An Empirical Analysis of Sponsored Search Performance in Search Engine Advertising

1. An Empirical Analysis of Sponsored Search Performance in Search Engine Advertising Anindya Ghose Sha Yang Stern School of Business New York University

2. Search Engine Marketing • Internet has changed the way consumers search for information. • Search engines act as intermediaries between firms and users. • Refer consumers to advertisers based on user-generated queries. • Paid Search Places keyword advertisements on the search engines based on auctions. • Organic Search Users find web sites having unpaid search engine listings, as opposed to using the paid advertisement listings.

3. Search Engine Marketing Paid Search Listings Organic Search Listings

4. Search Engine Marketing % Online Advertising Revenue Very Cost Effective Growth of Search Other Rich Media Sponsorships Classifieds Banners Search Search now the most dominant form of online advertising Source: Interactive Advertising Bureau, PricewaterhouseCoopers (September 2006)

5. Characteristics of Keywords Classification of user queries in search engines (Broder 2002) • Navigational • Transactional • Informational • Presence of Retailer information (Retailer name) • “K-Mart bedding” • Presence of Brand information (Manufacturer/Product specific brand) • “Nautica bedsheets” • Specific search or Broad search (Length of keyword in words) • “Cotton bedsheets” vs. “300 count Egyptian cotton bedsheets”. Prior theory to motivate study using keyword attributes

6. Implications? Prior theory to motivate study using keyword attributes • Presence of Retailer information • Presence of Brand infhormation • Specific search or Broad search Competitive/ Searchers/ Yellow Pages Loyal/Aware Consumers/ White Pages

7. Research Agenda • How does sponsored search advertising affect consumer search and purchasing behavior on the Internet? • Consumer: What attributes of a sponsored advertisements influence users’ click-through and conversion rates? • Advertiser: How do the keyword attributes influence the advertiser’s bidding decision? • Search Engine: How do the keyword attributes influence the search engine’s ranking decision? • Is there any evidence of cross-selling using sponsored search advertisements? • Conversion/purchase from the searched category • Conversion/purchase from the non-searched category but led by the searched category

8. Summary of Findings and Contributions • Hierarchical Bayesian model to empirically estimate the impact of various keyword attributes (Wordographics). • Also analyze the impact of these covariates on firm level decisions – `CPC’ and `Rank’. • Policy simulations suggest that the advertiser can make improvements in its expected profits from optimizing its CPC. • Search engines take into account both the bid price as well as prior CTR before setting the final rank of an advertisement.

9. Empirical Methodology Framework • Hierarchical Bayesian model • Rossi and Allenby (2003) • Markov Chain Monte Carlo methods • Metropolis-Hastings algorithm with a random walk chain to generate draws (Chib and Greenberg 1995) • Consumer level decision: Click-through • Consumer level decision: Conversion • Advertiser decision: Cost-per-click • Search Engine decision: Keyword Rank Models of Decision Making

10. Model N= number of impressions n = number of clicks m= number of conversions p = probability of click-through q = probability of conversion conditional on click-through • First, a user clicked and made a purchase. The probability of such an event is pijqij. • Second, a user clicked but did not make a purchase. The probability of such an event is pij(1-qij). • Third, an impression did not lead to a click-through. The probability of such an event is 1- pij. • Then, the probability of observing (nij,mij) is given by:

11. Empirical Models Consumer Decision Advertiser Decision Search Engine Decision

12. Data • Large nationwide retailer (Fortune-500 firm) with 520 stores in the US and Canada. • 3 months dataset from January 07 to March 07 on Google Adwords advertisements. • 1800 unique keyword advertisements on a variety of products. • Keyword level (Paid Search): Number of impressions, clicks, Cost per click (CPC), Rank of the keyword, Number of conversions, Revenues from a conversion, quantity and price in each order. • Product Level: Quantity, Category, Price, Popularity. • These are clustered into six product categories • Bath, bedding, electrical appliances, home décor, kitchen and dining.

13. Results

14. Results

15. Policy Simulations Overview Findings • Differences between optimal bid and actual CPC • Average deviation is 24 cents per bid • Generally CPC higher than optimal bid price (94%) • Differences in ‘Expected Profits’ and ‘Actual Profits’ per keyword • Regressions with optimal prices show that firm should increase bid price with Retailer or Brand information, and decrease with Length. • Determine optimal bid price • Impute profits with optimal bid and actual CPC

16. Ongoing Research • Expanded the dataset • Controlled for time trend effect • Collected information on landing page quality • Collected information on competition bid prices • Other robustness tests (quadratic terms for Rank, lagged profits, lagged conversion rates, etc)

17. Investigate Impact on Cross-Selling • Advertisements that had specific product information in them • These are clustered into six product categories • Bath, bedding, electrical appliances, home décor, kitchen and dining. • Each order leads to a purchase from the searched product category and a purchase from any of the other five non-searched product categories.

18. Model: Two Stage decision process • First stage consumer decides on how much to spend in ‘searched’ category. • Second stage they decide how to much spend in ‘other categories’ conditional on expenditure in ‘searched category’. • The latent buying intention in the searched category is: • The latent buying intention in the non-searched category is: • Non-linear fully non-recursive simultaneous equations

19. Results on Cross-Selling • Non-Searched Category • Searched Category

20. Takeaways • Empirically estimate the impact of various keyword attributes on consumers’ search and purchase propensities. • Retailer-specific information increases click-through rates and brand-specific information increases conversion rates. • Implications for products of interest to “loyal consumers” versus “shoppers/searchers”. • What are the most “attractive” keywords from a firm’s perspective? • How should it optimally bid in search engine advertising campaigns?

21. Takeaways • Analyze the impact of these covariates on advertiser and search engine decisions such as CPC and Rank. • While the advertiser is exhibiting some naïve learning behavior, they are not bidding optimally. Policy simulations suggest that the firm can make substantial improvements in its expected profits from optimizing its CPC. • Search engines take into account both the bid price and prior CTR before setting the final Rank.

22. Takeaways • Firms can set up cross-selling opportunities on their own websites by advertising keywords. • Branded keywords play an important role in influencing spillover effects in sponsored search. • There could be synergies in promoting both categories simultaneously rather than separately.