Some quick notes: if you are going to be at the Emetrics Summit, we are having a table for WA bloggers on Monday. Any WA bloggers are invited, although we might need more than one table…

IE6 can be pretty bad for blogs with lines that go over the margins, I learned. Many thanks to Chris at Moody’s Economy.com and Joe at More Visibility for pointing out the problem to me. I was just kind of lazy when Chris wrote me (because I go home every night and work on rethreading almost 300 posts on Wordpress, it is hard to worry about this blog too.) So Joe, you were just the fire I needed to fix the blog.

Coming soon, the next part of the multi-part series on GA filters. If you are having trouble with a filter, send it to me, I am doing so many tests that I might as well include yours too.

Robbin

Just imagine what it must be like to be the lone SEO working in a company that is devoted to conversion science and web analytics.

That’s the situation that LunaMetrician Taylor Pratt faces. And wouldn’t you know it, we have made him drink our Kool Aid. So he wrote his article for Marketing Pilgrim (they are having another SEO contest) about how much your web analytics can bring to your SEO/SEM.

I hope you will check out his article (especially because I think they rate based on traffic. I wonder if they include engagement. Comments? Diggs? Can’t you wait to go to the Summit and talk more about visitor engagement? Ooh, but I digress, let’s support SEOs who care about WA.)

Robbin
LunaMetrics

I promised to write about Google Analytics (in this video). But first, I want to talk about profiles and Regular Expressions, because they will make your work so much easier.

Profiles. So you’re learning about filters, and you’ll probably make some mistakes. Join the crowd. But why make mistakes on the data that you’ve been using for a year now? Keep that “production data” holy, and experiment on a sandbox profile. Even if you think you are an expert at GA, always have at least once sandbox profile, and preferably two.

(Need to understand what profiles are? Well, certainly, you can use a profile within an account to measure a second website. But here, we aren’t talking about profiles for a new website, we’re talking about profiles for the same website. This is one of those concepts that is hard to understand at first, but is trivial once you get it. The idea is, you have multiple copies of your web analytics, all measuring the same thing, and if you set them up exactly the same, they will look exactly the same. However, you don’t have to set them up the same — you can keep one as your “good” copy, and the others can be used to learn. Need to learn how to configure a second profile?)

Having two clean (i.e. no filters) sandbox profiles will help you in a variety of ways: First, you don’t need to worry that the other filters on that profile are messing you up somehow. Second, they both start (one with and the other without the filter) at the same time, so when you write me and ask me why your filter doesn’t work, I promise I won’t ask if you chose a time period that pulled in unfiltered data. Third, since you won’t have taken yourself out of the data (because most people use filters for that, all except those who build special cookie workarounds), you can test it yourself doing all the strange things you’d like to check out.

Regular Expressions. Most filters require regular expressions. Now that I’ve gone through fourteen posts on Regular Expressions (RegEx) for Regular People (and specifically, for GA), I will be referencing that data. And if you already know it, you’ll think that this filter stuff is easy, easy.

Robbin

Did you know that you needed five non-conversions for each of your combinations in Google’s Website Optimizer? Yeah, me neither.

When I was at Google last week, I was showing the results of a 72 way test to Eric, one of the engineers who is behind Google’s Website Optimizer product. (In case you are not familiar with WO, it is Google’s free multivariate software. It enables website owners to test multiple pieces of a page and even a funnel or series of pages to see which combinations of the variables convert the best. Read why multivariate testing is like a game of Clue.)

Well, anyway, I showed it to Eric, and he pointed out that even though we had a few combinations with greater than a 97% chance to beat the original, one variation was’t even showing up in the Combinations tab. “And that’s your best one so far,” he said. It converted 14 out of 17 times, the dashboard proclaimed. The reason that the combination data didn’t show, he explained, was that there needed to be enough data in general for that test, and there needed to be at least five conversions and (here is the part I didn’t expect) five non-conversions. In the case of this particular combination, there were only three non-conversions. (17-14, right?)

So why the need for non-conversions? After all, the more conversions, the merrier, right? Well in fact, no. “It does best when the conversion rate is 50%” Eric emailed to me.

In order to begin to wrap your head around this (assuming you aren’t a statistician), you have to stop thinking that conversions are good. Instead, there are two states here, a or b. Conversion or non-conversion. Heads or Tails.

So let’s say that we take a finite, maybe only 20, visitors and estimate the conversion rate based on the fraction of the 20 that converted. Is that a good estimate of the true conversion rate? The holy “Law of large numbers” in statistics says that the average conversion of a finite set of visitors becomes a good estimate of the true value as the number of trials becomes large. But, the fine print in this law states that the number must be really large when the true mean (the true conversion rate) is very small (very few convert) or large (nearly all convert). In fact, for the estimate based on finite visitors to be good you need to have enough counter examples. “Counter examples” are non-conversions when the conversion rate is high, or conversions when the conversion rate is low.

I know, you want to know where they got the number five from, and why it’s an absolute number and not a percent. Me too. I’m thinking that the issue is, it can’t be a percent, because if you have only 20 visits, you need a high percent, and if you have 100 visits to that combination, you need a low combination. By fixing a specific amount, you make sure you get something, for both conversions and non-conversions. For both heads and tails. But that last part is speculation. Now if you want to learn something really cool about WO, go over to ROI Revolutions’s blog and read Shawn Purtell’s magnificent piece on the marriage of GA and Website Optimizer.

Robbin Steif

(Were you looking for real training? We’re having a Google Analytics Training Day in NYC on June 4, 2008. Check out the agenda.)

Last week, I was at Google Analytics Approved Consultant training. It was so wonderful; I got all sorts of tips and tricks. As always, one learns the best stuff in between the sessions, from the other consultants. (Yet another reason to go to the Emetrics Summit.) A special thank you to Caleb Whitmore from POP – if you divide his IQ by his age, he is absolutely the smartest yearly analyst around. (Did I just create my third KPI in the last six weeks??)

By the end of the training, we were so wiped that Justin and I just sat down to create this video. (You might have to click through to the blog site to actually see this. That’s always been my own experience with other peoples’ videos.)

UPDATE: Join us at the Harvard Club in New York City on June 4th, 2008 for our Google Analytics training event: Getting Ahead with Google Analytics. This is a great opportunity to learn from the experts in a smaller conference setting - you won’t get lost in the crowd!