webcrawler" used then used by allothers i.e. Lycos (1994, project by carnegiemellon university)

Google (2001)

Key Players

Google

Pupularity:Numer of Linkspoining to the page

Asspt. good pagesare pointed to morethen not so good apges

--> count of number of linkspointing to page = rank

Other Ranking techniques

Presence of Keywords

Have people revistied an clicked on s.th. else?

Yahoo

< 2004 Google ran in thebackground

Since 2004: own search engine poweredby inktomi and AltaVistawhich Google bought

MSN --> Windodw Live Search--> Bing

built own search enginge

SEO Links

English

SEO Fast Start (Dan Thies)

German

Determining Relevance

Googles concept of Signals

"Signals" determine how high site appears on the retruredsearch list

Google's Search Rank(200 Variables!?)

Google Page Rank (how important is this page in general)

Relevance from 0-10

Psyma eBusiness (3/10)

low

ToolBar that supposedly tells how valuable a linkfrom that site is - however this is doubted and the rank shown is notup-to-date (3to4 month) while real value is updated daily

Companies however pay money based on that toolbar rating ("I will pay less"for my link on your page because your rank hasdropped)

Pay Sites high on Goolge listList URL money for Usability link to our page

Determine Important pages

Page rank from search words

Test page using googlecontructing diff: comb. ofkeywords important foryour site and seehow good page performs

Determine Importance of Pages in whole site

Rank titles of several pages using google

Determine quality of visitors on a certain page (prepared to buy?)

Another form of payment is link exchange Two PR4 links for one PR7

the real page rank=recursive algorith:

Qualtiy and Quantity of Linkspointing to your page

Quantity: How many

Quality: How many and howimportant are those pages

Non Google

Other search engines also pay attn. to meta tags

items receive more weighting

MetaTAGs etc.

Name of URL(i.e. usability.de)

Title

Meta Tags

Description

Keywords

Beginning words of the content

Frequency of words searched for:Words repeated several times

Concept of Factors

By Off Page Factors

Google:

Clickthrough Rateof search results

Google statisticlyknows what kind of descriptiontexts are most clicked when usersused specific keyewords

How good is the "scent"of information created by the data retrieved by Google search (a) Title b) Description )

Determined by descriptivnessof Title, Description Tag and Content

Psyma.com has empty description tag

ebusiness research tell about frank knapp

Quality factor

Quality and number of inbound links

no

Number of external links that point to our site

Links that directlypoint to our page fromthe outsinde

Inbound Links(a major determinandof Googles quality factor)

add Possibility to let user add Newsfeedsto your site

Blogs

no

RSS

no

Use social bookmark services buttons

Add button to add page to social bookmark services

Delicious

only me

Mr. Wong

Ask member or friends to book mark

no

Sponsoring

Place Logo on UPI

no

Exchange Links

i.e. with other psyma daughters

?

Write Wikipedia article about Psyma

yes, but only ebusiness not usability mentioned

Distriubute Articles that point to your site

no

Problems with inbound links

Yahoo tells thruth

check !

Google

Toolbar only part truth

no inbound links!

Web-Directories

Listing of Site/ Page in Webdirectory(Example for psyma eBuisness page)

web.de

y

dmoz

n

yahoo directory

n

allesklar.de

n

Subtopic

By ON-Page Factors

Keywords

Its not that "I'm numberone for this keyword" ...

... but which Keywordsdo users use?

Do internal keywordstormingfor Target group

let users make suggestsion

unlcear

Check Competitors keywordskeyword density using seaquake

no

Check keyword tools

i.e. seaquake

Check google adwords tool

Check competitors

Check for other related and hight ranked terms

Best may be combination offew competitors and highsearch volume

no

Keyword Concepts

Single

Hight Potentials

Many competitors or get lucky

Long Tail

Use (rare)combinations

more probable to getfew but qualified users

"user experience optimization"

What are websites goals?

Psyma goal not specified

Branding

Appear often in searches on top ranks

sol. use google ads

no

Traffic

Ad has high rank

use google ads

no

Conversion

Attract only qualified users,too many users cost too much

Landing page optimization

unclear

Usabilty - lead users to relevant content

no testing

Integration of (previsously checked) important (=high volume or qualified) Keywords

Overall strategy: optimize pages from users perspective aking:what keywords would the use?

embed (high volume )keywords in text

optimize pages for single, not for manykeywords

i.e. "usability testing"

no

"international qualtiative market research" (start page)

no

in title

in header

in links

in tags

no

integrate alternative versions of keyword (i.e. ebusiness)

singlular

no

plural

no

Quality of Landing page:

use of keywords

not enough compared to comp.

not enough of the right ones

usability

not really tested

Indexing: How many pages of thesite are indexed?

How many urls of mydomain has google savedsite:psyma.com

Prob: Google saved allmain links but not usability

solutions

Psyma is ranked .... even if you search for "international market research"

sol: get more inbound links

no

sol: do internal deep linking

no

sol: create rolling sites

unclear

Too many overview pages

Sol: Flat site strucure

unclear what the problem is

sol: keep creating content

?

Not enough deep links

Internal solutions

Use "Similar pages"

customer who saw this also saw that

no

sol: give no follow to human links (i.e. impressum)other links will become more strong

no

sol: create dynamic sitemaps

not recognized by google

sol: create glossary

no

Technical quality

Bad Programming?

other tools show programming errors

double domain (www or non- www)diminishes valu of both

yes

Links cant read because theyare graphics

Sol: Use alt tags

Other problems(repetitions and cheating)

Cheating: False Keywordsto increasevisibility / page rank (i.e. Sex Sex Sex Money Money etc.)

