While Google Analytics is so widely used, most marketers only scratch the surface when it comes to the reports they use. You can find insights for conversion optimization from lots of different reports & some of the juicier reports are lesser known.
I asked some of my friends in the industry to share under-utilized reports they often turn to when looking for insights.
#1: Advanced Segments Based On Buyer Personas
Brian Massey, the Conversion Scientist:
One of our more unusual Google Analytics reports is the buyer persona report. We create four advanced segments and see what their purchase rates are.
This report is built by taking the “site average pages per visit” & “site average time on site”.
We then create four segments:
- Methodical Marys visit the site for longer than the average and view more pages than average. They are taking their time to make a decision.
- One-hit Juans visit few pages, but stay on the site for a long time. For sites with video, there may be a lot of One-hit Juans.
- Lost Lucys hit a lot of pages in a short time. They seem to be looking for something and not finding it.
- Bouncy Bobs visit fewer pages than the average and spend less time than the average. Bobs typically include the least qualified traffic on the site.
Once we have the segments, we can examine their purchase rates, page flows & devices used.
It gives you the first segmentation approach as you begin to understand the visitors to a site.
#2: Network reports and motion charts
Yehoshua Coren, Analytics Ninja:
Here are two examples of some lesser known, or less often used, reports in Google Analytics.
In both of these examples I will show how they can be used first hand with very clear action items to take.
The first report is the Network Report – specifically, we’re looking at the Service Provider dimension.
The Service Provider is simply the name of the Internet service providers (ISPs) used by visitors to your site.
7 Ways that Predictive Analytics Is Transforming Ecommerce
Predictive analytics help you understand what your customers are going to buy before they do.
Something to notice about the example above is that Amazon and Microsoft are showing up as ISPs that have very high bounce rates.
This is just one example of how Service Provider report can be used to improve data quality.
What I find most useful about this report is the ability for B2B companies to gain insights with regards to their leads or prospects (could be applied to lead scoring within a CRM).
For example: The site we’re looking at is in the healthcare industry. As such, I can quickly see what other healthcare related businesses or organizations are browsing the site with some simple regex.
The good folks at LunaMetrics have a post where they suggest excluding the following regex in order to have the “real named” corporate ISPs float to the top.
After you’re done reading here, take a look over there for some simple but very useful suggestions around leveraging GA in a B2B environment in this regards.
The other “lesser-known” report is really an aspect of many different reports within Google Analytics; namely, Motion Charts.
While I suppose I should really use a moving GIF to describe what this can do, in short Motion Charts are great at visualizing where specific change may have happened. Phrased differently, motion charts visually help segment data much better than the standard charting in Google Analytics.
The following example is exploring e-commerce conversion rate over time segmented by browser type.
Since Firefox comprises about a quarter of the website traffic (even more when you take into account that a third of the websites traffic was bot related), I am tuned in right away to what appears to be a cross browser compatibility issue that is impacting the sites conversion rate.
This would lead me to explore funnel abandonment and exit pages for the Firefox browser segment before and after week 42 of 2013. I would then have insights about what the potential problem might be to then passed back onto the site’s developers to fix the issue.
Yehoshua Coren is the Founder & Principal of Analytics Ninja LLC, a pure play digital analytics consultancy that provides advanced strategic implementations and analysis utilizing Google Analytics and Google Tag Manager.
#3: Integrating Split Test Data with GA Custom Variables
Michael Aagaard, ContentVerve:
Getting insights from Analytics during your research phase is crucial to identifying optimization opportunities and developing solid test hypotheses.
But it’s just as important to stay on top of things while you’re running your A/B test experiments. Moreover, it’s vital that you are able to get the full picture after you conclude your tests.
As Peep says, “Averages lie”, and if you are looking only at overall conversion rates in your testing tool, it’s safe to say that you are not getting the full picture.
A number of factors impact the decisions and actions of your potential customers, so it’s risky to assume that they will all display uniform behavior.
- Have they visited your website before?
- What was the traffic source?
- Did they find your website via a campaign?
- What day of the week did they visit your website?
- Did they use a laptop, a tablet or a mobile?
- What was the browser version?
If traffic from one particular source or device behaves significantly different than the rest, it’s enough to screw up your test data completely – and thus also what you’re taking away from the test; in terms of insight as well as revenue.
In cases like this, it’s crucial that you can identify the anomaly and take it into consideration before you make any conclusions about what to implement permanently on your website.
Some testing tools offer a degree of segmentation, but it’s nothing compared to the level of insight you can get via Google Analytics if you make sure to integrate the data from your test variants in your GA account.
