The most important thing to know about the way Raven's Keyword Ranking tool works is that we're providing average rank data, rather than specific SERPs.
Traditionally, rankings are gathered through automated processes, rapidly searching Google (along with other search engines) and skimming the data from those searches for storage in separate databases. This practice, known as scraping, is explicitly forbidden by the terms of service of most major search engines, which means we don't do it.
Average Rank is calculated by looking at all of the times Google/Bing users saw your website in searches for a given keyword phrase, finding the highest point your domain appeared in those impressions and then averaging them all together. This winds up being a bit more reliable than scraped SERPs, as it takes into account geolocation, personalization and other factors which may change your position for an individual user.