Topic 4 - Search Engines

Question 1

Describe search engines as multi sided platforms

Answer 1

Users -free-> search engine <- advertisers

Question 2

What are the factors that explain the concentration in the search engine market

Answer 2

Economies of scale - high fixed cost for infrastructure (economies of sale in the design of algorithms), indirect network effects - larger the number of users = more attractive for advertisers, switching costs

Question 3

Describe switching costs on SEs

Answer 3

Users - easy, personalisation?
Advertisers - incentive to advertise on a small platform

Question 4

Pagerank and how it works

Answer 4

How many important sires link to a given site
Works by counting the number and quality of links to a page to determine a rough estimate of how important the website is
— more important sites will be likely to receive more links from other sites

Question 5

GoTo.com business model

Answer 5

Advertisers bid for first placement on the site
Paid per click

Question 6

Basic overview of google ad auctions

Answer 6

Everytime someone types a query, auction is run which determines which advertisers get up to 11 sponsored links on every results page - google gets 98% of the profits
Grows with internet usage
Each keyword - advertiser submits bid - if high enough.- add will be shown

Question 7

Google ad auction steps

Answer 7

Allocate positions among advertisers
Price the positions

Question 8

Explain googles position allocation scheme

Answer 8

Allocation scheme determines position of ad on results page
Calculates ad rank - bid x ad quality score
Bid is submitted by advertisers fir keywords relevant to users search
Ad quality score - measure of relevance of ad
Google has interest in ad quality score
Advertisers prefer highest ad rank - higher prob of clicks (maximise the clicks)

Question 9

Describe googles pricing scheme (ad auctions)

Answer 9

Google calculates cost per click in each position
In position 1, solve: p1 * ad quality = b2 * ad quality
Price to maintain ad position: p1 = b2* ad quality2/ ad quality 1

Question 10

Summarize the pricing of ad positions

Answer 10

Advertiser submits list of keywords
User entered a query, google compiles list of words that will match the query
List of ads is ordered based on bid and ad quality - measure of relevance to the user
Highest ranked add is ranked and showed first and so follows
If user clicks - advertiser is charged a price that depends on the bid and the quality of the ad of the advertiser below him

Question 11

Discuss whether SEs should show more ads

Answer 11

Benefits - more revenue
Costs - less competition between advertisers (lower ad price), relevance goes down (less users click)

Optimal choice is a trade off between short run profits and long term run goals

Question 12

What is SE bias

Answer 12

Listing of search results in the SEs best interest
Platform manipulates search results strategically to promote its own interests

Question 13

Can a search engine influence consumer behaviour by biasing search results?

Question 14

what incentives are there for SEs to bias search results

Question 15

Discuss impact of ranking on clicks

Answer 15

Links at top get more clicks
Correlation or causation
-50% and up up -85% traffic between first and 10th rank

Question 16

incentives to bias search results

Answer 16

Diverting customers toward more profitable results
Increasing the value of paid links (thus bids)
Favouring its own services of vertical integration

Question 17

Bias toward sponsored links - they used to be placed on the side out of the golden triangle (usual eye tracking - in later years study)

Question 18

Bias towards sponsored links (Taylor 2013)

Answer 18

Displays organic results and sponsored content
Derives revenue from paid links
Organic - no revenue
— SEs can induce clicks on more paid links by making organic links less relevant

Incentives to decrease quality of organic links
— positive - reduces cannibalization of paid links by organic links
— negative - lowers quality of search engine thus search activity

Trade off between activity and revenues

Question 19

Example of revenue-activity trade offs by amazon, booking.com, Spotify

Answer 19

Amazon - diverting to products with higher margins
Booking - diverting to hotels that don’t offer lower prices outside of booking.com
Spotify - diverting to songs with lower margins

Question 20

Describe the idea and trade offs of relaxing competition between advertisers

Answer 20

Search engine obtains profits from advertisers - interest that valuation for clicks is high
Advertisers have higher valuation of clicks if competition is soft (fewer relevant links)
More product market competition - lower prices because lower chance of clicks
Search engines have incentive to manipulate search engine results to relax product market competition

Question 21

Describe bias due to vertical integration

Answer 21

Search engine links to publisher based on - consumer taste and amount of advertising
If one is a stronger competitor - biased against this publisher
Integration can increase or decrease bias
- increase revenue from own publisher
Less bias -`> higher quality results -> larger marginal revenues from increasing total demand

Question 22

Main complaints of the European Commission against google

Answer 22

Search bias - favouring own services
Violation of IP rights
Contraction restrictions for web operators (exclusivity and non transferability of advertisement campaigns)
Contractual restrictions for Android manufacturers

Question 23

What makes the case of the Euro commission against google so tricky?

Answer 23

Market dominance - what is the dominant market when searching
General search - google is dominant
Does the relevant market include vertical search engines e.g., amazon - google is not dominant in places like this

Abuse of a dominan

Question 24

Why is search discrimination an abuse of a dominant position

Question 25

Positional remedies to googles abuse of their dominant position

Answer 25

Mandate search neutrality
Unbundle googles search business
Open algorithm
Regulate algorithm
Share historical data