Archive for August, 2009

Mistakes with Include Filters

Posted on August 28, 2009 by John

There is one particular mistake I see over and over with regard to setting up filters in Google Analytics.  This has been talked about before, but it comes up so often that I don’t feel bad about addressing it again (and again, and again).

The Mistake

Multiple Include Filters, causing no data to get into the profile.

What it looks like

1. Include – Request URI – /page.html

2. Include – Request URI – /other.html

What is intended?

The person that adds these 2 filters to a profile wants to include both /page.html AND /other.html.

But why doesn’t that work?

The Technical:

When GA processes filters it does so, sequentially, one at a time.  Filters are like a series of gates, in-line, along a road. In order for data to get from point A to point B, it has to pass through each gate in turn. If there is any one gate that it can’t get past, then no point B and it doesn’t end

up in the profile.

Each Include filter can be thought of as an “Include Only” filter.  Let’s look at the 2 example Include filters above with that wording in mind, and see how far our data makes gets.

Data –> A) /page.html, B) /other.html, C) /index.html

1. Include – Request URI – /page.html
2. Include – Request URI – /other.html

The first thing that happens is our data is checked against Filter 1.

Filter 1 says “Include ONLY data that has a Request URI that matches /page.html”
Data A matches /page.html
Data B does not match /page.html
Data C does not match /page.html

Now we have
Data –> B) /other.html, C) /index.html   -> didn’t make it past the filter
1. Include – Request URI – /page.html
Data –> A) /page.html
2. Include – Request URI – /other.html

The only piece of data that made it past Filter 1 was Data A, /page.html

Now it is tested against Filter 2, which says “Include ONLY data that has a Request URI that matches /other.html”

/page.html, is tested against Filter 2, /other.html.  It does not match, and ends its trip short.

So, no data in our set can possibly make it past both filters, since there is no single page that can match BOTH /page.html AND /other.html

The Solution

The solution is to use a single filter that looks like this:

1. Include – Request URI – /page.html|/other.html

We put both pages that we want to include into a single filter and seperated them with a pipe.  The pipe is the vertical bar symbol usually located just above the Enter key.  This is a Regular Expression (RegEx) symbol that can be read as “Or”.  So the filter says Include ONLY pages that match either /page.html OR /other.html.

View Comments (3 Responses) | Categories: Filters, Google Analytics

Why is this Summit different from all other Summits?

Posted on August 18, 2009 by Robbin

The Summit (more correctly known as the eMetrics Marketing Optimization Summit) is coming to Washington, DC, during the week of October 19, 2009.  Recently, I interviewed Jim Sterne, Summit Organizer Extraordinaire and Analyst to learn Why This Summit is Different from All Other Summits.

The first thing I learned was that this Summit will be more practical, real life stuff. “You mean, how do I make it with just one guy doing analytics (and he has to do SEO too) instead of three full time analysts?” I asked Jim. The answer was, yes, some of that, but “We aren’t hitting the bell and saying, times are terrible,” Jim said, “We did that last year.”  Now, he pointed out, there will be really concrete ideas.  To that end, there were will very few consultants speaking and many more practitioners.

Also, Jim pointed out that, “The tracks that we have are more focused. Instead of trying to do many things at the same time, there will be only three tracks, and they will be different every day. The first day will be multi-channel data, voice of the customer, and search analytics. Day two: Gov’t and non-profit, ecommerce and data-delving,”  (Not quite data diving, he noted.) Day three will be Marketing management (i.e. how do you organize corporate culture? How big should your staff be? How do you set up a global analytics dept?), BtoB marketing and  Emergent Media (socia, video, mobile, “all the new stuff.”)  Jim told me, “We are offering a one-day pass to the third day and is called “Bring the Boss day,” because this is the stuff that the senior managers care about more.”

I asked about auxiliary sessions — i.e. sessions on the Monday and Friday before and after the Summit, and there will be plenty. A GA session, a couple of sessions on predictive analytics, a landing page session.

Also, I was told, “We’re expanding the clinics and labs. We’re asking vendors to show us real life situations, how a client did a thing. Coremetrics, Hitwise and Google are going to do labs to show how a customer had a problem and it got fixed.”

