Main goal: Produce a web site with psychological and psychiatric advice and information for non-experts.
Find the right answerOur own manual assessment (100 randomly selected real user questions and answers, October 2003):
Recall = Percentage of relevant answers found, 43/53 = 81 % Precision = Percentage of found answers which are relevant |
|||||||||||||||||||||
User satisficationControlled user testing (11 users):
|
The English site was opened in July, 2003. The German site was opened in November 2003.
|
German |
Greek |
English |
Italian |
Swedish |
|
Total in FAQ data base | 515 | 290 | 799 | 220 | 661 |
Has text | 514 | 289 | 798 | 220 | 656 |
Lacks text | 1 | 1 | 3 | 0 | 4 |
Searchable and has text | 473 | 41 | 711 | 58 | 629 |
Not searchable and has text | 41 | 248 | 46 | 162 | 28 |
In link tree and has text | 707 | 616 | |||
Not in link tree and has text | 89 | 41 | |||
In link tree but not in data base | 22 | 110 | |||
FAQ and has text | 475 | 274 | 730 | 181 | 617 |
Menu and has text | 38 | 16 | 40 | 39 | 14 |
Heading | 2 | 0 | 26 | 0 | 26 |
Method |
Why it works |
How we used it |
High quality web site |
Search engine rank web sites |
Our main goal |
Get links from external sites |
Search engines count links to web site. |
We asked a number of high quality sites to set links
to our site. |
Submission to search engines and portals |
Users come from them, and search engines rank by them. |
Yahoo, The Open Directory, the Sunet Directory. |
Search phrase optimization |
Search engines ranks by count of phrases, especially
in titles. |
We found popular search phrases for 78 of our best FAQs,
and optimized them. |
FAQ identifier |
child-trauma-symptoms |
Original title | How children react to trauma |
Optimized title |
Effects of child abuse |
No. | Keyword | KEI Analysis | Count | 24Hrs | Competing .. |
1 | "effects of child abuse" | 51.124 | 187 | 70 | 684 |
2 | "child abuse statistics" | 43.210 | 402 | 150 | 3740 |
3 | "child sexual abuse trauma symptoms" | 16.000 | 4 | 1 | 0 |
4 | "signs of child molestation" | 9.846 | 16 | 6 | 26 |
5 | "articles on child abuse" | 9.074 | 82 | 31 | 741 |
6 | "symptoms of child molestation" | 6.750 | 18 | 7 | 48 |
7 | "child molestation law" | 4.500 | 15 | 6 | 50 |
8 | "information on child abuse" | 3.531 | 161 | 60 | 7340 |
9 | "child abuse laws" | 3.522 | 127 | 47 | 4580 |
10 | "child molestation statistics" | 3.411 | 18 | 7 | 95 |
11 | "child molestation" | 0.671 | 213 | 80 | 67600 |
12 | "child abuse prevention" | 0.564 | 241 | 90 | 103000 |
13 | "child abuse cases" | 0.445 | 112 | 42 | 28200 |
14 | "child sexual abuse" | 0.431 | 373 | 139 | 323000 |
15 | "child abuse and neglect" | 0.109 | 168 | 63 | 258000 |
16 | "molestation" | 0.058 | 121 | 45 | 252000 |
17 | "child trauma" | 0.002 | 3 | 1 | 5900 |
So we put "Effects of child abuse" in <TITLE>, <H1> and once in the text of the FAQ.
White text on white background |
Obviously unethical |
Cloaking = showing special page to search engines |
Dangerous, ethics depend on how you use it |
Doorway pages = special pages to attract visitors | Depends how you use it, we have one or two such pages |
Approximate percentage |
|
Browsing our static pages |
60 % |
Using Google and other search engines to get directly to one of our answers | 25 % |
Browsing our forums and ask-the-expert areas | 8 % |
Using our natural-language question-answering system | 5 % |
Other methods | 2 % |