Month: January 2011

Google Analytics API – Now With New Dimensions and Metrics

/ Ophir Prusak / 2 Comments

Google just added a boatload of new dimensions and metrics to the Google Analytics API:

http://googlecode.blogspot.com/2011/01/127-new-dimensions-and-metrics-through.html

I’ll spare you the technical details (you can read the official post) but I do want to comment on what I think is the most important change – 10 new Adwords dimensions.

Here’s why –

I admit I’m not an expert regarding Adwords administration and optimization tools, but until recently, they’ve had what I consider one very big flaw. Initially Adwords tools would look at the beginning of a visit; what happened on Google and the Google network such as impressions, CTR, CPC, etc and then what happened at the end of a visit, IF it ended with a conversion and you had Adwords conversion tags.

Then Google integrated Google Analytics goals into the mix which provided some additional data, but we’re still looking at the start and the end of a visit.

For sites that have a zillion visits and a few hundred conversions a day, you have enough data for analysis, though for the average small business site, there just isn’t enough data if you’re just looking at the end goal (sales, leads, etc).

In order to analyze the vast majority of visits that don’t end up in a conversion, you really need to look at metrics that serve as indicators for traffic quality such as bounce rate, time on site, page views, viewing key pages, etc.

This means that either the Adwords tool has it’s own internal analytics system (and you need to install yet another tag on your site) or it can take advantage of your existing analytics data.

I know a few vendors recently added Google Analytics metrics to the mix, which is a very welcome addition, but some key Adwords dimensions were still missing form the API.

Now that we have almost every Adwords dimension you could want in the API, I foresee a new wave of Google Analytics / Adwords integration, and eventually tools that will truly be able to automatically optimize your campaigns.

Advertisement

The Future of Split Testing and Conversion Rate Optimization

/ Ophir Prusak / 3 Comments

I’ve been fortunate enough to see and experience first hand the evolution of the Internet, from even before the web till today.

I’ll spare you a lengthy history lesson explaining how we’ve gone from brochureware sites to where we are today, but I do want to share some thoughts and perspective on where I think things are going.

When marketers started to understand the potential of dynamic web sites, there were two terms everyone was throwing around:

Personalization & Customization.

Fast forward to today (2011). The user experience is still exactly the same for all visitors (other than on a handfull of sites such as Amazon.com).

For the most part, web site Personalization has failed. Sure it sounds good in theory, but trying to tailor the web site experience at the individual level is extremely difficult. It is difficult both from a technological perspective but mostly by trying to create an optimal user experience based on data from a single individual.

There is no doubt in my mind that in the future (and to some extent today) the user experience when visiting a web site will be created dynamically based on what gets the best results, but based on “anonymous” information which is common to large groups of visitors, and not based on a single person.

This reminds me of the concept of Psychohistory from the science fiction series “Foundation” by Isaac Asimov.
Wikipedia explains it better than I can:

The premise of the series is that mathematician Hari Seldon spent his life developing a branch of mathematics known as psychohistory, a concept of mathematical sociology (analogous to mathematical physics). Using the law of mass action, it can predict the future, but only on a large scale; it is error-prone on a small scale. It works on the principle that the behaviour of a mass of people is predictable if the quantity of this mass is very large. The larger the number, the more predictable is the future.

I also like to think of this in terms of what usually happens at (successful) brick and mortar stores.

When you walk into a store, the salesperson probably doesn’t know you personally, but will probably try to help you based on certain public traits such as gender, age, if you’re by yourself or with someone else, etc.

Which brings me back to what actually prompted me to write this article in the first place :)

While I’ve been split testing since 2005 in order to improve conversion rates, the majority of the time, it’s still about what works best for the site as a whole, opposed to split testing together with segmentation (which is what we really want).

Until recently, there haven’t been many options out there to achieve this level of targeting and testing (at least not priced for small to mid sized businesses) but over the past few months, I’ve been starting to see more and more startups trying to bring this level of sophistication to the masses.

While I haven’t had a chance to use any of these services first hand, there is no doubt in my mind that business that truly embrace this level of targeting and split testing will eventually lead the pack and leave most one-size-fits all web sites in the dust.

Google Analytics on Intranets and Development Servers / FQDN

/ Ophir Prusak / 7 Comments

Just a quick posting about using Google Analytics on pages that don’t use a fully qualified domain name.

If you’re using Google Analytics on a site with a URL like http://intranet/ or something like http://mydevserver:12345 it won’t work.

Specifically, the Google Analytics JS code will not send the tracking hit (__utm.gif) to the GA servers.

I don’t really know the specifics, but I’m guessing that the domain hashing code looks for at least one period in the hostname and won’t work if it doesn’t find one.

Two alternatives come to mind:

1. Use an IP address if one will work. If you’re testing on a local machine 127.0.0.1 should work fine (that IP always resolves to the machine you’re on)

2. Turn off domain hashing. Simply using _setDomainName("none") in your code should also fix the issue.

Hope that helps someone who might be pulling their hair out trying to figure out why the page is not being tracked :)

Split Testing and Return Visitors

/ Ophir Prusak / 3 Comments

Just a quick post about a phenomenon I’ve personally seen happen but don’t recall ever seeing mentioned in split testing articles.

I’ll start by saying that ideally you should always look at the results from any split test by segmenting your visitors.

It’s not enough to know that overall version X did better than version Y. Ideally you should check how the different versions performed for various visitor segments. For example, users from organic search might behave differently than visitors from a referring site or direct traffic.

There is one segment though where merely the fact that you’re doing a split test can have an impact on the results:

New vs. Return visitors

Even if you weren’t doing a split test, you would probably see a difference between the two segments based purely on the fact that return visitors already know something about your product, service or site.

I’m talking about a different phenomenon though. The “something has changed” effect.

For new visitors, your site will be new regardless of which version of a test page they see.

For return visitors who have some level of familiarity with your site, if they see something new or changed on the site, they’ll probably pay more attention to it – merely because it’s different.

For example, if you’re homepage does not currently have any video on it and you test a new version with some video on it, return visitors who get the version with the video might watch the video simply because it’s something new.

Conclusion: Always segment visitors by new and return visitors.

If both groups show the same preference, it’s safe to say that you have a winner. If you’re seeing a large variance between new and return visitors, it might be worth it to let the test run for a while to see if the variance changes over time as more of the return visitors first visited the site after the split test started.

[UPDATE]
I was just thinking that this would be a feature that split testing tools can / should support. Segmenting not only new vs return users, but return users who’s first visit was before the split test started vs return users who’s first visit was after the split test started.

Heck – you should be able to only include visitors who’s first visit was after the test started if you want to. Are you listening optimizely, visual website optimizer, and the rest of the gang?