Author Archive
Posted on August 30, 2012 by Phil Anderson
Google Analytics is a big tool, loaded with lots of features. It’s possible, even for experienced users to not have discovered all of the nooks and crannies where features hide in Google Analytics. As someone who trains and consults in the tool, I’ve used everything at least once, but sometimes I forget these features are there as well.
- Term Clouds – Term clouds are a great way to visualize the relative importance of words. It can also spice up an ordinarily dry Analytics presentation. This option is especially useful for search terms people used to find your site, which is why I suspect it’s only available in the Traffic Sources area (Remember to exclude “(not provided)” and branded terms).
You can find the Term Cloud function by clicking this button:
when you are on a Traffic Sources report. To see more words, change the results per page drop down. (more…)
View Comments (3 Responses) | Categories: Google Analytics
Posted on July 19, 2012 by Phil Anderson
Companies often come to us to correct their Google Analytics configurations. The problems we encounter vary from the very simple to super complex, but by far the most common problem I see is the “self-referral” configuration problem.
(more…)
View Comments (17 Responses) | Categories: Google Analytics
Posted on July 3, 2012 by Phil Anderson
It seems like every time I login to Google Analytics these days, something has changed. Google is really following the “iterative” in Iterative Development lately. Some of these changes are major, and get full-on blog posts, and others Google seems to just sneak into GA in the middle of the night. Here’s the ones I could find, comment below if I’ve missed anything.
The Big Ones
Website Optimizer is dead, Content Experiments are the new A/B Testing. We covered this in a lot of detail in that linked post, but it’s worth reiterating what a big change this is, and I’ve already noticed much of the web using these (you see the telltale utm_expid query parameter).
(more…)
View Comments (7 Responses) | Categories: Analytics, Google Analytics
Posted on June 12, 2012 by Phil Anderson
Last Friday I gave a talk on Getting Analytical with Google Analytics at the Netroots Nation event in Providence. Here is a summary (with links to resources) for those that attended or could not make it.
If the way you use Google Analytics is to just look at general traffic trends in the Visitor Overview area, you’re seriously under utilizing the abilities of this tool. Here are 5 things you should definitely be doing in Google Analytics right this minute!
Profiles
Accurate data is one of the most important starting points for the person serious about doing data analysis. Profiles are the best way to segment out different types of traffic. Here’s a great example, what happens if you have an organization where everyone’s browser opens to your homepage by default?
Well mix your internal numbers:
With your external numbers:
And you’ll get a really skewed number of your true engagement! So before you do anything else, create two new profiles with just internal and just external traffic! ALWAYS REMEMBER TO KEEP AT LEAST ONE COMPLETELY UNFILTERED PROFILE! Also consider adding profiles for different portions of your site if they bring completely different types of traffic (Blog vs Non-Blog for example).
Read more about Profiles in Google Analytics
View Comments (1 Response) | Categories: Analytics, Google Analytics
Posted on May 29, 2012 by Phil Anderson
You use Motion Charts in Google Analytics, don’t you? Wait, you don’t know what I’m talking about? This guy, with all the pretty colored balls?
No? You’ve never even heard of it? Well it’s not surprising that most people don’t even realize this feature exists. Until the most recent UI update to Google Analytics it was hidden in a small icon you wouldn’t have noticed unless you were looking for it. They’ve made it a little more obvious now, and you can get to motion charts by clicking the “three circle” icon
(more…)
View Comments (4 Responses) | Categories: Google Analytics
Posted on April 26, 2012 by Phil Anderson
The final part of the Infinite Conversion Loop is A/B Testing. A/B testing is like exercising, everyone knows they should do it, it seems harder to do then it really is, and once you do it you wonder what took so long to get started.
First, an introduction. The concept of A/B testing is as old as marketing itself. An A/B test is simply showing one variation of a web page to 50% of your visitors, and another variation to the other 50%. You then measure what is more successful, and that one is “the winner”. Slightly more complicated is a multi-variate test. In this case, various specific parts of the page are changed, maybe a different picture shown, or the text justified differently, and every possible combination of the various parts are alternated to find the absolute best combination.
