�ݺ�ߣ

Editor's Notes

• #2: Thank you, Dana, and before I start I just want to say it’s an honor for me to be on the same panel with Bill Hunt, who I first saw at SES Chicago in 2004 on the Big Sites/Big Brands panel. Today I’m going to talk about big data, and give you all a little something that you can do today to better understand your audience.
• #3: First of all, I’m assuming that most of you have seen this photo, since it was the most retweeted ever on Twitter. We did just get through a brutal election, but this is significant to marketers regardless of your political party.
• #4: Rand Fishkin mentioned it on Twitter after the election. For those of you who haven’t seen it, he says…
• #5: The future of marketing. What does that mean? Well, for those of you who haven’t read the Time magazine article, Obama campaign staffers used data more in this election than has ever been done in order to understand their audience. Why? To quote one staffer…
• #6: One example of how they used data: they could tell from modeling that the people who gave the most for the Have dinner at George Clooney’s house with the president contest were women in their late 40s to early 50s from California, so they targeted them with new contests and new celebrities. The result: they raised $1 billion dollars from donations and won the election.
• #7: Another winner that night: this guy. If you don’t know him, this is Nate Silver, who was pummeled by political pundits prior to the election, and then correctly predicted the vote in all fifty states with statistics. He has since become a hero to geeks everywhere and inspired a Nate Silver facts campaign on Twitter, including this winning entry…
• #8: How did Nate Silver do it? In a similar way to how the Obama campaign did it: by harnessing the power of data to model likely audience behavior. This is significant because data trumps opinion, as Google often says, and there was no opinion in Nate Silver’s prediction.
• #9: Believe it or not, we’ve been doing a version of this in search marketing for a while. Some of you may remember the Overture Keyword Selector Tool at the right, which is what we used for keyword research before Google and others gave us a better option. Historically we have used keywords as a proxy for the consumer. Looking at this we can say that it’s likely that most people who are researching SEO online are looking for the exact phrase “search engine optimization” as opposed to optimisation with an ‘s’. Also, people who are looking for SEO are mostly looking for help, either in the form of an agency or an individual consultant. As SEOs and SEMs we have used this information historically to make our web sites and our ad copy more audience-centric, which has certainly gotten us a long way as an industry.
• #10: However, there is a lot more ground to cover. One of the biggest problems with using keywords as proxies for the consumer is that you don’t know if the keyword is qualified. By looking at the audience, and not just the keywords that they use, you can have a better idea of whether
• #11: At Resolution Media, we go beyond the keywords to show the specific data, including the keywords, giving businesses a better sense of who their audience is. We have a product called ClearTarget that…
• #14: I know that not everyone has the budget or is ready for this level of audience targeting, but everyone here can start building data-driven personas today. Today I’m going to show you how you can build one in fifteen minutes, and then talk about some of the applications of that persona.If you have a URL, you can go into Quantcast and get demographics and pyschographics easily, so we’re going to do things the hard way.
• #15: In keeping with the George Clooney theme, let’s say we’re promoting his new movie Gravity, which comes out in 2013. We don’t have a URL because there’s no official site, but we want to get a better sense of who our online audience is so we can market to it to a relevant audience.
• #16: At Resolution we build personas by asking questions about our audience. Questions such as…
• #17: Unfortunately none of the usual sources can give me demographic data for something with so little traffic, but by looking at the demographics for the official trailer I can start to get a better sense of who’s viewing my video. You can see there, the top demographic is Male, 35-44.
• #18: To get as much information as possible, we can enter related keywords in Google and see what else this audience is searching for.
• #19: We can then take that data into Excel and run Bing ad intelligence on the list to see if demographic data is available. Here you can see the data is somewhat spotty, as several categories are empty, but effectively the audience looking for the movie trailer on Bing is mostly females over 50.
• #20: Because there’s not a lot of data around this specific movie, it helps to get creative, and expand into the category of things of which this movie is one. For example, “georgeclooney movies”. For that we know it’s mostly females over 45. Or “sandra bullock movies”, which is mostly males of the same age. And also fans of science fiction movies, including 2001: a space odyssey. And we know that those people tend to be males over 45.
• #21: We can put these keywords into comScore Search Planner and get keyword demographics, as well as sites that these types of people spend their time at.
• #22: Using all this data we can deduce with only the data that our most likely target for this particular movie is a 45 year old male from Cleveland who is a senior web producer who makes over $100k and has two kids. This is the easy part.
• #23: Once we have that data we can use the Google Keyword Tool and Google related searches to build a semantic map of his interests, which include the following…Ideally we would also have access to conversion data from analytics, but since there is no site available we can at least determine that this gentleman is interested in watching movie trailers and reading reviews, which can be KPIs for a new site.
• #24: We can also dig deeper beyond the concepts into the keywords and see that someone like our persona is much more likely to search for concepts they know, like georgeclooney movies and sandra bullock movies than a movie title that they don’t yet know. We would probably target “gravity movie”, but also “georgeclooney movies” for awareness.
• #25: These days we know our target is probably not just accessing content from a desktop computer, so we need to know how he accesses media and where. To discover this we can use Bing Ad Intelligence again to understand the breakout of query volume by platform…
• #26: Doing that we can see that our target actually does mostly use desktops for these types of queries. This may change once the movie is released, but for now we know that desktops are key to this campaign.
• #27: Putting it all together…
• #28: What can we do with this information? A lot. We can use it to inform search campaigns, but it can be used to inform other types of media as well…
• #29: There are a number of benefits to this approach…
• #30: There are also some drawbacks…But ultimately, knowledge is power, and the more information you have about your prospect beyond keywords, the better off your campaigns are going to be.
• #31: Thank you! I look forward to answering your questions after the panel.