i290-7 Fall 08

We begun the discussion with understanding what it means for a consumer to be ‘better’ served in 2-sided markets – where ‘better’ could take several forms, including improved service and cheap price. In addition, if platform provides more functions, a customer would have more benefits from a platform.

The first mover (like EA) to video game console may require high costs. But, once it gets lower, more players get in. Because a game console and a game maker sometimes make multi-year contracts, it prevents a game maker from moving to another game console, which in turn causes high switching cost to another game console.  This was compared to the DVD market and found that video content makers don’t have to take care of multiple technology platforms. So, it allows them to reduce cost in producing and distributing video contents through the DVD platform.

Consumers will find it convenient to use mobile phones, video, internet access, etc. through one single device – i.e. convergence. In addition, the platform providers can take bundling strategy by selling some service together with cheaper price. It was wondered if Apple’s offering via iPhone, which bundles various services like telephony, internet access, music, applications etc. is a conglomeration platform envelopment attack on adjacent markets of handset manufacturers like Samsung or Nokia, which offered these services individually.

We then had an interesting discussion comparing Apple’s iPhone and Google’s Android. While iPhone includes its own hardware and software platform, Android is an operating system which is independent of a mobile phone device. Metaphorically, it is a good example to look back on Apple’s Mac and Microsoft’s Windows. In this game, MS dominated the related markets over Apple.

We then moved on to talk about the mobile phone market – which is a 2-sided market in which the platform is the service carrier and the two sides are the handset manufacturer and the consumer.  In this market, the same-side effects for consumers are positive, as more users on the network often lead to increased investment in the infrastructure of the network.  Same-side effects are negative for handset manufacturers who would rather have fewer other manufacturers to compete with.  Cross-side effects for both the consumers and the manufacturers are strongly positive.  Consumers would like more manufacturers from which to choose a handset, and manufacturers would like more consumers to whom they can sell handsets.  In this market, the network subsidizes the consumers, charging high service fees and giving handsets away for free or with deep discounts.

The above was then discussed analogously with Better Place – a VC backed startup developing an electric car charging system to work with electric cars manufacturers. This will create a 2-sided market similar to the mobile phone market in which Better Place is the network provider (of batteries and charging facilities) and the two sides will be the electric car manufacturer and the consumer/driver. This market features similar network effects – same side effects are positive for consumers and negative for automakers; cross-side effects are positive for both consumers and automakers.  In the same way the network providers subsidize the mobile phone consumer by providing cheap handsets and expensive network service, the scheme proposed by Better Place features the network provider subsidizing the automobile consumer by providing cheap automobiles and expensive recharging services.

As this potential new market (electric cars) emerges, it will face competition with the already existing and powerful gasoline-fueled automobile market.  It will have to compete with both gasoline suppliers and gasoline-fueled automobile manufacturers.  Assuming a case in which electric cars are given away, we considered the options gasoline suppliers would face.

Gasoline suppliers could begin to offer electric recharging in addition to gasoline refueling.  This would be a conflict of interest, so companies might be very slow to adopt.  Additionally, it is not clear that electric power will be the eventual leader in gasoline-alternative fueling.  Gasoline companies might wait to see which alternative(s) will be most common and lucrative and slowly begin to introduce those.

Better Place is introducing its electric car products (batteries, charging spots, battery switching stations, and software) in pilot programs in Israel and in California.  The recharging stations will use renewable energy sources.  We will watch this market unfold over the coming years.

We then moved on to the final discussion comparing Facebook (FB) and iPhone’s platforms. We soon established that the 2 sides for both these networks are consumers and application developers. FB is a traditional 2-sided platform with consumers and lots of apps whereas Apple also makes money by selling these apps to the consumers. While for FB, advertisers subsidize the cost for the app developers, for Apple, consumers pay directly for the app.

