SEM or paid search is key driver for online marketing success. The response is directly linked to the keyword and success is quickly measurable.
Google has for a long time squeezing the organic content to lower position on the search result page for desktop. On the mobile device the number of slots are extremely limited so paid search is often given the priority over organic contents. SEM is becoming even more important as marketer cannot rely on SEO as much as in the past.
Many Internet metrics follow the power law distribution. The search query and number of clicks by keyword are just two examples of the power law. That means while the business manages millions of keywords the clicks distribution has a very long tail. Simply put, vast majority of the keywords in SEM campaigns will have very limited number of clicks while only a small fraction of the keywords will get large amount of clicks.
This classic problem of SEM campaigns needs to solve the problem of low probability events (ecommerce transactions, lead gen etc) and insufficient click history associated with the vast majority of keywords in order to provide expected revenue on a statistically significant basis. How to manage the long tail of SEM bidding keywords has been a challenging problem for business of all sizes.
The combination of data science and customized bidding approaches has yielded very effective results across multiple industries with proven record of successes doing data driven marketing optimization across broad range of advertising platforms including Google, Bing, TripAdvisor and Trivago.