Personalisation of search: take back control Karen Blakeman, RBA Information Services 5th June 2012 Pre-conference workshop, 11th Southern African Online Information Meeting, Sandton Convention Centre

1
Personalisation of search take back
controlKaren Blakeman, RBA Information
Services5th June 2012Pre-conference workshop,
11th Southern African Online Information Meeting,
Sandton Convention Centre

• Slides are available at http//www.rba.co.uk/as/
• karen.blakeman_at_rba.co.uk
• Twitter _at_karenblakeman
• http//www.rba.co.uk/

This presentation is licensed under a Creative
Commons Attribution 3.0 License
2
General plan for the session

• Getting to know one another and feedback on
  search issues
• Slides are a basic framework for the session
  (download from http//www.rba.co.uk/as/ now if
  you wish). I'll create an addendum of key,
  additional information after the session.
• Ask questions as we go along, or write on the
  notelets and we'll have QA slots throughout the
  session
• Summary of "stuff" we've learned and what to take
  back home and to work

3
How it all started

• Before 1992 priced electronic databases indexed
  by humans. Many still exist for example
  LexisNexis, STN, Dialog
• 1992 the Internet can be accessed by anyone but
  2-3 years before significant information started
  appearing on the web
• Increase in amount of data and information led to
  the development of tools that indexed and
  searched the content of web pages
• Lycos, Excite, AltaVista, Hotbot

4
How the web search tools worked (and still do in
part)

• "Crawl" the internet looking for new and updated
  pages by following links
• Copies of pages and documents added to a database
  that is publicly searchable
• Results sorted according to
• how often the words you looked for appear in the
  page
• where they appear (words in the title and first
  few sentences given higher ranking)
• and other criteria not disclosed by the search
  engines
• They do not cover
• password protected sites
• databases or sites where you have to fill in a
  form to find the information

5
Then along came.....
11 November 1998 From the Internet Archive
www.archive.org
6
How was Google different?
Links (citations) a major part of sorting search
results http//www.seobook.com/learn-seo/collater
al-damage.php
7
Google 2011
Revenues 37,905 millions Net Income 9,737
millions http//investor.google.com/financial/tabl
es.html 2011 96 of revenues are from
advertising Google has a problem....
8
Google's problem "How People Spend Their Time
Online Stephen's Lighthouse" http//stephensligh
thouse.com/2012/03/14/how-people-spend-their-time-
online/
9

• Search engines and social networks need to keep
  you on their "properties" captive audience
• Always trying to deliver new services to stop you
  wandering off to other sites
• To keep you on-site need to deliver content as
  relevant as possible to YOU
• need to know more about you
• need to know what sort of information you are
  interested in
• need you signed in to your account
• need to know who your contacts are on-site and
  elsewhere
• need to know about your activity elsewhere
• Personalise results

10

• The Filter Bubble What The Internet Is Hiding
  From You
• Eli Pariser
• Publisher Viking (23 Jun 2011)
• ISBN-10 067092038X
• ISBN-13 978-0670920389

11
Trends in search

• No longer straightforward text searching of web
  pages and documents
• Localisation
• Personalisation
• Social
• Mobile

12
How far does personalisation go?

• Google can seriously damage your news
  http//www.rba.co.uk/wordpress/2011/09/03/google-c
  an-seriously-damage-your-news/
• Is Google really filtering my news?
  http//www.librarianoffortune.com/librarian_of_for
  tune/2011/09/is-google-really-filtering-my-news.ht
  ml

13
How far does personalisation go?

• An Awfully Big Blog Adventure The answer to your
  question... depends on who you are (Anne Rooney)
  http//awfullybigblogadventure.blogspot.co.uk/2012
  /04/answer-to-your-question-depends-on-who.html

Borromeo vs
Borromeo
14
Word cloud of top 20 results for a Google search
on Prague (web history and social networks
switched off, cookies cleared)
Word cloud of top 20 results for a Google search
on Prague (signed in to Google account, social
networks and web history enabled)
15
Google loses the plot? Search on
goats
16

• Dear Google, stop messing with my search
  http//www.rba.co.uk/wordpress/2011/11/08/dear-goo
  gle-stop-messing-with-my-search/

