Abstract
Reported literature in sponsored search advertising markets asserts that at equilibrium an advertiser has no incentive to swap her position with another advertiser and her bid on a keyword would be bound with the click value acting as an upper bound. We investigate a closer-to-practice case where advertisers do not have an ex-ante known value per click and her bid on a keyword is an outcome of simple cost-cap heuristics on a portfolio of keywords. Using simulations and an experimental setup containing advertisers that have the same upper-cap on cost, we show that the distribution of advertisers’ cost per click and bids are emergent in nature. Keywords exhibit ex post heterogeneity in observed valuation even when all advertisers bid under the same cost-cap constraint. We explore the dynamics of the market, such as temporal stability of advertiser’s bids, and advertiser’s rank based on the click-share along with the distribution of ex post valuation of keywords associated with this closer to practice setup. The results call for a richer understanding of these markets that can incorporate temporal interdependence between auctions of a keyword as well as boundedly rational behavior of advertisers working under imperfect information.
Similar content being viewed by others
Notes
Data is part of sample data shared by an advertising agency (for a research project) and is suitably anonymized without changing the essential characteristics.
Advertisers do not have an individual value per click, rather, an aggregate value defined for the portfolio of keywords.
It is commonly assumed that the click through rate decays monotonically and by same factor as we move down the ad slots (with slot 1 referring to the slot at the top of the page) for all advertisers (Gonen and Pavlov 2007).
Target_CPC may be visualized as the maximum value that an advertiser gets from a click. Therefore, an advertiser would like to keep her cost below this value.
Academic (Animesh et al. 2011) and independent research on the Internet (Tabeling 2014) suggests that click through rate decreases as an ad is placed lower down the page. This is also consistent with the heat-map studies conducted using eye tracking movements on the search engine results page (Sandberg 2012).
The parameter was selected to ensure a sufficiently wide range of values while maintaining the exponential drop (Chaffey 2012) in CTR with position. The exponential drop in CTR was identified by empirically examining data obtained from an advertising agency and other independent analysis available on the Internet. While the advertising agency’s data gathered as a part of this research indicated that the value of the parameter varied between 0.1 and 0.4, independent research on the internet reported the value to be approximately 0.2.
The reason we selected 20 advertisers was based on the value of the parameter which controls the decrease in probability as a function of location. The value of 0.17 selected for this simulation physically means that the probability of clicks on the last position (20 here) is only 0.64%. With 30 advertisers this probability decreases to 0.12%. Thus, there is only marginal effect of additional advertisers.
See Levene (1960).
The three categories comprise three equal divisions of the range of values observed. For example, the minimum average CPC was about 7.2 and the maximum about 9.7. We divided the range [7.2–9.7] into three equal categories. The standard deviation of the observed heterogeneity was about 0.42, therefore, the three categories roughly correspond to low = (mean \(-~2~*\) std. deviation, mean), medium = (mean, mean \(+~2~*\) std. deviation), and high = (mean \(+~2~*\) std. deviation, maximum).
References
Abrams, Z., Ghosh, A., & Vee, E. (2007). Cost of conciseness in sponsored search auctions. In Internet and network economics (pp. 326–334). Berlin: Springer. Retrieved from http://link.springer.com/chapter/10.1007/978-3-540-77105-0_34.
Al-Najjar, N. I., & Pai, M. M. (2014). Coarse decision making and overfitting. Journal of Economic Theory, 150(1), 467–486. doi:10.1016/j.jet.2013.12.003.
Animesh, A., Viswanathan, S., & Agarwal, R. (2011). Competing “creatively” in sponsored search markets: The effect of rank, differentiation strategy, and competition on performance. Information Systems Research, 22(1), 153–169. doi:10.1287/isre.1090.0254.
Borgs, C., Chayes, J., Immorlica, N., Jain, K., Etesami, O., & Mahdian, M. (2007). Dynamics of bid optimization in online advertisement auctions. In Proceedings of the 16th international conference on World Wide Web—WWW’07 (pp. 531–540). New York: ACM Press. doi:10.1145/1242572.1242644.
Cary, M., Das, A., Edelman, B., Giotis, I., Heimerl, K., Karlin, A. R. et al. (2007). Greedy bidding strategies for keyword auctions. In Proceedings of the 8th ACM conference on electronic commerce–EC’07 (p. 262). doi:10.1145/1250910.1250949.
Chaffey, D. (2012). Comparison of google click through rates by position. Retrieved 14 July 2016 from http://www.smartinsights.com/search-engine-optimisation-seo/seo-analytics/comparison-of-googleclickthrough-rates-by-position/.
Dhar, V., & Ghose, A. (2010). Research commentary–sponsored search and market efficiency. Information Systems Research, 21(4), 760–772. doi:10.1287/isre.1100.0315.
Dugundji, E. R., & Gulyás, L. (2008). Sociodynamic discrete choice on networks in space: Impacts of agent heterogeneity on emergent outcomes. Environment and Planning B: Planning and Design, 35(6), 1028–1054. doi:10.1068/b33021t.
