Posts Tagged 'Search Engines'

*How Much Information Does A Search Query Reveal About A User?

One search query on its own might actually not reveal too much information about a user. If you however, keep logging the queries from one particular user one might very soon be able to gain interesting insights. At some point early this year Google extended its ‘personalized search’ function onto all users, not matter if they were logged into any Google service or not (explanation found >>here<<).

I was first confronted with this topic at the Suma e.v. conference in Berlin 2007. Hendrik Speck (a indeed humble man who even has a ‘My Quotes’ section on his website) mentioned in his speech search engine log files and talked in detail about how much information search engines could gain out of analysing them. In the example provided by Speck he talked -as far as I remember about an overweight, sick old lady who had some kind of fixation on cats. I didn’t really like the example and labelled the whole idea as ‘Google-Bashing‘.


My next encounter with this topic was when I was playing around with the Google Dashboard and was surprised to see that how precisely Google kept track of what I did and was even kind enough to tell me on which days I had been lazy, not doing much research for my blog or my thesis.

.

Then, last week I stumbled across a ‘cute’ YouTube-video, on the German Basic Thinking Blog, telling a romantic story, just by showing the queries a user had typed in.

.

Cute, as I’ve said, right? But… let’s take the idea a bit further:

I have repeatedly reported lengthy a service called TweetPsych which allows users to create kind of an psychological profile of any twitter feed, analysing the language used, the topics covered, the frequency of the posts, etc… The first worrying thing about this service is, that it works quite well. The second worrying thing is the idea of spreading this idea from a person’s twitter feed, which he deliberately had decided to publish freely on the internet, to a users search queries. Doing this we would be able not only to analyse a person’s interests but also its mood and even living habits.

.

Most of you will now say, yes that’s what Google (Google Insight) does anyway. True. But the difference is that Google, at least that’s what they’ve communicated only do this on a large scale.

Interest in the search term 'Michael Jackson'... not so much apparently until his unexpected death

.

Doing the same with just a single user takes the whole thing to a completely new level. I am not saying this because I am just another privacy-prayer hoping to get ‘street cred‘ for his words but to rephrase an idea I’ve heard from THE Austrian privacy activist (Hans Gerhard Zeger). [I know this idea isn’t entirely new, but I do think its worth being repeated many, many times…]

.

If data/information about everybody is available, authorities will start searching the data for unusual patterns to be able to investigate or even predict potentially malicious behaviour. So, the second a user types in ‘uncommon’ queries, he/she thus would be under suspicion. And here comes the point; under such circumstances, the whole principle of “presumption of innocence” (ei incumbit probatio qui dicit, non qui negat) is actually turned over. Then the authority won’t have to proof that the user has done anything wrong, but the user would be under the obligation to prove that he/she hasn’t.

.

I guess, nobody is feeling comfortable about being tracked/logged. At the same time we all appreciate the benefits of this technique. So as always a compromise has to be found as stubborn search engine bashing will just blur the whole situation and allow competitors to use the confusion and find loopholes to put things into practise, big corporations are still struggling to be allowed. Example? While some 70 year old Austrian even attacked an Google Streeview car last week with a pick-axe, nobody seems to care that an Austrian company (Herold Straßentour) has already recorded most of Vienna’s inner city, using pretty much the same technique. So shall Google be punished for at least openly speaking about their plans while others ‘just do it‘?

*Empirical Research on the User’s Perspective of Keyword Advertising

skills:age

When speaking about the essence of the Keyword Advertising dispute from a trademark as well as from a law of unfair competition point of view, 3 central questions arise. These were also explicitly listed by GA Maduro in his opinion on the three pending French Keyword Advertising trademark infringement cases:

i) Whether the use takes place within the course of trade.
ii) Whether the use is made in relation to goods or services which are identical or similar to those covered by the trade mark
iii) Whether the use affects or is liable to affect the essential function of the trade mark, by reason of a  likelihood of confusion on the part of the public

So, let’s focus on the last criteria/question: the likelihood of confusion. As I am not aware of any court decision really dealing with this question on an empirical level, I turned to the current literature on the topic and read that:

The important question then is: would the consumer be confused into thinking the trade mark owner was the origin of the goods advertised in the sponsored links? This is a difficult question to answer in the absence of empirical research. [1]

Although I truly believe that the chapter of Bednarz/Waelde is (apart from a small issue concerning the placement of an ad on the SERP) the best and most comprehensive piece I’ve ever read on this topic, I find it very difficult to agree with them on the point that there is empirical research available which could serve as a ground for further discussion of this issue. Having said this, there is no relevant study that I know of which explicitly answers the question if users are or will be confused when being confronted with advertisement on a Search Engine Result Page (SERP).

Here are some facts I came across while doing research for my thesis. As you will see in a second the range of topics covered is lsightly broader than the question at hand itself as I wanted to allow the reader to get an overview about not only the empirical research available but also the multitude of factors to be -possibly- taken into account then discussing this topic. Concerning the quality and reliability of the studies I am going to refer to in a second, I’d like to point out that I have serious doubts about some. Still, comparing all the studies available on the topic, there seems to be a discernable”trend” and this is exactly what I am aiming for. In case you are interested in more details, please feel free to contact me – I’d be happy to discuss this issue.