(more…)
View Comments (1 Response) | Categories: Conversion
Posted on April 5, 2012 by Phil Anderson
This is part five of a multi-part series on The Infinite Conversion Loop. If you have not read the previous articles, you should check out Introducing the Infinite Conversion Loop, 10 Things to Check in Your Google Analytics, Who are these people visiting my website?, and How to Legally Spy on Your Visitors.
Last time we talked about Passive User Testing, that is, examining user behavior without their knowledge or involvement. Passive user testing is great for finding large problems, but it’s hard to get the “why” from this method of testing. For example, you may notice everyone is abandoning at the first step of the shopping cart, but if there is not a JavaScript error or something obvious, it may be hard to fix this problem.
Active User Testing is the best way to get more answers. It’s more expensive then passive user testing, which is why I recommend it only after you’ve fixed the real big problems. You really don’t want to waste money on people telling you the site doesn’t work at all in Internet Explorer (which can easily be identified using automated means). Active user testing involves (usually) paying people to use your website and verbalize what they are doing.
(more…)
View Comments (No Responses) | Categories: Conversion, Usability
Posted on March 8, 2012 by Phil Anderson
Last week, most of us logged in to a big surprise in our Google Analytics, and it wasn’t a traffic spike, it was the new, new look:
Whoa! Google Analytics, you lookin’ fine.
Now that we’ve had a week to digest the changes, here’s a few of our favorite things that Google has added in this substantial update.
- Sharing of Custom Reports and Dashboards has never been easier.
I think this most affects people who work with clients or have teamwork among analysts, but if you’ve ever wanted to share a custom report with someone before, it was a pain (usually ending up in the wrong profile), and the ability to share Dashboards was just nonexistent. The new interface lets you choose what profile to apply the dashboard/filter easily.
- Less flash!
You can now view Google Analytics on your iPod/iPad and see the line charts. I suspect this also explains why GA feels a little peppier. They are most likely phasing out all flash in favor of HTML 5. (A peek into the source code shows that graphs now come from an Iframe which is rendering things with scalable vector graphics)
- Toggle Buttons are More Clear.
A bunch of features that have been in GA for some time are more apparent now, thanks to a clean new layout of buttons. For me, the Compare Metric and Plot Rows options are much more obvious now. I also really like the Day/Week/Month toggle, it seems cleaner and is definitely less buggy then the old drop-down.
- Sampling Slider.
While you still can’t turn off sampling (unless you use GA Premium), at least it’s possible to show your customers that sampling can be pretty accurate (or inaccurate) depending on the amount of sampling. I’ve played around with this to see what percentage Transaction data gets the most inaccurate. I’m not sure why anyone would want .02% of their data sampled in exchange for speed, but it’s possible now.
- Pretty Thumbnails.
This small bit of eye candy makes the sections in GA really pop out, and makes the overall tool look a little more exciting. Although Daniel Waisberg makes a great point that the $ symbol should be by goals instead of Advertising, but I guess Google doesn’t see it that way!
Still using the old GA? We’re offering a free online webinar to help with the transition. Sign up now, we’re limiting this webinar to the first 90 people.
View Comments (1 Response) | Categories: Google Analytics
Posted on February 20, 2012 by Phil Anderson
Welcome back to the Infinite Conversion Loop. Hopefully since the last post, you’ve not only got your Analytics accurate, but also are getting more visibility into who is visiting your website.
Have you ever found yourself staring at your Analytics data, wondering “What the heck is wrong with these people on my website? Why would 10,000 people come to my site, but only 100 of them decide to purchase anything? I mean my site LOOKS awesome, and I’d give me money.” Unfortunately, we can’t actually view what people are doing without resorting to something illegal… or can we?
ClickTale
Enter ClickTale. ClickTale records a sampling of your visitor’s behavior for an entire session using JavaScript to detect mouse movements, clicks, time on page, etc… it’s able to do this in such a way that it really feels like you are watching a video of someone using your website. ClickTale is pretty pricey ($99/month for the lowest level), but I think running it at least for a month or two will give you all sorts of insight you just can’t get looking at the numbers.