By sending your test data to GA as custom variables, you can do in-depth segmentation, run custom reports and get the full insight that your analytics setup has to offer. This means that you get the full picture of how your test variants affect potential customers and their behavior on your website.
Here’s an example of a VWO experiment where the data has been collected under custom variables.
As you can see, factors like bounce rate, page visits, total revenue, and ecommerce conversion rates for the individual test variants are displayed offering a deeper insight than you’d get from you test tool by default.
In this next example, we’re digging into Variation 1 in order to compare performance across devices.
Notice that bounce rate is significantly higher (and ecommerce conversion significantly lower) via tablet than desktop and mobile.
And here’s an example where we’ve delved into performance across different browsers.
If you use Optimizely or Visual Website Optimizer, setting up your integration with Google Analytics is super simple.
However, if you use a custom implementation of GA, you might run into some issues.
#4: Site Speed Reports
Craig Sullivan, @OptimiseOrDie:
I’m a performance junkie – because I started life as a network and security engineer with routers, TCP/IP and lots of cool tools for analyzing and fixing problems.
I carried this into my life as a UX and CRO practitioner, because the performance (and perception of performance) plays a vital part in how people rate their experience. Aside from the pain of sucking millions of hours of productivity from people’s lives as they wait for your stuff to load – you’ll also make lots of money by fixing performance issues.
In the past, people used to use synthetic tools – bots placed on the internet which would measure your site. Now you have something much better. Real user performance monitoring of how long the pages really took to load and render. Where? In Google Analytics of course.
In Google Analytics, go to the ‘Behavior’ section first. Choose ‘Site Speed’ and then ‘Page Timings’.
At the top of the report, click on the ‘DOM Timings’ tab.
You can now see 3 metrics called
(1) Avg.document interactive time’
This is the time it takes the browser to start handing control back to the user. This metric is really important because it’s when people can start navigating, clicking and scrolling.
(2) Avg. document content loaded time
Depending on how your website pages build, you may need to use this measure instead. It’s a technical thing but basically this measurement refers to running scripts in the page. If you’re using stuff that blocks a lot of user interactions until all the scripts are loaded, this may be a better metric. I tend to look at both of these.
(3) Avg. page load time
This is a less useful metric these days – this is the ‘entire time’ to load everything on your page. However, this gets skewed by all the 3rd party tags, tracking stuff and social buttons you might have on your site. It simply isn’t reliable for real performance metrics (IMHO) but it does tell you if you’re loading heaps of crap on slow connections.
By focusing on these metrics and then creating yourself a ‘suck index’ report, you can start to chip away at things. You can create a ‘suck index’ by simply ranking each page by page views * average document interactive time. That way the slowest and most used pages will bubble to the top.
Running performance reports around the pages that really suck in delivering fast load time will do wonders for your bounce rate, user delight and your conversion rate. If you want to prove all this to the boss, the resource list below has everything you need.
Some resources on performance and materials to convince your boss:
- The mother-load of tools and tips : https://developers.google.com/
- New findings about page views, time on site & bounce rate for desktop and mobile browsers http://bit.ly/yu2z6D
- The “performance poverty line”: What is it and why does it matter? http://www.webperformancetoday
.com/2012/06/05/web- performance-poverty-line/ …
- Case study: The impact of HTML delay on mobile business metrics http://www.webperformancetoday
.com/2011/11/23/case-study- slow-page-load-mobile- business-metrics/ …
- AB testing slows your site down : http://blog.convert.com/ab-
- 4 slides showing how page speed correlates to business metrics at http://Walmart.com http://bit.ly/xhsAjz
Really Excellent Tools for diagnosis:
- Performance insights from Google : http://developers.google.
- For Desktop and Tablet experiences : www.webpagetest.org (Note : Please use the timeline photo feature – it’s ace)
- For Mobile experiences : http://mobitest.akamai.com/
#5: Flow Reports
Judah Phillips, Vice President Analytics and Data Sciences of Karmaloop shares his go-to lesser known reports:
Google Analytics has an inordinate number of reports, both standard and custom, that when considered as an abstraction of your customer behavior can be overwhelming to interpret.
In fact, many GA users don’t examine all reports because they either don’t think they are necessary, worthwhile, or relevant to their business goals.
While that opinion may or may not be true, there are several standard reports in GA that are “lesser known” – and thus used less frequently, which can be helpful to the analyst.
These lesser known reports yield additional insights if they are understood properly and in context. On such example requires a bit of history to understand: in traditional “web analytics” the notion of a “path” – that is the discrete series of clicks (i.e the clickstream) taken by a visitor during a visit to a web site – became somewhat disdained as not very useful about 8 years ago.