1. Overall internet savviness of users:

Internet savviness depends on two things; Age & Experience:

There seems to be a distinct correlation that younger people are doing better on the net than older ones.

skills:ageStill, the most important factor concerning internet is simply experience (time/experience online). Experience is far more important than gender, money, eduction, social background and family situation… It just seems to be important that users have been online for a while and made their “own” online experiences. [3]

Something else to be kept in mind is that older and younger people use the internet in different ways.  Younger users use more IM etc, middle age eCommerce and old users email. [4 ]

In relation to their internet search capabilities, users regularly tend to overestimate themselves (the paradox of the active user). [5]

2. Trust in Search Engines

Users place an impressive amount of trust in search engines and the ranking of their search results and as a result search engines are attaining the status & trust of public institutions, such as the government, hospitals etc. As a consequence users blame themselves, not the search engine if they are unable to find something on the internet. (Even if some tricky scientists have actually reversed the order of the search results.)

However, users are very biased towards ads on SERPs, but if a company is ranked within the top search results AND its ads are displayed next to the search results at the same time, the likelihood of a user transaction is increased and the brand displayed in the search result and the ad are remembered better. This of course raises questions as to whether the search engine benefits from ranking the website of the advertiser and its ad very high, as users apparently see Google’ ranking as a kind of “neutral opinion” and are thus more likely to carry out an transaction. However, Google strongly denies that it favours websites in its ranking which also display ads through it Adwords service. Nevertheless there is a fertile market for conspiracy theories…

Although “people just ain’t no good“, I am of the opinion that Google most probably does not generally accept money to change the ranking of search results (Keyword Buying), as such a practise would, if proven before the public, cause an immense uproar among users and trolls and users might shift to other search engines. Losing users and market-share would then of course also mean losing advertising revenues. Accordingly, I don’t believe that manipulating search results on a large scale would prove to be a good and sustainable business practice.  [6.]


3. How many people know about Keyword Advertising:

According to Fallows (2004), only 38% of the users actively knew about the existence of paid ads on search engines. Only about 16% of users said that they were sure they could tell the difference between paid advertisements and the search result. [7]

Other studies come up with pretty much exactly the same percentages, but point out that knowledge increases with internet experience. Whilst only 24% of people who have used the web for less than 6 month knew about the existence of paid ads, 46% of users who have been online fore more than 2 yrs were aware of the phenomenon. The highest degree of knowledge, 63%,  was however found among college students who said that they use the internet more than once a day. [8]


4. Knowledge about How Search Engines finance themselves:

User knowledge about how SEs finance themselves is very, very poor. Some authors found that less than 10% of users know that SEs are financed through advertising. Still, most users just don’t really seem to care HOW search engines finance themselves. [9]


5. Do Users notice the Ads?

Yes, but not consciously, as apparently at an early stage of their internet experience users learn that most of the ads displayed are not relevant for them and thus ignore them. A German study has found that only 5,4% of all “fixations” (fixation: the act of not only looking at an object, but also looking at it long enough to consciously process the object) are targeted at ads. This percentage was higher for transactional/commercial searches. When measuring attention in clicks, during a first search the first three search results were clicked more often than the ads, while for an extensive search with multiple searches the ads were clicked on less often than the fifth search result. [10]

Other studies find that only about 15% of all clicks are on the ads and most people seemed to ignore the ads. In more than one third of all searches there is no click on any ad. [11]


6. How relevant are ads -objectively- for internet searchers?

Users are biased towards ads on SERPs and tend to rate ads as being less relevant for their search then they objectively are. Studies have shows that more than 77% of all users favour organic links more than ads. [12]

Still, when being asked how relevant the sites were to which the ads have let the users, the users rated these websites to be as relevant as the links from the organic search results. So in fact ads prove de facto to be as relevant for searchers as (organic) results.

This result truly surprised me, but a factor that should be taken into account in this respect is that Google has stipulated stringent guidelines about how a landing page (that is to say, the website you navigate to after clicking on the link in an ad) has to look like (Landing Page and Site Quality, e.g. the product has to be  easily identifiable and the price has to be clearly stated etc.)  So it seems that the rise of quality/relevance for the user through these Google Adwords guidelines even compensates for the fact that advertisers can influence the ranking of an ad by changing their bids on the keyword.

Summing up the findings of Jansen/Spink, users are unjustifiably biased towards ads. However, if ads were not marked as ads users would notice no difference. Therefore, so as to maximise search efficiency the authors suggest that the separation between ads and search results should be abolished, as labelling the ads as advertisements triggers bias in a user’s brain and as a consequence hinders the user from performing more efficient searches. Furthermore, the authors point out the fact that ads are kind of an insurance against the users’ “blind trust” in the ranking of search engines (primacy effect, *Does Google’s AdWords Trademark Policy impair Freedom of Speech?) [13]


CONCLUSION: Having written and  cited all the studies above I’d like to point out again that there is no simple explicit study relating to the question of the likeliness of confusion – insofar I agree with Bednarz/Waelde. Still, I think the German Supreme court and the AG might have been a bit too quick when stating that users will under no circumstances establish a link between the ads and the search term/brand entered. However, on the other hand the other extreme view as brought forward by  e.g. French Courts is outdated and most probably not correct either.

