I just love a good mystery, and to be candid, I love being the one who gets to solve it! Solving mysteries and putting together the proverbial pieces of the puzzle is a critical skill in the field of web analytics. You almost have to like the torture that comes with trying to figure out a problem, in a weird and demented way.
So when my industry colleague Matt asked me on Twitter to help him solve his Google Analytics quandary, I was ready in a nano-second.
You can read the full post here, but essentially, Matt needs to know what the best way to “isolate” page data would be. He has a sub-directory on his web site, which include pages, and needs to be able to create a segmented, sliced-up view(s) of that sub-directory, and needs to be able to view how each sub-directory’s pages are performing in relation to other sub-directory pages.
Creating a duplicate, filtered profile for all of this sub-directory’s traffic within the same Google Analytics account (using the same website domain) will create your isolated view of only those sub-directory pages. You will only see visits and page views that happened on those sub-directory pages. It’s good for looking at your sub-directory data in a silo, and you can compare the high-level data by using the profile overview screen (assuming you are planning on creating additional filtered profiles for the other sub-directories). You can also download the data offline and mash it up, either via the Google Analytics API or by simply downloading PDF or CSV files.
Creating an advanced segment that displays any pages that match your sub-directory name will show you any visits which included at least one page view on any one of the pages within that sub-directory. This definition – visits instead of pages from the previous paragraph – is an important differentiation. As commenter Amanda has already astutely observed, you will see other pages appear in your Content report section, because this segment will show you those other pages, as they were a part of these visitor’s sessions that viewed at least one page within your desired sub-directory. You can create an advanced segment for each sub-directory and compare up to three (plus the “All Visits” segment) at the same time, and get an on-the-fly look at your sub-directory data. However, if your date-range is long, you may encounter data-sampling (not the biggest issue in the world, but something to be aware of).
If you create a Custom Report, in your main profile and without any advanced segments applied, you will be tailoring an original view of your data. You can combine metrics from different reports, like visits, bounce rate, goal start and goal completion percentage, and revenue / ROI metrics (if you do Ecommerce). You can then match it up with the page dimension, and even set it up so that when you click on a page, the report will show you the keywords, or the source / medium combo, or the visitor country, or whatever drill-down dimension you want to see. Then, if you really want to get fancy, you can apply an advanced segment while you are looking at your custom report to show you visits that have viewed at least one of you desired sub-directory pages, and really get cooking! You can then apply a custom report and an advanced segment to multiple profiles from within the main profile (Click on the respective “manage” links), and apply it to any of the other profiles within your account.
So, what would I do? I would create a custom report with an advanced segment applied to it. You can also create a filtered profile if you wish, but I would suspect you would not use it as much as you would a custom report / advanced segment combo. I would also insist that your report is meaningful and that you can take action from it (e.g. knowing that a page’s $Index value is a lot lower than the site average would point you in that page’s direction to optimize / refine it). Pick metrics like Bounce Rate, $Index and Goal Conversion Rate that help you understand page performance, and ditch trivial ones like Avg. Time on Site or Exit Percentage.
Hope I helped out Matt and others in a similar situation!
Yesterday I asked the following question on Twitter:
I need some ideas for my weekly [Web Analytics] blog post. Anyone?
To which Twitter user nedeniarecel replied:
What about custom reporting to measure entrances and bounce rate metrics by geography and time of day?
Excellent! I want to show nedeniarecel (and you, of course) how to set this up, but I also want to discuss what to do with this report once it’s set up. That is, the insights and decisions that we take and make from this new custom report that we are about to create. A report is only as good as the insights it can provide.
I’m going to use Google Analytics to show how to set this up, but regardless of what tool you’re using, you should be able to apply the same principles that I have here.
To set up this Custom Report,
1. Log-in to your Google Analytics account, select your profile of choice, and click on Custom Reporting from the left hand navigation
2. Click on Manage Custom Reports, and then on the subsequent screen, click on Create new Custom Report on the upper-right.
3. Give your report a name, and, if you want to, give your tab a name. Drag and drop metrics / dimensions from the left-hand side into the areas of the report on the right-hand side. Your report should look something like this screen shot:
4. Click on Save at the bottom of this screen and you’re good to go!
Or – are you good to go?
This kind of report will give you some good data by itself. We have entrances by country / territory (you can do region or even city level geo’s instead if you wish), and it will provide you the bounce rate for each country / territory – which is the percentage of single-page visits from each location.
You can even click on any country and get a breakdown of entrances and bounce rate by the hour of the day.
What is important to do at this point is to ask ourselves whether or not this report can answer a question, add a specific value, or, allow you to extract a valuable, usable insight. For example, can you take an action from this report on your website or on your marketing efforts? Is this report adding a value that you didn’t have before? Can someone make or suggest a decision with this report?
If you don’t feel that you can, then I would suggest a few modifications / additions to this report, while maintaining the integrity of the request. I would do something along the lines of this screen shot:
In a manner of speaking, I “pimped out” the original custom report to include:
1. An additional site usage metric, Pages / Visit, gives me a pinch more insight on average on how deep their engagement was.
2. Three goal-oriented metrics: Total Goal Starts, to see how many visits started the goal completion process; Total Goal Completions, to learn how many visitors completed an important action on my website; and Goal Conversion Rate, to obtain the ratio of visits to goal completions.
3. Two final metrics, Search Exits and Event Value. These last two give me clues as to the performance of both my website’s search function and any non-page view Events that I am tracking, like videos or PDF file downloads. Search Exits represent the percentage of visits that leave your website after performing a search – like a bounce rate for your website’s search function. Event Value is the total value of all events.
4. A new dimension, Source / Medium. This way I can get an informative breakdown of how each source of traffic + each medium viewed and performed on my website, and then I have the option of clicking on any individual source / medium to view a country / territorial breakdown. If I want to go deeper than that, I can click on any country / territory and get another breakdown by hour of the day.
Now, I can take this data for any source and know what I need to do next, because I know what country visits came from, what their bounce rates and conversion rates are, and what these visits thought of my search function, events, or PDF files. I know if I need to work on my SEO if my organic traffic sources aren’t matching my performance expectations, or if my pay-per-click efforts are doing fantastic and if I need to allocate more budget there. I can see if my E-mail newsletters tanked or if my direct mail piece was a smash hit.
I can do all of that from that custom report. The great part about custom reporting is that there are thousands upon thousands of different combinations that you can put together, just as I did above. What are some of your custom report combinations in your favorite web analytics tool? Leave a comment below to share your thoughts!