The question of what would happen if FB starts charging users an app fee was discussed next. It was noted that FB shouldn’t really charge the consumers as they would be too elastic and may not pay, but then it was observed that they should possibly charge the app developers instead. In fact, FB has started charging the app developers an app fee, to ensure that the quality of the application is good. This point of ‘charging the supply side (app developers) instead of the demand side (users) to ensure the quality of an application’ was discussed in detail and it was noted that it leads to weeding out the lousy applications. This was advocated as a strategy that FB could focus on. Another interesting recommendation for FB was to get users to pay for hard-to-get apps.  Another could be in the content discovery for apps – place the high quality ones higher up than the lower quality ones.

The discussion was then opened to other strategies in the 2-sided markets that are useful to study. eBay was considered to be a good example of a marketplace for buyers and sellers. The question of how Google is using its platform as a market was raised next. Google’s search engine connects eyeballs to advertisers. Google has recently added the feature of letting users promote or remove search results. It is not very clear how they will use these results. It was noted that this may affect page rank and may not have an affect on search results.

All in all, a very fascinating discussion on platform competition, two-sided networks and platform envelopment.

We began with a discussion of the different schemes used to price online ads, focusing primarily on cost-per-click (CPC) and cost-per-action (CPA). The various pros and cons of CPC and CPA were discussed, with Professor Chuang noting that CPA schemes address the click-fraud concerns associated with CPC. We observed that CPC schemes seem to be more favorable (as compared to CPA) to the search engine selling ad space. An interesting question was raised as to what constitutes an “action” aside from the purchase of a good or service. Suggested answers included signing up for a mailing list or providing personal information. Professor Chuang also noted some retailers’ use of special phone numbers tied to Internet advertisements as a way to measure online marketing effectiveness. However, because of the difficulty in capturing, or simple lack of, actions in many contexts, it was suggested that the industry is unlikely to ever become CPA dominated. Rather, CPA is likely to continue to gain popularity in arenas where actions are easily defined and recorded.

We next looked at an example of different ads used by Google on http://searchenginewatch.com/3631223 to see sample data on two and how they had fared. We asked the class to vote on which ad they thought would have the highest conversion rate. The results appeared counter-intuitive and we discussed why this was so. It seemed some people in the class did not find the polling conclusive, due to the fact that the examples used variants on generic search terms like green or industrial widgets. Still it was interesting to point out that the context where the ad itself was placed could influence the outcome.

Our next question dealt with the limitations of limited-time dynamic auctions where participants also maintain a fixed budget. The complexity of this pricing mechanism could explain why most search engines haven’t adopted it, in favor of simpler, yet often less rewarding schemes. We then jumped forward to discussing dynamic allocation, but this is an open research question with no specific answers. There does not exist a concrete dominant strategy.

On the topic of more individually targeting ads, we considered specialized search engines and whether or not they would play a role in the advancement of online search. The conversation then focused on the collection of personal information and viewing habits. No one seemed to think that privacy issues would overly constrain this tactic of targeting advertisements. Some type of incentivized “opt in” system that would collect personal browsing habits and better target ads was suggested as one approach to address privacy concerns. It was also noted that only short-term viewing habits are needed to effectively target ads.

QUESTION 1. We have read about how companies can succeed in 2-sided markets.  Let’s consider how consumers are served in these markets.  Are consumers better served by 2-sided markets than 1-sided?

From the Renesys blog: http://www.renesys.com/blog/2008/11/will-work-for-bandwidth.shtml

… Look to the wholesale carriers if you want to see an income statement wasteland. Level 3 lost $1.1b last year. They lost $120m in the most recent quarter alone. Cogent is thrilled because they reported a tiny, tiny positive net income last quarter on top of a yearly loss of $30m in 2007. Global crossing lost $300m in 2007 and $88m in the last quarter they’re reporting, which doesn’t include much of the recent downturn. Other wholesale networks are in the same boat. …