Edelman, B., & Ostrovsky, M. (2007). Strategic bidder behavior in sponsored search auctions. Decision Support Systems, 43(1), 192–198. doi:10.1016/j.dss.2006.08.008.
Edelman, B., Ostrovsky, M., & Schwarz, M. (2007). Internet advertising and the generalized second-price auction: Selling billions of dollars. The American Economic Review, 97(1), 242–259.
Feldman, J., Muthukrishnan, S., Pal, M., & Stein, C. (2007). Budget optimization in search-based advertising auctions. In Proceedings of the 8th ACM conference on electronic commerce—EC’07 (p. 40). doi:10.1145/1250910.1250917.
Filatova, T., Van Der Veen, A., & Parker, D. C. (2009). Land market interactions between heterogeneous agents in a heterogeneous landscape–tracing the macro-scale effects of individual trade-offs between environmental amenities and disamenities. Canadian Journal of Agricultural Economics, 57(4), 431–457. doi:10.1111/j.1744-7976.2009.01164.x.
Fleder, D., Hosanagar, K., & Buja, A. (2010). Recommender systems and their effects on consumers: the fragmentation debate. In: EC. Retrieved from http://archive.nyu.edu/handle/2451/29485.
Goldstein, J. (1999). Emergence as a construct: History and issues. Emergence, 1, 19–72. Retrieved from http://www.tandfonline.com/doi/abs/10.1207/s15327000em0101_4.
Gonen, R., & Pavlov, E. (2007). An incentive-compatible multi-armed bandit mechanism. In Proceedings of the twenty-sixth annual ACM symposium on principles of distributed computing—PODC’07. doi:10.1145/1281100.1281174.
IAB. (2014). IAB internet advertising revenue report 2013.
Katona, Z., & Sarvary, M. (2010). The race for sponsored links: Bidding patterns for search advertising. Marketing Science, 29(2), 199–215. doi:10.1287/mksc.1090.0517.
Krause, G. A. (1996). Agent heterogeneity and consensual decision making on the federal open market committee. Public Choice, 88(1), 83–101.
Lambert, N., & Shoham, Y. (2007). Asymptotically optimal repeated auctions for sponsored search. In Proceedings of the ninth international conference on electronic commerce (pp. 55–64). ACM. Retrieved from http://dl.acm.org/citation.cfm?id=1282112.
Levene, H. (1960). In I. Olkin et al. (Ed.), Contributions to probability and statistics: Essays in honor of Harold Hotelling (pp. 278–292). Stanford: Stanford University Press.
Liang, L., & Qi, Q. (2007). Cooperative or vindictive: bidding strategies in sponsored search auction. In Internet and network economics (pp. 167–178). Berlin: Springer. Retrieved from http://link.springer.com/chapter/10.1007/978-3-540-77105-0_18.
Maeda, T., Matsumoto, S., & Murata, T. (2015). Agent heterogeneity and facility congestion. Computational Economics, 46(2), 189–203. doi:10.1007/s10614-014-9451-x.
Maillé, P., Markakis, E., Naldi, M., Stamoulis, G. D., & Tuffin, B. (2012). Sponsored search auctions: An overview of research with emphasis on game theoretic aspects. Electronic Commerce Research, 12(3), 265–300. doi:10.1007/s10660-012-9094-8.
Salganik, M. J., Dodds, P. S., & Watts, D. J. (2006). Experimental study of inequality and unpredictability in an artificial cultural market. Science (New York, N.Y.), 311(5762), 854–856. doi:10.1126/science.1121066.
Sandberg, H. (2012). Google heat maps and SEO. In SeoCustomer. Retrieved from http://seocustomer.com/google-heat-maps-and-seo/.
Shah, S. (2009). How important is click through rate in Google’s quality score formula? In Search engine land.
Simon, H. (1972). Theories of bounded rationality. In Decision and organization (pp. 161–176).
Tabeling, J. (2014). PPC Click-through rate by position: Does rank matter? In Search engine watch. Retrieved October 12, 2015, from http://searchenginewatch.com/sew/study/2345638/ppc-click-through-rate-by-position-does-rank-matter-data.
Varian, H. R. (2007). Position auctions. International Journal of Industrial Organization, 25(6), 1163–1178. doi:10.1016/j.ijindorg.2006.10.002.
Vickrey, W. (1961). Counterspeculation, auctions, and competitive sealed tenders. Journal of Finance, 16(1), 8–37. doi:10.2307/2977633.
Yuan, Y., Zeng, D., Zhao, H., & Li, L. (2014). Analyzing positioning strategies in sponsored search auctions under CTR-based quality scoring. In: IEEE transactions on systems, man, and cybernetics?: Systems (pp. 1–14). Retrieved from http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=6960033.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Gupta, A., Saha, B. & Sarkar, U.K. Emergent Heterogeneity in Keyword Valuation in Sponsored Search Markets: A Closer-to-Practice Perspective. Comput Econ 50, 687–710 (2017). https://doi.org/10.1007/s10614-016-9637-5
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10614-016-9637-5