I just wanted to provide a visual equivalent to Avinash’s excellent explanation about the difference between Visits to Transaction and Path Length. Visits to Transaction is in the E-commerce > Time to Purchase report section, whereas Path Length is part of Multi-Channel Funnels.

Here is an actual example of Visits to Transaction:

Looks like 30 out of 33 transactions were completed in one visit. Maybe I don’t have to “worry” about attribution at all? Compare that to Path Length:

Same number of transactions, but only 23 had a path length of 1. The difference is due to the fact that Visits to Transaction only counts visits from the last campaign, whereas Path Length shows you all interactions for the past 30 days. You can reconcile the different attribution models in the Top Conversion Paths report, where I added how the data would look in Visits to Transactions:

The Path Length report really gives you a richer understanding of how visitors interact with your site. The takeaway is that there are typically more interactions prior to a transaction than you might otherwise believe by looking at the Visits to Transaction report.

More interactions means that you should help your visitors achieve their immediate goals at each interaction, even if those are not actual purchases. If a visitor “needs” six different visits from an organic search before placing an order, so be it. You’ll want to provide enough value during each of those visits to get the visitor to come back (and hopefully place an order).

I have always been a big believer in tracking mini goals. Abandonment occurs earlier than you may think: you are not going to convert someone if she is not able to get to the product detail page first. Or maybe a visitor is not ready to pull out his credit card, but gives you permission to continue the dialog by subscribing to your newsletter. That is a worthy goal to track.

There is another reason to add upper funnel goals: more data to work with. If your e-commerce conversion rate is 2%, are you going to ignore the other 98%?

Where this clearly comes into play is Multi-Channel Funnels. You only get conversion data from the goals you set up.  If you only have data for relatively few transactions you would be really limiting yourself.

Try this for fun: Set up a goal for non-bounce visits. I like that goal in any case because many visitors who bounce off your site are never going to convert anyway. Set your sights on those who give you at least one click!

After you have collected some data, fire up your Multi-Funnel reports.

Again, if you only have an e-commerce goal funnel you’ll only get data on those conversions.

In this case I have 249 Sale Confirmations, but over 2,400 non-bounce goals! Far more data to work with.

Lets take a specific example: the (not provided) keyword issue. If I look for (not provided) keywords in any interaction, I get 19 different results. By the way, I set up some custom groupings to illustrate the point more easily and not reveal the actual keywords.

Contrast that to the non-bounce goal:

I have 5 times as many data points than before and I can now dig a little deeper. How do visitors interact with Adwords, brand and non-brand terms etc? I may never know what’s hidden in those (not provided) keywords, but at least we can know a little more about those people who interact multiple times with our sites via multiple channels.

In any case, the point of this post is to create more goals – if only for the plain reason to get more data. I know that Google Analytics will have to work a little harder when you throw in additional goals, but I am sure they don’t mind…

The site speed report in Google Analytics got a nice upgrade recently with the introduction of the performance report and the ability to track virtual pageviews.

The performance report shows us how the page load times are distributed and you can see that the average page load time may not the best statistic to describe that distribution.  The trouble is that you can have outliers with extremely high/non-sensical values:

and these will skew the overall average page load time with one single sample:

Far better in my opinion to use the median with page load times. Just add up the individual percentages until you get to 50%, and track that value over time. In the above example the median would be in 1.5 – 2 second bucket, which is not affected by the outlier.

[Update 11/03/2011: We have since added goal conversion rates as well. However, please note that it uses the overall goal conversion rate, which is the sum of all goals. Use appropriately]

Can you be sure that a conversion rate of 5.56% is better than 4.87% if you only have a few data points? It’s important to think statistically when working with this kind of data.

I wanted to share a Z-Test calculator that you can use directly in Google Analytics via a simple bookmarklet. It uses the jstat library. You’ll need to use the new version of Google Analytics and have e-commerce tracking enabled as it works on the e-commerce conversion rate.