Even then, the nascent analytics industry was criticizing pathing reporting because constructs like identical links on the same page couldn’t be pathed accurately. Rich internet experiences (where there was no page view) were becoming popular to replace static web sites, and pathing couldn’t capture these events. GA addresses these concerns and more in “Flow” reports.
While GA doesn’t name it’s “Flow” reporting as “pathing,” the concepts, the presentation, and the utility of the reporting is reminiscent certainly of traditional pathing. Yet, GA’s recasting and evolution of “pathing” as “flow” has made this type of data representation and visualization more usable and helpful. As such the following “Flow” reports make my cut for usefullesser known reports that can be insightful, especially when combined with custom segments:
- Visitor Flow is another representation of traditional web analytics pathing with more features in it. This reports lets you see where visitors came from and what page was their next and subsequent interactions. You can hover over the page for more information – and select to view the flow from a particular page. A nice feature is the ability to segment the flow by available dimensions and custom segments. This report is helpful for marketing analysts.
- Behavior Flow is also a representation of where visitors started on the site and where they went next. This report is pre-configured to begin with landing pages (first page in visit) instead of an acquisition source (like Visitor Flow); thus, it is helpful for informing landing page optimization. It is helpful for those analysts concerned with verifying your audience is viewing the best possible content.
- Goal Flow is almost identical to the Visitor Flow but contains the goals you have configured as part of the conversion funnel. In this sense Goal Flow is similar to the Funnel Visualization report; however, it provides more information and enables deeper exploration. Hovering on a goal also provides additional information (like in all Flow reports). This type of flow is helpful for eCommerce Analysts and Optimizers.
- Event Flow is represented like the Goal, Behavior, and Visitor Flows but it represents those Events you have configured in GA to represent page-less events within an interaction. This Flow report is particularly helpful when you want to see how people interact with a rich application that doesn’t generate page views when you have implemented and configured it using GA Events. This type of Flow is helpful for Product Managers.
Also I think the Multichannel Funnel Analysis reports are helpful – and lesser known. They have some complexity in both concepts and data interpretation, which I covered in Peep’s other post.
#6: Site Search Insights
Justin Rondeau, WhichTestWon:
The ‘Site Search’ section in Google Analytics has a lot of information and will identify major areas of opportunity. This can be useful for really any site across any industry insofar as their site search is used often.
WhichTestWon doesn’t get a lot of search queries so I don’t gather a ton of info here, but another Anne Holland Ventures Publication gets a lot of queries and helps the content team create content people are looking for and the development team make it easier to get to the content people want.
On the ‘Overview’ section, select ‘Start Page’. You get the top 10 pages where search has occurred. People who search for content or items on your site are more likely to consume that content or purchase a product if they can find it on the page 1 results.
What is great about this very simple report, is it lets you know if you are missing critical pieces of content on that page, especially if you notice search trends.
To notice the search trends select ‘Search Terms’. This displays all the search terms that have been typed into your search box. Here I like to do two things:
1) Select ‘Start Page’ as a secondary dimension again to see the keyword next to the page that spawned the search.
2) Select ‘Exit Page’ if they left on a search results page, I know they couldn’t find what they needed. On the other hand I can see what pages people did end up navigating to if they didn’t exit.
I highly recommend making a custom report to anyone with a significant amount of search queries that includes the ‘Start Page’, ‘Exit Page’, and maybe even a ‘Conversion Goal.
#7: Content Grouping
Tim Leighton-Boyce from CXFocus:
This year I’m starting to get really excited about GA ‘Content Grouping’. I used to do this many years ago using view (profile) filters to rewrite the URI of pages in GA into groups which made sense to people working on the site.
Although this was valuable, it was also difficult to do effectively and so I found myself using the technique less as the years went by.
The new built-in features for doing this in GA are far more powerful and flexible. And even if I most often have to resort to using daisy-chained hard-coded rules to get the structure which works for the client, the effort is worth it.
The reason the effort pays off is because these functions now appear in so many places in the GA Behavior (content) reports, and I suspect they will appear in more soon. I dream of the day the groups appear in the Flow Visualization reports because the automatic grouping in those reports is so frustrating.
Here’s an example of a ‘grouped’ view when applied to a Navigation Summary report to illustrate just how useful grouping can be. The Navigation Summary used to be good for deep-dive forensics on a particular page, but with grouped content it starts to make aggregated user-journey reporting less of a route to wasted hours.
This feature is built right in to standard reports — that’s why I’m so pleased as this addition for those of us who are more interested in what people do on web sites and not only concerened with getting more and more people to the site.
When you’re looking for insights in the analytics, be ready to put in the hours. Whenever I start to analyze a new site, I easily spend 2 full working days in their analytics. Reports mentioned here will help you find the insights you’re after.