For example, one thing I always wanted to understand was the low usage of the advanced search on my site. I saw in analytics that few people were searching with these advanced options, but I didn’t understand why. ClickTale gave me the other piece of the puzzle, people were TRYING to use the advanced search, but getting confused by the options. As I watched them click radio button after radio button, only to click the Back button 90% of the time, the UI problem became clear.
Another benefit to ClickTale is it also sort of does your cross-browser testing for you, quietly in the night. It logs JavaScript errors, and lets you search on them, so you’ll discover the JavaScript error that you didn’t realize was occurring on IE7 or Firefox 4 for OS X.
The only frustrating thing with ClickTale is that you’ll see people who are clearly confused, but you can’t ask them why!! (I’ll cover this in the next post).
Google Analytics
Since we’re big fans of Google Analytics here at LunaMetrics, if ClickTale is out of your budget, another great option is available to you now in GA.
One of the best new features Google has added recently was Flow Visualization in Google Analytics. Jim does a great job in the videos on that link explaining the basics, so let’s put them to work in figuring out what people are doing.
When you go to the Visitors Flow report, it may be confusing at first. I believe the default view is by Country, which probably isn’t useful for the majority of you out there. I typically choose by “Medium”, since that’s usually the most interesting for understanding how different types of visitors flow (since organic search comes in at the index, affiliates go to the deepest level, and cpc goes to landing pages).
Quickly you can see some interesting things, a quarter of my affiliate traffic is going to 404 pages! Also, almost all of the organic and referral traffic is going to the index of the site.
Keep adding steps on the right until you see all the large patterns. Don’t get too hung up in the outliers that only have a few “connections”. Look for the trends… are people clicking to search and then going back to home? That probably shows they can’t find what they are looking for. Does it seem like an abnormally long set of steps to get to the Checkout? Maybe your sales funnel can be optimized.
I’m sure some of you might be thinking “What about GA’s In-Page Analytics?” Well, it’s there…
But because it doesn’t differentiate between clicks on different buttons (any links with the same URL will be counted equally), I would only use this for the most barebones analysis.
All of this is just the first pass at UI changes to increase conversions. Use these tools to solve the big problems preventing users from converting, next time I’ll show you how to get to really get down to the heart of things.
View Comments (2 Responses) | Categories: Analytics, Conversion, Usability
Posted on January 23, 2012 by Phil Anderson
This is part 3 of a multi-part series on increasing conversions from your website traffic. If you haven’t already, you should read part one, which introduces the Infinite Conversion Loop. And Part Two which gets your Analytics in order.
Now that you have high confidence in your Analytics, let’s discuss the next step of the Infinite Conversion Loop, identifying your visitors. The first thing to understand when you are trying to improve your conversion rate is that not all visitors are created equal, which is why simply looking at your overall traffic numbers is an empty metric.
On any site you are going to have a wide variety of visitors that all want and expect different things. Here’s just an example of a few of the types of people you may see if you run a website for used car shopping, but these same type of people exist for many sites.
The Cast
- Mr. Shotgun – This guy isn’t great at searching, just average. If he wants to buy a new 2012 Audi A6 with leather, his query is probably just “audi”. Then he clicks on the first search result he gets without reading, hits back, clicks the next one, hits back, until he finds something that kind of meets his expectations. These are the people that end up on page 5 of Google result and kill your bounce rate if you don’t have a clear message.
- Miss Untargeted – This is the visitor you get from Reddit, Hacker News, Drudge Report, etc… when you post a great blog. They give you that big awesome spike in traffic that never seems to result in increased conversions. They make you feel good about your traffic numbers, but really they have no intention of converting on your website. They also make your site-wide conversion numbers useless, and cause Sys Admins to pull their hair out.
- Mr. Confused – This person is running IE6, unpatched, he’s looking for exactly what your website sells, but can’t figure out how to find it on your site, and when he does he don’t trust you because you don’t have the Visa or BBB logo.