… Commodity markets are good for most of us. They provide necessary goods at a great price. They provide strong and direct incentives for companies to align investment with market demand, and keep costs down for everyone. Commodity markets tend to misbehave when some number of players can afford to lose money for very long amounts of time. This keeps prices below costs and has the potential to ultimately raise prices (and margins) significantly, provided many competitors go out of business. I believe that this is essentially what several of the low-price leaders in the marketplace are hoping will happen: that they can drive several competitors out of business with unsustainably low costs and afterwards can capture significantly more market share and make money through a combination of economies of scale and raising prices. It’s not a bad strategy provided they have a lot more money than everyone else, or a nearly infinite ability to lose investors’ money on their way to profitability. But it’s risky. …

… There may be another way forward here: quality differentiation. … In addition to competing on quality, there’s another way that IP networks may choose to compete: differentiated cost. Right now everyone pays per bit per second thresholded in some way regardless of where those bits come from or are going to. This billing model completely ignores the reality that carrying 1 Mb/s from Los Angeles to Karachi is dramatically more expensive than carrying that same traffic flow to San Jose. People who are carrying cheaper traffic are subsidizing those who have more expensive traffic, because the market has always worked that way and carriers don’t have any good way of accounting for the difference. …

Question 1:
As described in the readings, the process of allocating online advertisement space has evolved considerably in the last decade.  In addition to changes in determining how ads get placed (i.e. changes to the bidding process), the pricing scheme has also been modified.  Payment-for-impression has been largely replaced by cost-per-click (CPC).  While some search engines have experimented with cost-per-action (CPA), CPC remains the dominant scheme.

• Do you believe that the industry will move to CPA from CPC?
• Are there problems with CPC that would be solved by CPA?
• Are there new issues that using CPA would introduce?
• Do you think that one scheme (CPC or CPA) tends to favor advertisers or search engines more than the other?

Question 2:
Cost-per-click has become the dominant payment scheme used by search engines.  An important aspect of this approach seems to be the belief that advertisers derive benefit not from having their ads displayed, but from attracting traffic to their sites and then selling some good or service.  This raises the question: are there still categories of advertisers who may gain a great deal from impressions alone?  Consider the film industry — a recent google search for the phrase “new films” brings up ads that include: “‘Valkyrie’ Starring Tom Cruise. In Theaters December 26″ and “BOLT – New Disney Comedy. Opens Friday.”

• Is it possible for this type of advertiser to game the system?
• Or is it reasonable to think that the “weighting factor” (the estimated click-through-rate) accounts for these type of ads?  That is, are advertisers who fall into this category simply forced to internalize the publicity benefits of their ads into their pay-for-click bids?

Question 3:
Lahaie, et al. note that there may be a need for richer bidding models for search engine advertising.  Specifically, “ones that might allow bidders to express… distinct values for traffic from certain geographic regions, demographic profiles, etc.”

• What are some techniques that search engines could use to provide more targeted advertising opportunities? What are the costs and benefits to these approaches?  Are they likely to be accepted by consumers?
• What about the use of more specialized search engines as a way to perform market segmentation?  Consider, for example, Google Scholar.  While this search engine currently does not use advertising, could knowledge of its user base be leveraged to better target ads towards certain demographic groups?

Question 4:
How can we explain why the VCG mechanism, although a means to maximize revenue while ensuring each agent reports its true value, still hasn’t been adopted by most search engines? Is there some network effect due to the nature of web 2.0 that makes it prohibitively costly for non-dominant search engines to handle VCG pricing? Furthermore, how do the effects of the long tail and its attendant proliferation of niche markets erode the monopoly of search engines as a main gateway for advertisers to maximize profit from?

Question 5:
A blog post on ecoIron elicited an interesting discussion on the economic incentives for Google to move its high volume site to a green computing format (the likes of Blackle), thus potentially saving  750 Megawatt Hours a year. The argument, based on the large number of queries ran on Google each day (over 200 million), and the average amount of time each result is displayed (10 seconds), stated that the search giant could save a good amount of energy and dollars for changing a few color codes. Following on the question above, since the value of a user to an advertiser depends on how likely he or she is to buy, not how many users there are, do you think it could be economically viable for Google to follow this strategy? Why or why not? Given that a growing number of niche users might be impressed by this environment-friendly search, what would the ramifications be for Google in terms of modifying its Ad Auction mechanism to reach and serve this potential user base?