17
Google introduces the soft AND

• When you do a multi-term query on Google (even
  with quoted terms), the algorithm sometimes
  backs-off from hard ANDing all of the terms
  together.......its clear that people will often
  write long queries (with anywhere from 5 to 10
  terms) for which there are no results. Google
  will then selectively remove the terms that are
  the lowest frequency to give you some results
  (rather than none)....Soft AND is a way to reduce
  the overall frustration and give the searcher
  something to examine (and with luck, a chance to
  reformulate their query).
• Dan Russell
• http//www.rba.co.uk/wordpress/2011/11/08/dear-goo
  gle-stop-messing-with-my-search/comments

18
Google

• No more '' to force an exact match
• No more automatic 'ANDing' of your terms
• No more highlighting your search terms in its
  cached copy of a page
• No more clearing search history unless you are
  logged in to a Google account alternatively
  delete all search cookies from your browser

19
Google's new Privacy Policy
"Our new Privacy Policy makes clear that, if
youre signed in, we may combine information
youve provided from one service with information
from other services. In short, well treat you as
a single user across all our products, which will
mean a simpler, more intuitive Google experience."
"we're more excited than ever to build a seamless
social experience, all across Google"
Toward a simpler, more beautiful Google
http//googleblog.blogspot.co.uk/2012/04/toward-si
mpler-more-beautiful-google.html
20
What does Google know about you?
Look in your Google account dashboard
http//www.google.com/dashboard/ How Google is
targeting your ads http//www.google.com/ads/prefe
rences/
21
Impact of Google's policy and personalinformation
management changes on YouTube
Targeted advertising?!
I wonder how Google will customise my web search
based on my YouTube viewing?
Gives me videos based on my web search history,
linked to my location (Reading) plus a long list
of videos mentioned by people in my Google
circles.
22

• Google Enables Cross-Platform Local Search (As
  Carrot To Relinquish Your Privacy)
  http//searchengineland.com/google-enables-cross-p
  latform-local-search-as-carrot-for-web-history-113
  811
• Introducing a new local search experience across
  your devices - Inside Search http//insidesearch.b
  logspot.co.uk/2012/03/introducing-new-local-search
  -experience.html

23
(No Transcript)
24

• Google's new(ish) social network Google Plus
  (Google)
• http//plus.google.com/
• Google Now Forcing All New Users To Create
  Google Enabled Accounts
• http//marketingland.com/google-now-forcing-all-ne
  w-users-to-create-google-enabled-accounts-3912
• Search Plus Your World (SPYW) referred to as
  Search now available in Google.com and is the
  default. Gives priority to content from people in
  your Google network if you are signed in to your
  account.
• (And the next Google killer is.Google!
  http//www.rba.co.uk/wordpress/2012/01/30/and-the-
  next-google-killer-is-google/ )

25
SPYW currently being tested on Google.com
Top results (blacked out for privacy reasons)
were from one of my Google circles and from
people who restricted access to the postings.
Take care when providing information to users
or incorporating data as part of a report
26
Google Knowledge Graph

• Introducing the Knowledge Graph
• http//www.google.com/insidesearch/features/search
  /knowledge.html
• It isn't a graph!
• At present only on Google.com
• Only shows if you are logged in to a Google
  enabled account

27
Google Knowledge Graph
28
Google Knowledge Graph
No Knowledge Graph
29
Bing - Adapting Search to You http//www.bing.com/
community/site_blogs/b/search/archive/2011/09/14/a
dapting-search-to-you.aspx
Bing to use Facebook, Twitter more in fight
against Google ZDNet http//www.zdnet.com/blog
/facebook/bing-to-use-facebook-twitter-more-in-fig
ht-against-google/8631
Bing Relaunches, Features New Social Sidebar
http//searchengineland.com/the-new-bing-microsoft
-tries-again-with-search-meets-social-120728
"Its not just Facebook and Twitter that get to
play in the social sidebar, however. Social
suggestions might also come from LinkedIn, Quora,
Foursquare, Blogger and - wait for it - Google
Plus"
30
Bing Relaunches, Features New Social Sidebar
http//searchengineland.com/the-new-bing-microsoft
-tries-again-with-search-meets-social-120728
31
(No Transcript)
32
No social side bar for me (not in the US) but
social network 'stuff' appears in main results
33
So.cl http//www.so.cl/
Microsoft Launches Socl Social Network A Look
Inside http//marketingland.com/microsoft-launche
s-so-cl-social-network-a-quick-look-12499
34
What I see on my screen is not what you'll see on
yours!
35
How do search engines personalise results?

