The impact of organic links in position auctions with popular versus niche keywords

• Published date: 01 September 2018
• Date: 01 September 2018
• Author: Thierry Penard,Michael Arnold,Eric Darmon
• DOI: http://doi.org/10.1111/ijet.12156

doi: 10.1111/ijet.12156

The impact of organic links in position auctions with popular

versus niche keywords

Michael Arnold,∗Eric Darmon†and Thierry Penard†

Weanalyze a game-theoretic model to explore the impact of organic links (links generated by the

search engine algorithm) and popular versus niche keywords on bidding strategies in sponsored

link position auctions. In contrast to results in earlier literature, we find that with organic links

the firm with the highest value per click does not necessarily win the first sponsored position; it

may be optimal for a firm to bid an amount greater than the expected value from a click; and

firms may choose not to bid even if they would incur no cost for clicks on the sponsored link.

Key wor ds searchadvertising, keyword auction, sponsored link, bidding strategy, search engine.

JEL classification D44, L86

Accepted 5 July2016

1 Introduction

Consumers often access information about firms in online markets through a commercial search

website such as Google.com, Yahoo.com or Bing.com. Sponsored search advertisingenables fir ms to

display sponsored ads above organic results produced by the search engine.1These advertisements

improvea website’s visibility, increaseits audience, and help convert visits into sales. They also provide

a balance between several concerns. Because sponsored links are displayed with organic links, from

the user’s perspective they appear less intrusive than other types of ads such as pop-up windows or

e-mail advertising. From the advertiser’s perspective, sponsoredsearch provides the ability to target

customers based on a search query which results in more qualified traffic viewing sponsored ads.

Finally, sponsored search typically entails a cost-per-click payment (also called performance-based

payment) under which advertisers only incur a charge if a consumer clicks on the sponsored link.2

In 2013, sponsored search advertisements generated over D13.4 billion in revenuefor search engines

in the European market (up 13% from 2012) and $18.4 billion in the US market (up 9% from

2012). These advertisements accounted for 43% of online advertising revenue in the USA and 49%

in Europe (Interactive Advertising Bureau, www.iab.com).

∗Department of Economics, Alfred Lerner College of Business and Economics, Universityof Delaware, Newark, Delaware,

USA. Email: marnold@udel.edu

†CREM, University of Rennes1, France.

We would like to thank the anonymous referee and the participants at the Searle Center Third Annual Conference on

Internet Search and Innovation for helpful discussions and comments.

1See Evans (2008, 2011) for a survey on the economics of online advertising. See also Dhar and Ghose (2010) for a research

agenda on sponsored search markets.

2See Taylor(2011) for an economic rationale for current payment schemes in the online advertising industry.See also Moon

and Kwon(2011) and Zhu and Wilbur (2011) for the advantages of cost-per-click versus cost-per-thousand-impressions

schemes for the advertiser and the publisher.

International Journal of Economic Theory 14 (2018) 257–277 © IAET 257

International Journal of Economic Theory

Organic links in position auctions Michael Arnold et al.

Most sponsored ads are sold through keyword auctions. Our paper explores howfir ms strategi-

cally compete in keyword auctions when customers can access a firm’s website by clicking on either

a sponsored link or an organic link appearing in the search results produced by the search engine.

Following the seminal papers of Edelman et al. (2007) and Varian (2007), a growing literature has

analyzed keyword auctions and search advertising strategies (Agarwal et al. 2006; Chen et al. 2009;

Animesh et al. 2010; Edelman and Schwartz 2010; Athey and Ellison 2011; Yao and Mela 2011;

Zhang and Feng 2011; Taylor 2013; White 2013; Ye et al. 2013; Gomes and Sweeney 2014). However,

these papers typically have not incorporated the role of organic links. Our analysis demonstrates

how the relevance of organic links impacts bidding for sponsored links. Sponsored links shift clicks

away from organic links because sponsored links are placed at the top of the searchresults page and

move organic links further down the page. This location effect may be exacerbated or offset by the

degree to which sponsored and organic links serve as substitutes or complements in the consumer’s

search experience. In particular, regardlessof location, the nature of the keyword (e.g. generic versus

brand-specific) and the text accompanying the sponsored link may affect the relationship between

sponsored and organic links. Evidence suggests that for some keywords (e.g. branded keywords),

sponsored links serve as a substitute for organic links, whereas for others, sponsored links com-

plement organic links and lead customers to click on both the firm’s sponsored and organic link.3

Consistent with actual markets, our model allows organic and sponsored links tobe either substitutes

or complements. Our analysis also distinguishes between “popular” keywords (for which demand

for sponsored positions exceeds supply,as in the existing literature) and “niche” keywords (for which

demand is not constrained).

Wefind that the presence of organic links can lead to more aggressive bidding in keywordauctions

if sponsored and organic links are sufficiently strong complements. We also show that the firm which

is most relevant or has the highest value per click does not necessarily win the first spot in the

sponsored search listings. Under certain conditions, a less popular firm may use a sponsored link

to increase its traffic while a more popular firm relies only on its organic links to attract customers.

Our results imply that equilibrium outcomes for a given keyword can generate no sponsored links,

a single sponsored link, or multiple sponsored links. The extent to which sponsored and organic

links are complements or substitutes along with location effects also creates an important role for

the minimum cost per click (CPC) established by the search engine, particularly in the case of niche

keywords. Most of these results are not consideredin earlier models that do not incorporate organic

links.

Our analysis is closest in spirit to Katona and Sarvary (2010) and Xu et al. (2012). Katona and

Sarvary (2010) show that under certain conditions a less relevant firm may outbid more relevant

firms to win the top position in the sponsored listings. In contrast to their analysis in which all firms

participate in the keyword auction, the firm’s decision to participate is endogenous in our model

and depends upon the minimum CPC as well as the relevance of both organic and sponsored links

to searching consumers. Xu et al. (2012) analyze two firms that differ with respect to their organic

ranking and compete in the product market. Like us, they find that bidding strategies depend on

the relevance of the firm’s organic and sponsored listings, but the keyword auction they consider

3Yangand Ghose (2010) show that organic and sponsored links tend to be positively interdependent. In particular, total

click-through rates, conversions rates, and revenues are significantly higher when both sponsored and organic links to

the firm appear on the search results page. However, Reiley etal. (2010) find that sponsored links may substitute for

organic links. Other factors beyond the scope of our analysis, such as the presence of images on the search results page,

also impact click-through rates (see, for example, Metrikov et al. 2014). Empirical analysis of these factors is typically

limited to sponsored links and does not consider the role of organic links.

258 International Journal of Economic Theory 14 (2018) 257–277 © IAET