Question 6:
Dynamic online allocation poses interesting challenges based on the maximum budget available to advertisers for a period of time and their willingness to pay per impression. Both the greedy and budget smoothing methods of assigning slots to advertisers fall short in maximizing revenue. Focusing on the budget constraints feature of these auctions, is there an alternate mechanism that would help maximize social surplus by not relying solely on user clicks? How about establishing an incentive scheme to have advertisers defer the timing of their ad impressions in exchange for auction credit or display at a later time? (for example, like when airlines have passengers exchange tickets for flying vouchers or seats on later flights through some type of matching/mediation mechanism to ensure user utility is maxed)

Question 7:

Varian says, “The answer, at least in my opinion, is a much older economic concept called ‘learning by doing’ that was first formalized by Nobel Laureate Kenneth Arrow back in 1962. It refers to the widely-observed phenomenon that the longer a company has been doing something, the better it gets at doing it.” What implications does this have for new upcoming search engines? Can you think of real-world examples where a new player in the market has overthrown existing well-established players?

First, we discussed the asymmetric information market which exists between employers and job applicants. Various strategies to reduce this asymmetry were suggested in discussion, such as checking applicants’ Facebook pages, conducting rigorous on-site interviews, and using internships.  Professor Chuang contrasted actions initiated by the uninformed party (Screening) from actions initiated by the informed party (Signaling).  He explained that for signaling to be effective, the cost of the signal must be high (e.g. attending university).  The analogy to signaling in evolutionary biology was also described.

We next discussed the response by the National Association of Realtors (NAR) to the Freakonomics’ critique.  People did not find their response convincing for a number of reasons.  Joyce questioned whether sellers would be able to accurately evaluate their realtor’s service.  Josh was skeptical that people would have an opportunity for repeat business.  Professor Chuang related this to the iterated Prisoner’s Dilemma and pointed out that with such a small value of n, the realtor’s optimal strategy is unchanged.

Third, we talked about the effect of guarantees in markets with asymmetric information.  Contrasting government guarantees and guarantees by sellers/manufacturers, one distinction is that government regulation is related to the concept of best effort, which meant 95% of products checked by the government were good but not all of them, but companies’ guarantees such as refunds were applied for all products that the companies sold. Another distinction that was mentioned was that government guarantees generally certify production processes. When investigating/evaluating is costly, or requires special skills, there is a demand for government certification.  Josh mentioned the bizarre case of Creekstone which everyone should read about.

The final discussion topic was risk aversion.  Kesava described how risk aversion affects online behaviors and solicited suggestions of how companies might reach out to potential customers.  Two different dangers were identified, (1) dissatisfaction with the purchase and (2) fraud.  Many of the suggestions were directed at actually reducing these risks — for example by offering a money-back guarantees and return shipping (like Zappos) or by getting credit card companies to offer better protections against fraudulent charges.  Josh also suggested that people could be comforted by official looking security certifications (e.g. TRUSTe).

http://startuplessonslearned.blogspot.com/2008/11/using-adwords-to-assess-demand-for-your.html

Very timely for today’s class discussion.

Interesting forray into ‘p2p’ on the part of the ISP

http://arstechnica.com/news.ars/post/20081103-comcastic-p4p-trial-shows-80-speed-boost-for-p2p-downloads.html

Asymmetric Markets

The job market is a typical “asymmetric” information market for both employers and applicants. It is difficult for employer to find a “real” quality of an applicant until he/she start to work, and it is also difficult for an applicant to find “real” quality of a company.

(1) To reduce asymmetry of information, what strategies are adopted by each of them?

(2) How about a “marriage” market?

(3) There are some agents who are specialized in job-matching or marriage. Are they helping to reduce information asymmetry?

(4) Can internet or information technology help to reduce (or widen) these information asymmetry?

Information Asymmetry in Real Estate

The National Association of Realtors responded to the Freakonomic critique by saying:

“Levitt and Dubner assume real estate businesses are built around one-time transactions. In fact, successful professionals build relationships with customers for life. Homeowners move once every seven years on average and are likely to use an agent they have used before.”