Please drag the button below to your bookmarks bar (you’ll need to be viewing the post on my site directly, not in an RSS reader):

Z-Test

And then click on it in any report that has Plot Rows checkboxes and select the e-commerce tab. Or take a look at the 2 minute video:

(You may want to watch a larger version directly on youtube)

A couple of notes. I decided to show the confidence interval so that you can decide if you want to use a 95% confidence level. Hat tip to my friend Kevin Potcner for giving me feedback on this.

Doing a statistical analysis on observational data will perhaps not give you quite as strong a signal as from a controlled experiment, but it’s still a valid approach. Try to isolate the question you are trying to answer by segmenting your data.

Have fun!

Michael

P.S. I tested it in Chrome and FF on a Mac, but if it doesn’t work somewhere else I am sorry. It’s just a tool we built for ourselves, which we are now sharing in the hope that it may be useful for other folks. Feedback is of course appreciated!

The impact of the SSL Google Organic change can be nicely visualized using Motion Charts since we can use up to 5 dimensions at the same time. For example, are the (not provided) keywords more like branded navigational terms or more like longtail? What’s the bounce rate? What’s the percentage of new visits or the impact on revenue? And since this is a gradual rollout it’s interesting to see it unfolding over time…even if you don’t like what you are seeing

Lots happening in the world of Google Analytics in the past few days: Realtime, Premium, Webmaster Tools integration and yet another design update coming soon.

I actually experienced an emotional response when I first saw the numbers and charts updating live. It feels very personal to see a certain number of people from different parts of the world right now on my site, giving me a little bit of their valuable time. Is my site making them feel welcome? Are they finding what they are looking for? Please, no 404 page

Back to business. Having realtime data means you have to be able to take action in realtime.

Realtime data is really useful when you know what to look for, such as checking that a new campaign has been implemented properly. After you launch a new Adwords campaign, are the tags firing properly on the landing page? If they are not, we can quickly add tags or pause the campaign.

Another cool use is if you have implemented a new AB experiment. With Website Optimizer it currently takes quite a bit of time between the launch of the experiment and getting data. Now you can know right away if you have messed up the tagging:

Realtime is great for tactical stuff like that, but I would love to see a couple of additions:

1) Allow regex in the search box so I can monitor groups of pages.

2) Add Page Titles to the realtime Content report. Sometimes I can’t identify a page by its URL, but rather by its title, such as 404 Page not found pages.

Given the pace of development I am sure they are working on the next version already!

Forget all the fancy new report capabilities…the real reason why Google Analytics launched multi-channel funnels is because they got tired of having to explain why Google Analytics showed fewer conversions than Google Adwords reporting. OK I made that up, but in my years of doing web analytics this is probably THE number one question I get and I am glad I can now point folks to this post.

Despite the fact that Google Analytics and Adwords use different tracking approaches and attribution models I guess it feels “wrong” to many people that Google Analytics would show different/fewer conversions than Adwords. But instead of me trying to explain why this is the case I can just show you by way of a simple example.

I picked a week and looked at Adwords reporting

Then I looked at the same date range in Google Analytics.

Look familiar? Fewer conversions attributed to Adwords – there must be something wrong! Maybe bad tagging?

Enter MCF.

We know that Adwords takes credit for a conversion within 30 days of the initial click. Doesn’t matter if it was first touch, assist or last touch. To simulate this behavior I created a conversion segment where ANY interaction is paid search.

Now look at the data in the Assisted Conversions section.

Drumroll…..tada!

The Paid Advertising/Adwords channel actually shows fewer Last Interaction Conversions than Google Analytics – 204 vs 219, but we know from the Google Analytics attribution model that Direct visits do not overwrite a campaign cookie. Add Paid Advertising and Direct channels and you basically get the same number: 218 vs 219. Voila – that explains the Google Analytics data.

Now, add up all channels and you get 273 conversions vs 276 from Adwords reporting. I will take that margin of error any time.