In my opinion the truth lies somewhere in between and I most agree in this respect with the District Court of Massachusetts, which brought forward  a 7 factor test for checking for a likeliness of confusion, which allows taking into account factors like the the web-savviness and sophistication of the plaintiff’s potential customers as well as the content of the search results webpage that was displayed, including the content of the advertisement [Top-Ad, Side-Ad; Adv+, Adv-], etc… [14]

Endnotes:
1: Bednarz/Waelde, Search Engines, Keyword Advertising and Trade Marks: Fair Innovation or Free Riding? in Edwards/Waelde (Pub), Law on the Internet (2009) p.304
2:Hargittai, Second-Level Digital Divide: Differences in People’s Online Skills (2002) p. 12
3: Howard/Massanari, Learning to search and searching to learn: Learning to search and searching to learn: Income, education, and experience online (2007); Hargittai, The social, political, economic, and cultural dimensions of search engines: An introduction (2007)
4: Eimeren/Frees, Internetverbreitung: Größter Zuwachs bei Silver-Surfern (2008) p. 334; Fox, Generations Online December 2005, PEW Intertnet & Amercan Life Project (2005)
5: Fallows, Search Engine Users: Internet searchers are confident, satisfied and trusting – but they are also unaware and naive (2004) p. 12; Hargittai, The Role of Expertise in Navigating Links of Influence, in Turow/Tsui (Hrsg), The Hyperlinked Society: Questioning Connections in the Digital Age (2008) p.  11; Spink/Jansen/Wolfram/Saracevic, From e-sex to e-commerce: Web search changes (2002) p. 108
6: Fallows, Search Engine Users – Internet searchers are confident, satisfied and thrusting – but they are also unaware and naive (2004) p. 27; Fries, Suchverhalten im Internet (2007) p. 72; Pan/Hembrooke/Joachims/Lorigo/Gay/Granka, In Google We Trust: Users’ Decision on Rank, Position, and Relevance (2007); Yang/Ghose, Analyzing the Relationship Between Organic and Sponsored Search Advertising: Positive, Negative or Zero Interdependence? (2009) p. 26; White, Search Engines: Left Side Quality versus Right Side Profits (2008) p. 29.
7: Fallows, Search Engine Users: Internet searchers are confident, satisfied and trusting – but they are also unaware and naive (2004) p. 17; Hargittai, The Role of Expertise in Navigating Links of Influence, in Turow/Tsui (Hrsg), The Hyperlinked Society: Questioning Connections in the Digital Age (2008) p.  11
8: Princeton Survey Research Associates, A Matter of Trust: what Users Want from Web Sites (2002) p.  17; Hargittai, The Role of Expertise in Navigating Links of Influence, in Turow/Tsui (Hrsg), The Hyperlinked Society: Questioning Connections in the Digital Age (2008) p.  12
9: Machill/Welp (Pub), Wegweiser im Netz (2003) 190; iCrossing: How America Searches (2007) p. 12.
10: Erlhofer, Informationssuche im World Wide Web: taktiken und Strategien bei der Nutzung von Suchmaschinen (2007) p. 125, pp. 160
11: Jansen/Spink, Investigating customer click through behaviour with integrated sponsored and nonsponsored results (2009) p. 74, 90; Höchstötter/Lewandowski, What Users See – Structures in Search Engine Results Pages (2009) p. 12; Jansen/Resnick, An Examination of Searcher’s Perceptions of Nonsponsored and Sponsored Links during Ecommerce Web Searching (2006) p. 1958
12: Hotchkiss/Garrison/Jensen, Search Engine Use in North America (2005); Greenspan, Searching for Balance (2004);
13: Jansen/Resnick, An Examination of Searcher’s Perceptions of Nonsponsored and Sponsored Links during Ecommerce Web Searching (2006) pp. 1958
14: Schubert/Ott, Adwords – Schutz für die Werbefunktion einer Marke?, MarkenR 2009, 341.

This Satelite Doesn’t Beep But It ‘Tweets’

Please click here if you want to follow this blog on Twitter.

Enter your email address to subscribe to this blog and receive notifications of new posts by email.

Join 94 other followers

Author’s Rights

Stopline.at

Stopline.at - Online reporting hotline for child pornography and nationalsocialist content on the internet
JuraBlogs - Die Welt juristischer Blogs
Herdict.org

Previous Posts:

RSS Goldman’s Tech & Marketing Blog

  • An error has occurred; the feed is probably down. Try again later.

RSS Class 46 Blog

  • An error has occurred; the feed is probably down. Try again later.

RSS WIRED Epicenter

  • An error has occurred; the feed is probably down. Try again later.
wordpress stat