Good morning, folks! Today is a big day.  We will have seven sessions coming to you live from SES NYC.  That’s right, SEVEN!  We’re starting off with a keynote from Avinash Kaushik on how to be awesome.  I think, that’s what it’s about. I’m a little intoxicated after a truffle incident.  There’s also a giant strobe light that keeps hitting me in the face while I Got A Feeling continues to blast in the background.  It’s making me feel more intoxicated. And giving me a headache.

Dear flashing purple light, please stop.  Please? I’ll share my champagne truffle?     Okay, the keynote’s about to begin. Mike Grehan is up on stage looking dapper. And British. Though, I guess that’s redundant.  Why did I eat that truffle?Okay, Avinash is up.  He talks about the Market Motive course and says that at the end of it you become a ninja. Ooo, ninjas!

He wears many hats. He works at Google.  He’s on the board of a few companies. Today he wants to step back and go a bit higher. He’s going to talk to us as the author of his two books. He’s going to share why search requires us to dramatically rethink our fundamental beliefs about marketing.  He’s going to tell us four stories.  I hope there’s milk and cookies after. Because I could go for a nap. It’s going to be a really long day.

Break Down Limits…visualize and interpret

When you think about Web analytics, you think about spreadsheets and numbers. The problem with tons of data is that our eyes can only see so much of it.  It’s hard to look at lots and lots of data.  There are tricks you can use to help you understand the data. You can do hypothesis testing to help you find the keywords that are super important to you. By creating it, you get something that tells you how many new visits are coming to your Web site and how many of those keywords meeting a certain criteria. He talks about keywords he wants to make love to because they’re so damn good to him.  Oh boy. This is getting steamy for 9am.  Avinash must have had the same truffles I did this morning.

Tag Clouds: He’s a big fan of tag clouds. He says they’re a great way to visualize tons of data.  You can take thousands of keywords and see how your SEO and PPC strategy is going.  It can help you find the long tail of what you’re doing. It can help you determine if you’re targeting what you should really be targeting.  If you want X and you have Y… does that match up?

He mentions Juice Analytics. They have a concept of keyword trees. You connect their visualization tool into your analytics and it will analyze your data and create a tree mapping out all your keywords. It will show you keyword that lead to a term and ones that lead away from a term. It will help you show were you suck a lot. Those are his words, not mine.  By using this kind of data you can look at tons of data and look beyond the top ten rows that you normally see. And you get to do it all in one screen. Most people’s search strategies are so horrible because they’re so obsessed with ten rows of data.

Outcomes, Baby!

The most common complaint is that people don’t have enough resources. The common thread is that most of the time people are focusing on crappy things like page views, visits, time, and other useless things that are not important. If you want to change the way your people think about the Web, you have to focus on outcomes.

Why don’t we get love?

Story: Convincing The Boss (his wife).  The only time he can blog is after midnight when his kids are sleeping. His wife loves him a lot and tells him every day to go to bed. He says he can’t because he has to blog… and she tells him to go to bed.

He says he needs to blog he’s ‘kind of a big deal’. His wife looks at that and says, “go to bed”. Because that data is not useful. We do search the same way. We say we should do things because “it’s cool”.

He will tell her that last month he had 73,000+ visits from 176 countries. It’s data proof of the statement he made a few minutes ago. His wife still doesn’t care and tells him to go to bed. The data isn’t great because it doesn’t connect what is actually important. It’s a meaningless number.

He goes back one more time and shows her the analytics for his listed goals. He wants people to read two pages on his site because that’s how many it takes to convince them of his greatness.  He also wants you to read his About page because he wrote really nice things about himself there.  He wants you to subscribe to his blog.   Then he can tell his wife how many dollars he made from his blog.  She tells him to work harder. Hee!  That number helps her to see the value and gets him the pass to blog that night. You want to talk to people in language and numbers that mean something to THEM.

You want to have multiple goals.

It’s really important that you’ll never achieve greatness on the Internet unless you can quantify the impact of the work you do. What makes the Internet awesome is the economic value.