- Señor Advanced – This person found your website with a query like: used -new acura tl -reviews “low mileage”. Your website is exactly what they’d want, but your refinements don’t let him find what he’s looking for in an efficient way so he leaves.
- Mrs. Researcher – She got to you by clicking your ad, she’s looking for what your site sells, and she even finds product she wants, but wants to make sure she gets the best price, so she’ll bookmark your site for now while she looks at some other sites.
- Botman 9000 – Some percentage of your traffic will be bots, spammers, or others who have no intention of using your website like a real person, let alone buying something. While the bots won’t usually show up in your analytics, they will often submit forms and do other things that makes your “back end” numbers never quite match up with the Analytics numbers.
This cast are the reason your overall visits mean nothing. These people represent the 98% of people you get daily but are not converting. I could give you a million hits today, but if they are the people in this group, you’ll see $0 out of it. The key to improving conversions is first getting a handle on what type of visitors are likely to convert in the first place, and then optimizing your site to convert those that are not converting now for some reason (some of the above have the potential, but your website has deficiencies that prevent them from doing so).
The Numbers
So first, let’s figure out who IS converting on your site. The first thing I would look at is your landing pages report, ordered by visitors. Specifically the Bounce Rate (The percentage of people that visited that single page then left).
What do these stats mean? First, these are from a profile I use that broadly groups types of content into single URLs. This site has over one million URLs, and I find most useful to categorize everything into types of pages (you can do this with filtered profiles in GA, which is probably the topic of another post). You can see the Index page of the site is how most people are entering by far. 22% bounce immediately, meaning something about the home page felt completely wrong for whatever those people were looking for. We’ll examine how to improve that number in future posts. You’ll notice a few popular blog posts have much higher bounce rates. These are the “Miss Untargeted” that are coming to your site for one-off content, but are probably not interested in buying whatever you are selling. It is interesting to note that one blog only has a bounce rate of 50%, indicating there is something about that post that invites people to dig further into your site.
Next I would look at the Exit rate (The percentage of people that left on a given page). The Exit rate is useful for identifying common pages that people are leaving your site on. This often points to problems on the page or shows you that people are not finding what they want (for example if a search result page is a common exit point). I order this report by Exit rate, filtering out outliers first.
Next look at where your most valuable traffic is coming from. I would do this by going to Sources->All Traffic, clicking ECommerce at the top, and then sorting by Ecommerce conversion rate. In this case, once you take out a few outliers, it looks like search engines (both paid and organic) are giving us the most valuable traffic. Another case of how blogs, social networks, etc… will increase overall traffic but not necessarily contribute to the bottom line (these things may be valuable for SEO though, so I’m not saying don’t do it, just don’t focus on overall traffic numbers as a KPI)
Of course it’s very difficult to read people’s intentions just by looking at numbers. The only real way we can try to find their intent is by looking at what they searched for to get to us (hurry before the number of logged in Google users increases!). Create a report of Organic Search Traffic ordered by Ecommerce conversion rate to see what terms are the most valuable. You’ll usually see some good long tail keywords you never would have considered. I would also sort by 0% conversions to see the keywords you may be wasting your time with. Remember to normalize these results by filtering to see keywords with at least 10 visits or so.
Those are some beginner ways of segmenting users or figuring out what they are trying to do on your site, some proactive ways to segment is the use of Custom Variables. Most people use these to track “logged in users” or what affiliate someone came from, but what about tracking those people who found no search results? That’s what I did here, and there is no surprise there is $0 revenue from those users… perhaps an opportunity to use spelling correction or show similar products.
Using Custom Variables is a great way to create segments of content and visitors to determine easily what types of users, content, etc… leads to the best conversions.
Hopefully using these tips will get you started in figuring out where your traffic is coming from, the next post in the Infinite Conversion Loop series will delve deeper into actually seeing what people are doing on your website.
View Comments (No Responses) | Categories: Analytics, Conversion
Traffic
Analysis
Action
Subscribe to this blog's feed