Aren’t you just itching to delve into understanding the relationship between Paid and Organic channels? But even before we attempt that, we can now reconcile Adwords and Google Analytics reporting. That’s pretty sweet…

On August 11 Google Analytics changed their attribution model and it’s had a bit of a dramatic effect in some accounts. Visits are way up, and metrics that depend on visits such as e-commerce conversion rate are consequently affected as well. There was a bug that was fixed on August 17 that did inflate visits even when the campaign source did not change, but visits seem to be back to “normal”.

Some people seem quite upset about the change, but I am not sure whether this is about the temporary bug or about the underlying change, which in my opinion is great news. I totally disagree with comments such as this one:

As a data analytic, the most important thing is data consistency. The ability to measure changes in traffic, conversion, bounce, etc. is the most important aspect of an analytics effort.

As with any model, it’s not so much about being “right” or “wrong“, but whether the model is useful. Does your model help you understand your customers a little bit better and make smarter decisions? If there is a better model then you should change to that one.

Ironically, the data causing higher visit numbers was there all along! It was just hidden due to the way Google Analytics used to count visits, pageviews, etc. Anyone remember 0 visits?

Here is what is happening by way of an example.

1) Grab this custom report for the new Google Analytics and pick a date before Aug 11

2) Filter the data to only show keywords where visits = 0

I see something like this:

Almost 6% of visitors have such 0 visit keywords. Now go try and find any of these keywords in your regular keywords reports. You won’t be able to find them because they have been taken out.

Now pick a date after Aug 11.

Almost all of the 0 visit keywords have disappeared because they are now attached to a visit. Hence they will show up in your regular reports. You should therefore find that you actually have more keywords now.

So there you have it. 6% more keywords in my example, which indeed opens up the possibility of having better attribution management as was pointed out in the post. The bigger the change, the more you should worry about attribution.

Finally, I am not a huge fan of visit-based metrics anyway. Who cares if your dear customers take 1 or 2 visits before placing an order? The main thing is that they accomplished their goals during each visit, i.e. browse in the first visit, have a cup of tea, then buy in the second visit.

Segmentation is great, but segmentation in context is even better. Just wanted to point out one subtle update in the new version of Google Analytics.

Let’s say I want to compare engagement metrics for two different segments. Specifically, I want to compare buyers vs non-buyers who have viewed more than 20 pageviews in their visit.

Here is what it looks like in the “old” version of GA:

And here it is in the new version:

With the exception of the All Visits row, the bars for the two segments look quite different (nb: I removed the pageviews bars for clarity). How do you get from 1.11% to 36.35%?

In the old version each dimension is evaluated in the context of overall visits. Nothing wrong per se, except when you work with segments that are very small as is the case with the “new and buying” segment; it gets “dominated” by the All Visits segment and the small 1.11% can easily be overlooked or dismissed.

Contrast that to the new version, where each dimension is evaluated within the context of the segment itself and not overall visits. I can now see that over a third of visitors within the “new and buying” segment viewed 20+ pages. All we needed to see this valuable information is a subtle change of context.

In general you should be particularly concerned with tracking elements on your site where your visitors (hopefully) take action, such as adding an item to cart or filling out a form. Same thing with social buttons – you want to know if visitors click your Like button. These are mini-goals and should be tracked. So it’s great to see that Google Analytics now has a dedicated report for tracking social plugins, but it might be even better if you could add goal tracking to Social Actions, just like they did with event tracking.

There is of course a cost to adding external scripts on your site, such as increasing page size and giving out data to the button providers. Remember that external scripts can log valuable web analytics data such as referrer info, so you just have to weigh the pros and cons. But only add buttons if you actually want visitors to see and use them. There is no point adding buttons on your site, only to hide them inside tabs where they will never be clicked.

And of course, you should test whether having buttons at all helps your site goals. You’ll probably find that a small minority of users actually click on social buttons, but maybe just having them provides “conversion-boosting” social proof.