Due to fake keywords Google will not look here?

Repetition

Cheateing?: use of too many(the repetition) of search words

Sol: Avoid Repetition

Problem of 2 equal domains: www. and no www

Sol: Permanente Umleitung einrichten

Title

Non-descriptive titels

Make users wonder if thepage is relevant

Sol: Make titles relevantto page content

same titles for several pages

Sol: Every page needs individual title

Title apperas as first line insearch engine

using spelling or termsthat users wouldn't use

Content

repeating paragraphsaccross different pages

i.e.overview pages

Sol: "follow, noindex" tag

MetaTags

Meta Tag Keywords (whichare however ignorde byGoogle)

use of same keywordsfor different pages

Sol: Every page needs individual keywords

Sol: Keywords that describe site purpose bestcan be used on every page to increase impression ofusefulness of page

Sol: Write content first, then takekeywords from content

User use differnt keywordsfor a page

Sol: Keywords can also contain synonyms (eBusiness instead ofE-Buisness)

Keywords are not linked

Sol: Link key words alternatively

Only Main Keywords are supplied for which we have many contenders

Sol: Use specific kombined keywords i.e. Mafo Nürnberg

Sol: Check with Google Key wordstool for alternative kombinations

Sol: Create specific pages for users highly used combinations?

Case specific search

Sol: Use correct case

Description

The description is usually shown beneath thesearch results title. If no description tag isfound then other info is taken.

If not filled out, the firt paragraph is used where the keyword is found in

Often the navigation is the item first read bysearch engines. Then this is what is diplayed

However not with Google or Bingwho ignore the description Tag.

Worte, für die Ihre Seite optimiert werden soll, müssen in der Beschreibung enthalten sein. Die Worte müssen sich auch im Seitentitel, den Keywords, dem Seitentext,... wiederfinden

Repeated keywords are treated as Spam

Sol: Don't repeat

Sol: Write individual keyword per page

Case specific search

Sol: Use correct case

Headlines

Words in Headlinesare very relevant - Keywordsshould appear in the headline

Important for relevance are onlylarge headlines (h1 h1)

Sol. Use <h1>Überschrift</h1>in html code but control headlines in CSS

Basic Tech Functioning

Front End

Search Engine Software

Problems: Being up to date /"live"

Size and Time

Updates: Frequent updates (i.e. News pages)

More and More pages

Dynamically created pages

Solution: Crawl only top noch 15-20% of the web every day - rest every two weeks

Spurious Results

Therefore Keywords searched forshoud be found in the same paragraphmake it a relevant page.

Back End

Webcrawler

Indexing the a website(=following Links) and copyingcontent to index of a page.

SEOFastStart

Factors thatdetermine "Page Rank" (=quality of site or pagefor search terms)

On Page (dirct control)

Keyword Postion makes asite more relevant: "Where" is thesearch term

Heading, Page Titlemust promote search termsthe page is optimized for

Page Titles

Precise Matchingof Search Terms

Headers

Maching of Search Term

Must be "bold"

Meta Tag "Description"

... not important for Google

Correct/ Standard programmingensures readability by Spiders

Images without "alt" tags are problematic

Off Page (Indirect Control)= more influentialthen OnPage factors

Other Websites

Pages linking to you

How many other pagespoint to your site

Their popularity / relevance forthe subject

the text in the article that contains the link

Ex: Hamsterfood on Hamster Sitebetter than link from non Hamser Site

The text of the linkitself ... buy hamsterfood

special term is called"link reputation"

User Experience/User Satisfaction

Do users returnto search result page?

Tagged by by Social BookmarkingServices?

So now now that you've got a good rank/ visbility ...

Does your search result text attract users?

Match users searchbehaviours by either ...

Be yourself an "authority" in the subject = high rankingin Googles Organic Search

Be a link on the"authority page"

"If you’re selling products that are available elsewhere, you can get a lotmore attention and traffic by providing such valuable resources as reviews,independent testing, and side-by-side or feature comparisons"

Concept flowing intoGoolges Page Rank

Page rank sorts the results on thesearch results page: The more page rankthe higher on the organic search list

Quantity of Links

The more linksleading to your page the better

Quality of Links

The higher the Weigthting of the page the link is coming from, the betteri.e. the page contains the word hamster food

The higher the Popularity of the page the link is coming from, the betteri.e. the page is often selected from the organic search page

"the quality of the inbound links to your website, in most cases, will be moreinfluential with the search engines than the content of your website itself."

The higher the link relevance, the betteri.e. the link is called hamster food

Keywords matching Search Query

=weight of page

How Google probably does it ...

1. Each page is split into array of words and each word is weightedfor the document.

Rare words have more weightthan common words

2. Each page rank is calculated based on incoming and outgoing links

Outgoing and ingoing the page rankis subdivided betweenthe number of links flowingfrom page (100 links= 10% each)

3. Final results are calculated using weight given by basic ranking multiplied by the PageRank (which is between 0 and 1 always) for the particular page.Ex. The weight of result page A was 2 but pagerank was 0.2 and weight of result page B was 1 but pagerankwas 1 sorting would be B, A (1, 0.4). Without pageranksorting would be A B (2, 1).

Psychological Study: Which words are strongly assotiated with each other?Using word assotiation test psychologists found outthat Google Page rank model predicts assotiations better than other models