(1) How might repeated transactions (and reputation systems) work to protect sellers?

(2) How do the effects of these compare to efforts which aim to reduce  information asymmetry?

One approach to eliminating information asymmetry is Zillow.com.  Although entusiastic about the idea, Levitt says:

“Unfortunately, I don’t think you want to put too much stock in the house valuations they come up with. The information appears to come from the assessor’s office and thus is neither particularly accurate or informative. They do also seem to be using information on actual home sale prices as well, which is a much better source.”

(3) What are the consequences of biased estimator that undervalues/overvalues homes?

(4) Why aren’t real estate appraisals used more often during home sales?

(5) If unreliable, why are appraisals required by lenders?

Regulated Markets

Some product’s/firm’s qualities in some industries are regulated or “guaranteed” by a government agency such as FDA (foods, drug), FRB (bank) or SEC (investment bank). However, some products have not.

(1) What are the differences between these?

(2) Is the government’s guarantee enough to persuade buyers to believe those products/companies quality? Is it more trustworthy/trusted than guarantees made by companies themselves?

(3) Is it possible for government to do some “moral hazard” activities? For example, can you explain one of causes of current economic turmoil by a moral hazard or a principal-agent problem, in terms of bank’s moral hazard or SEC and FRB’s omission of regulation?

(4) If you can explain (3), what should Obama to do to prevent these problems again?

Paying for Information

According to the economics of information, it is rational for us to pay money for information. However, most reputation system and informational web pages don’t receive money directly from information recipients.

(1) Can operators that run “reputation” systems or web pages filled with economically useful information collect money directly from recipients of that information? If not, why?

(2) In practice, what types of information are people willing to pay money for?

The discussion began with thoughts on how the game theory applies to information goods.  The main idea taken from the previous lecture was that there are tradeoffs between individual rationality and collective welfare.

The case of eBay removing buyer feedback was then discussed.  Currently, buyers and sellers find mutual compromise without any kind of arbitration from eBay.  The new move to remove the buyer feedback may undermine this current setup.  However, it could also foster strong trust because more relevant and truthful feedback for the sellers is given.  There is less buyer fear for seller reprisal if they give a negative response, and the buyer has nothing to gain by giving a bad rating if they got good service.

A discussion on Craigslist vs. eBay or Amazon followed where a couple of points were raised.  One point was that craigslist is largely used for local purchases – where you can meet the actual buyer and negotiate in person, while eBay or Amazon is used for non-local purchases because of the trust issue.  An analogy was also introduced, wherein eBay was like a fair or flea market where startup companies were trying to establish reputations, while craigslist was like a garage sale with individuals selling their wares.  In summary, it was decided that the absence or presence of a reputation system was only one dimension on how these two markets are different, but that it somehow contributes to the different types of sellers, buyers and goods being transacted in those markets.

In terms of people’s contributions to feedback systems, like in eBay, for instance, a very good point was raised by Ashwin, who said that even if he never plans to buy from a certain sellers again, he would still give feedback, because he wants eBay as a platform to succeed.  By contributing feedback and seller ratings, he increases trust in the system.  It was also raised that in the online realm, the cost of giving feedback is very low, thus making it easier for people to contribute.

A short discussion on asymmetrical reputation systems, as in the case of banks, was done.  An example was given involving students and professors.  How a professor is judged by the school (in terms of research and teaching skill) is different from how students judge their professors (how much they actually are able to learn), and so a single source is often not reliable enough.  A hypothesis was also presented, wherein it was stated that in an asymmetrical statement, a third party (like the government) is needed to prop the system up, but when a certain level of asymmetry is exceeded, the system fails.

As an aside, the class got into a very short discussion regarding research citations, and how researches somehow trade on reputation.  It  was concluded that the diversity of people and institutions citing a specific work was needed to get a more accurate view of a research’s
reputation, instead of just a quantitative count (which includes self-citation).  Google Scholar apparently does not have such distinction, but Citeseer does.