• Depends on
• your location
• past searches
• which sites you have looked at in the past
• your 1s
• your likes
• your shares
• sites blocked by you (Google)
• which social networks you are signed in to
• who is in your social networks

36
To allow personalisation or not?

• Not necessarily a bad thing
• Use a second browser with search history enabled
  and logged in to accounts for a different point
  of view
• It does bias results
• What I see on my screen is not what you'll see on
  yours
• Be aware of potential privacy issues regarding
  friends and contacts in social networks when
  providing results to your users

37
Want to switch it off?

• Disable and remove web/search history
• - but in Google, no option to erase 'signed out'
  histories
• Actively manage search cookies, automatically
  delete cookies after computer log out or switch
  off
• - How to delete cookies
• - http//aboutcookies.org/Default.aspx?page2
• Log out of all search engine and social media
  accounts when searching
• Use Chrome Incognito (Chrome owned by Google!)
• In Google use Verbatim in the left hand menu on
  results page
• Use advanced search commands if relevant
• Use a search engine that doesn't track or
  personalise

38
(No Transcript)
39
Google search settings
40
Google web history
41
Verbatim

• Forces Google to run an exact match search. Run
  your search first and then select Verbatim from
  the left hand menu on your results page
• Cannot be combined with time options in the side
  bar
• Google Verbatim for exact match search
  http//www.rba.co.uk/wordpress/2011/11/18/google-v
  erbatim-for-exact-match-search/

42
Signed out
Chrome incognito
Verbatim
43
Try a search tool with less or no personalisation

• DuckDuckGo does not track, does not personalise
• http//duckduckgo.com/
• Yandex.com International version of the Russian
  search engine
• http//www.yandex.com/
• Blekko
• http//www.blekko.com/
• Million Short omits the top million most
  "popular" sites from results
• http//www.millionshort.com/

44
Use more than Google anyway

• The Disruptive Searcher (Sanity checking Google
  http//disruptivesearcher.wordpress.com/2012/02/27
  /sanity-checking-google/)
• if I hadnt searched across more than Google for
  data on a small, new company that I was asked to
  research recently, I would have missed out on
  some very significant information that Google
  just wasnt showing me.

45
Bing

• http//www.bing.com/
• Does personalise search and include social
  network content in results
• Most of the interesting developments and features
  are only available in the US version
• Results tend to be more consumer/retail focused
  unless using advanced search features
• Coverage not identical to Googles - sometimes
  yields important unique content, especially in
  research and business
• Sometimes more up to date than Google

46
Bing

• Link to minimalist advanced search options now
  vanished
• Advanced Search Operators
• http//msdn.microsoft.com/en-us/library/ff795620
• Main ones
• site
• filetype
• intitle

47
DuckDuckGo

• http//duckduckgo.com/
• DuckDuckGo silly name but a neat little search
  tool http//www.rba.co.uk/wordpress/2011/11/07/d
  uckduckgo-silly-name-but-a-neat-little-search-tool
  /
• No tracking, no filter bubble
• Commands
• site inbody intitle filetype
• sortdate to sort by date (uses results from
  Blekko)
• regioncc (e.g. za) to boost a country
• Syntax and keyboard shortcuts at
  http//duckduckgo.com/goodies.html

48
Yandex

• http//www.yandex.com/

49
Yandex advanced search
OR Search operators http//help.yandex.com
/search/?id1113759
50
Blekko

• http//blekko.com/
• slashtags for sorting by date (/date), searching
  for images (/images) and videos (/videos)
• Use public slashtags to search a group of web
  sites covering a particular topic or type of site
  e.g. /library or create your own to search your
  specified list of sites (similar to Google Custom
  Search Engines)
• wind turbine electricity generation
  /karenblakeman/renewable
• Musings about librarianship Using Blekko to
  search across thousands of library sites
  http//musingsaboutlibrarianship.blogspot.com/2010
  /11/using-blekko-to-search-across-thousands.html