The Looooooong Tail

He read the book by Chris Anderson and didn’t see how the phenomenon applied to the Web.

His blog had 7,861 visitors last month. About half that traffic comes from search.    Those 40,000 that come from his blog from search came from 26,137 different keywords.  He only writes about one topic – and yet people found 26,000 different ways to come to his blog.   That’s the story of the long tail.

Once you get into the data you’ll find your head is filled with your brand terms.  People who come from the long tail typically use more keywords. They’re using keywords that are all related to your category, not your brand. The people who live in the long tail, don’t know what they want. They haven’t made up their mind. And if you can get in front of them first,  you get to convince them. In order to exist, you have to show up.

Most companies cannot find the right people at the right time. The bitch about the cost of [digital camera] but they miss the cheaper keywords that would bring even more people. The problem with attracting the people who already know you is that at some point those people will DIE! And then what are you going to do?  [Hee!]

You find the right person at the right time by using data.  Whether you show up or not, the salmon is going to go by. All you have to do is open your mouth. That’s data-driven marketing.

Intelligent Attribution

You can use data to be very intelligent about what you do.

Don’t freak out. Is this a problem for you. Run a Visits to Purchase report. It tells you the number of visits it takes for someone to convert on your Web site.  If people convert after 9 – 14 visits, you may have a problem. Run the report, but don’t freak out yet.

Instead of trying to boil the entire ocean, focus your efforts where you’ll have the highest ROI.  You want to only freak out about ONE thing. Giving the credit to the first click is like him giving the credit for marrying his wife to his first girlfriend. It’s really stupid. [Love. It!] Now we use a different model called Make Crap Up. We create a random set of rules to figure out how to deal with things.

He shows an eCommerce Web site. The model they use is that they give the last click 75 percent of the conversion.  Then they decided the remaining 25 percent would go back a certain number of days and clicks.  For the really impulse purchases, they don’t care about attribution. They lose money on it. For more impulse to less considered, they go back 7 to 14 days. For less impulse to more considered, they go back 14 to 30 days.  Considered items go back 30 days or more. They can go back in history to find clicks and create a model quick enough that it’s still relevant.

Who gets credit for how much is the wrong question.

You want to ask yourself: what is the mix of media that produces the highest amount of ROI, regardless of the crediting.  You want to know what portfolio of channels you should be using.

The Future

1. Media Mix Marketing:  Pick areas around the US that are related to your market and run experiments to see what media is most effective.  You’ll find the optimum mix for your portfolio using real life data. Now you can change your spend without worrying about credit.

2. Marginal Attribution:  If you add one thing to your mix, what is the number of marginal implications? In the end, all of the politics to go away.

Always remember to do the basics right.  If you suck, that’s all that matters.


About the Author

Lisa Barone

Lisa Barone co-founded Outspoken Media in 2009 and served as Chief Branding Officer until April 2012.


5 thoughts on “Be Awesome: Ideas for Approaching Search Analytics Differently


  • Luke Regan on said:

    Thanks for the great write-up Lisa, hope you find a way to power on through the day. :)

    Question: do you think that, now an Analytics expert from Google has delivered a polemic like this on attribution, senior marketers will change their views? Away from fancy attribution models which are ultimately bankrupt towards a more pragmatic and scientific approach, albeit without the (emperor’s new) security blanket?


  • Zach Prez on said:

    Definitely going to try out the Juice Analytics Concentrate thingy for long-tail search that connects to Google Analytics. Great tip for simplifying that outlandish report. Thank goodness there is a free version for small blogs.


  • Data Entry Services on said:

    I see analytics becoming more and more sophisticated. Used to just try something and see what happens. Now it’s a full time job. I just use Google Analytics and check once a week or so.


  • SLR Lounge on said:

    Hilarious style of writing and great information. Juice Analytics Concentrate is a gem that I’ll definitely be trying out. Thanks for the info.


Leave a Reply

Your email address will not be published.

*


*

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Comments links could be nofollow free.