Statistical Management of Pay-Per-Click Processes for Search Engines
Suppose you want to buy a dishwasher. What you might do is go to a search engine such as Google or Bing and type “dishwasher” in the search field. If you are using Google, what you may see are some sponsored links (adverts), as well as the results of natural search offered by Google’s search algorithm. The sponsored links appear because the keyword you typed is a keyword that a company has paid Google to display whenever someone searches for it. Broadly what happens is that if you then click on the advert, the sponsoring company will pay Google—or the search engine you used—a small amount. This is called Pay-Per-Click (PPC). We describe statistical models and methods which are used to automate and optimize daily PPC bid-price setting over portfolios which can contain hundreds of thousands of products and keywords, with the aim of maximizing the flow of customers and revenue to online retailers.
KeywordsOnline Retailer Forecast Revenue Generic Search Engine Multiple Choice Knapsack Problem Natural Search
Part of this research was funded by Knowledge Transfer Partnership KTP–7499, funded by Summit and by the UK Technology Strategy Board. We are grateful to Summit for providing data and to colleagues there for providing expertise.
- 1.Summit.: Back to the future for performance marketing. White paper on predictive analytics (2015). http://www.summit.co.uk/predictive-analytics-for-performance-marketing/