Well, there you have it. The early bird registration ends on September 4.  You can get a 15% discount by using our code, LUNA09, or just by using your WAA discount.

View Comments (No Responses) | Categories: Industry News

Tracking Your TCT (Twitter Click Through)

Posted on August 17, 2009 by Jim

Love it or hate it, Twitter can be a useful tool to send traffic to your site. But how do you know if all that time spent tweeting is paying off? You may take a look at your Google Analytics and see some referral traffic coming from Twitter.com, but that’s not the whole story. It’s not even half the story.

Most people who use Twitter receive and send their tweets from an application on their desktop or smart phone. For example – Twhirl and Tweetdeck are popular desktop apps, and Twitterific is popular for the iPhone. Here’s the rub: if you tweet a link to your site and someone clicks on that link from within one of those apps, it will show up in your analytics as a direct visit – no referral data.

So, how do you know how many people visit your site from links that you send out via Twitter? Enter the world of URL shorteners. You’re probably already aware of URL shortnening services, like TinyURL, Bit.ly, kl.am, Snurl, and there are many more out there. Basically, what they do is take your long URL (http://www.example.com/blog/page/category/year/month/day/article/id=123456) and condense it into a shorter version (like http://bit.ly/8FapX). This is especially important for Twitter, where you only have 140 characters per tweet (the long URL above would only leave you with 73 characters – not even enough to say what you had for breakfast!). The short URL just redirects to the page with the long URL.

Some URL shorteners also have fancy features that let you track important stats – like how many people clicked through. Here’s a more complete guide to URL shorteners that lists the features of each one. I like to use Bit.ly most often, so let’s stick with that for the rest of this post.

Using Bit.ly to Measure Your Twitter Traffic

So how do you use Bit.ly to track your Twitter Click Through (TCT) rate? It’s simple, really. Before you tweet that link to your latest blog post, just head on over to Bit.ly, shorten the URL, then copy and paste the new, shortened URL in your tweet and send it on.

Once you’ve done that, you’ll get some great data (from within the Bit.ly interface) about your tweeted link, including:

• how many people clicked on your link
• how many people clicked on other bit.ly links (that point to the same page)
• referring sites (where your Bit.ly link was clicked on)
• location (country) of people who clicked through

Here’s a screenshot of what you can expect: (click to view larger)

*Disclaimer: Don’t get too attached to the numbers – they can lie to you. For more information on which numbers to trust, you must read Danny Sullivan’s analysis on Twitter traffic numbers – which looks specifically at Twitter, Bit.ly and Google Analytics.

Now, don’t forget to follow @LunaMetrics and if you liked this post, why not tweet about it? (I will personally thank you via DM – no auto responder crap – you deserve only the best!)

View Comments (9 Responses) | Categories: Google Analytics, Statistics

And what do we do with all that data, anyway?

Posted on August 13, 2009 by Robbin

This week, I am feeling the same kind of pain that those who come to our training sessions often feel. What do we do with all that data, I sometimes hear. What does it all mean?

Doing an analysis “cold.” I feel the pain because we have created a benchmark for a special set of websites, and now I have to do the evaluation. It was easy to do the overall evaluation, but when I sat down to do individual evaluations, I often found myself scratching my head. This was a particularly hard exercise, because I don’t work with these sites on a day to day basis and don’t keep up with the analytics — I approached them “cold,” you might say.

Asking the right questions. What I found was that when I happened to ask the right questions and look in the right places, I often came away with incredible nuggets of information. Golden data, you might say. But it wasn’t linear (first you look at this report, then you look at that report, etc.  Nope, not like that at all.)  There was a decent amount of luck involved. I was also able to bring a lot of experience from other sites that helped me ask the right questions.  And I had the good fortune to be surrounded by others who knew the project and were able to make suggestions.

Patience, patience. So be patient with yourself.  Maybe you can start by looking at your own benchmarking (in Google Analytics, it is at the top of the Visitor’s section of the navigation on the left) and ask, “Why are we doing well (or not doing well) in each of these areas?  I think web analytics includes discovery, but first, you have to get to know the territory. My current project requires me to work with the analytics cold, but many of you have the luxury of working with the same site and same analytics every day.  Get to know them.

View Comments (No Responses) | Categories: Regular Expressions, Web Analytics