bq. On YouTube, ads can be targeted based on either category — news, politics, entertainment — or geography, by city or ZIP code. But Bachmann’s campaign was able to target much more narrowly, cross-referencing voters’ IP addresses with voter files that have long directed mail, phone and door-knocking efforts…
bq. …Bachmann won the straw poll, and though the effect her ads had on the tiny electorate is impossible to measure, the statistics are impressive: More than a half-million Iowans saw the videos, and roughly 75 percent sat through them to completion — a rate that suggests the targeting worked, Everhart said.
From this Ben Smith post. The statistics would be more impressive if there were actual measures of preferences or, even better participation in the straw poll, not just exposure to the ads. But given that presidential primaries see much larger swings in preferences than many other kinds of elections, and given that it’s early and people’s opinions about the GOP candidates are hardly crystallized, the persuasive effect of Bachmann’s web ads is at least plausible.