51
Blekko

• Cannot do filetype, inurl, intitle searches
• Drop down menu next to page in results list for
• site search (or use /site)
• similar pages (or use /similar)
• inbound links to the page (or use /links)

52
Million Short

• http//www.millionshort.com/
• "Imagine a search engine that simply removed the
  top 1 million most popular web sites from its
  index. What would you discover?"
• Can use
• filetype
• site
• intitle
• "..."
• Can add individual sites back in

53
(No Transcript)
54
Using Google features to best advantage
55
Location

• Country versions of Google to prioritise local
  content
• for example google.co.za, google.fr, google.de
• usually two letter ISO code for the country
• Change location in left hand menu on results page

56
Google.com and SPYW
57
SPYW hide personal results
58
Personal results only
59
Verbatim
60
Google results page side bar
'Everything' does not search everything
Videos is not YouTube
'Social' is not just Google. Look in your
dashboard to see who Google has decided to
include.
Not clear how discussions are identified
No Twitter option anymore and Twitter coverage is
sporadic.
61
Google side bars
Images
Videos
News
Books
Blogs
62

• Start using advanced search commands and Google
  gives up on personalisation although you may have
  to use Verbatim

63

• Looking for a particular type of information for
  example statistics, research report, expert
  presentation?
• Use the filetype command
• For statistics
• world oil consumption filetypexls
• world oil consumption filetypexlsx
• world oil consumption filetypexlsx OR
  filetypexls
• For government, research, industry reports
• oil consumption forecasts filetypepdf
• For conference presentations or trying to locate
  an expert
• renewable energy UK filetypeppt
• renewable energy UK filetypepptx
• renewable energy UK filetypeppt siteac.uk

64
Numerical range search

• Anything to do with numbers
• Use advanced search screen
• or
• 1st number followed by two full stops followed by
  2nd number followed by unit of measurement (if
  applicable)
• Norway oil production forecasts 2012..2020
• Norway oil production forecasts 2012..2020
  filetypexls OR filetypexlsx

65
Advanced commands continued

• inurl for example
• inurl"carbon capture" targets
• intitle for example
• intitle"carbon capture" targets
• asterisk () to search for terms separated by 1-5
  words (may have to use quotation marks)
• solar panels
• "solar panels"
• Picks up solar PV panels, solar photovoltaic
  panels, solar water heating panels

66
Synonyms

• Google often looks for variations of your terms
  but you cannot rely on it always happening
• Use the tilde before a term to look for what
  Google considers are synonyms
• energy will pick up oil, fuel, gas, electricity
• No information/documentation on how synonyms are
  created
• Very general, consumer oriented rather than
  scientific
• Can be used with Verbatim

67

• When you really DO want to search social media
  the main search engines don't make it easy!
  Social media is an essential part of many types
  of research.
• Search within the network itself means you must
  have an account
• Use specialist search tools
• come and go
• not comprehensive
• need to use more than one
• a few examples are shown in the following slides.
  For more information on social media search tools
  see Phil Bradley's presentations
  http//www.slideshare.net/Philbradley/

68
LinkedIn.com
69
LinkedIn.com
70
More on searching LinkedIn

• Boolean Black Belt-Sourcing/Recruiting
  http//www.booleanblackbelt.com/
• Mary Ellen Bates Ten Top Tips for Searching
  LinkedIn http//www.batesinfo.com/meb123/index.htm
  l PDF http//www.batesinfo.com/extras/assets/link
  edin.pdf

71
Bing social http//www.bing.com/social/
72
http//search.twitter.com/
73
http//search.twitter.com/
74
Topsy.com
75
Topsy.com
76
Icerocket.com
77
Socialmention.com
78
Keeping up to date

• Inside Search http//insidesearch.blogspot.com/
• Official Google Blog http//googleblog.blogspot.c
  om/
• Google Scholar Blog http//googlescholar.blogspot.
  com/
• Search Engine Land http//searchengineland.com/
• Search Engine Watch http//searchenginewatch.com/
• Boolean Black Belt-Sourcing/Recruiting
  http//www.booleanblackbelt.com/
• Karen Blakemans Blog http//www.rba.co.uk/wordpre
  ss/
• Phil Bradley's weblog http//philbradley.typepad.c
  om/