AAA's webmaster Sara Snyder distributes our monthly Google Analytics report; I have attached both. I am also cutting and pasting here from her last email that accompanied the reports. Sara has devoted a great deal of thought to this and I’m sure she would be happy to share her insights.
All data in this report comes from Google Analytics. Traffic from within the Smithsonian network has been filtered out.
Summary Analysis of Web Traffic for April 2010
The excellent presentation by Seb Chan from the Powerhouse Museum on April 20th has really been inspiring me to try and take better advantage of the advanced audience segmentation tools in Google Analytics. For example, did you know that we can segment visitors not just by country, but by state or even city? By what kind of computer, mobile device, and browser software they have? By how they found us, by what pages they looked at (or didn’t), or by whether or not they chose to use our site search box? And (best part) that we can combine any of these facets to create as many custom segments as we want?
To illustrate the utility of this, for this month’s report I want to focus on an audience segment referred to by marketers as branded traffic. These are people who already know about AAA: they either have us bookmarked, typed our URL into the address bar directly, or else searched for us by name (“Archives of American art” or a highly similar variation). In addition to our regular report, I have attached a summary report of the top content for April showing the branded traffic segment.
On average, these visitors spent more time on our website (3:32 compared with normal average of 2 minutes) and viewed more pages per visit (5 pages average, instead of 3)—in sum, they are just more into us. While branded traffic accounted for 21% of our total pageviews, it made up 30% of our pageviews within Collections, 36% within Collections Online, and 34% of pageviews within the “Support the Archives” and Membership pages. Branded traffic hit two notable peaks in recent months, directly after AAA’s exposure on the BBC website and in the Apartment Therapy design blog, neither of which included a link to our URL but which mentioned our name repeatedly (hence the need to Google us). Tracking branded traffic visits over time may provide us with the ability to measure the effectiveness of offline advertising/PR initiatives, as well as the growth of a loyal visitor base when examined in conjunction with repeat visits.
Blogs and Social Media
We gained 63 new fans on our Facebook page this month, more than double the 27 we gained last month, thanks to the “find us on Facebook” button that was added to the Primary Source e-newsletter. We had 513 non-SI visits to the Archives of American Art Blog, also up from last month by 27% thanks to a referral on April 13 from Stumbleupon.com, the personalized website recommendation tool.
It seems like there is always more to learn about how we can better use our web traffic stats to gain insights about our visitors. As always, I’m happy to work with you and your department to figure out how these numbers can support you in your work.
Barbara D. Aikens
Chief, Collections Processing
Archives of American Art, Smithsonian Institution
email: [log in to unmask]
Archives of American Art, Smithsonian Institution
PO Box 37012
Victor Bldg., Suite 2200, MRC 937
Washington, DC 20013-7012
From: Encoded Archival Description List [mailto:[log in to unmask]] On Behalf Of Jennie Anne Levine Knies
Sent: Wednesday, June 23, 2010 6:53 AM
To: [log in to unmask]
Subject: Re: Google Analytics
I think your suggestions are very helpful, as have been other comments on this topic. Thank you - something as seemingly simple as putting a brief explanation of a finding aid at the top of the page may have huge benefit. Yes, I understand that when it says that 58% of our users come from Google, it means that 58% of our users are coming from a Google search! However, you interpreted my comment correctly in that I was wondering "What does that tell me about how users are accessing our finding aids and what can I do to make them more usable for that large audience?"
Another interesting tidbit - in the last month almost the same number of people came to us from our online catalog as did from Wikipedia... and it wasn't many in either case. That said, taking three minutes to look into this explained to me why one of our more "obscure" collections keeps getting hit - it's because of the wikipedia link that someone (not me) entered.
I would really love to have the time to think about this more. But am grateful to the thoughts already presented in this thread, and am hoping that, combined with Chris's article, can get me motivated to focus on it more...
Jennie Levine Knies
Manager, Digital Collections
2216 Hornbake Library
University of Maryland
College Park, MD 20742
Email: [log in to unmask]
Sent: Tuesday, June 22, 2010 4:01 PM
Subject: Re: Google Analytics
In answer to your question "What does it MEAN?" I have several thoughts.
The first is that this is exactly the kind of situation where
collective power would be far more effective than individual
institutions attempting to dredge up resources to analyze their data.
You went to the effort of implementing analytics and you've got 3
years of data, but you may not have the time or staff power to conduct
analysis on your data. There are many others out there who possess
data, but who do not have time/staff/data analysis skills to conduct
meaningful analysis. For each person out there with data but no
analysis, there is someone else with the time/money/staff/data
analysis skills, but not the data. There would be enormous benefit if
we as a profession had a centralized location where data could be
freely shared. Not only would this encourage collaboration and reduce
time spent reinventing the wheel (in terms of people trying to figure
out if someone's already collected data on a specific topic that the
want to analyze), but it would encourage cross-institutional data
analysis, producing more meaningful findings for the community that
would aid us all in data-driven decision making.
To answer your specific question, the fact that 58% of your finding
aid visitors came from Google means that 58% of users interacting with
your finding aids probably weren't looking specifically for archival
materials and probably have never heard of a finding aid when they
landed on your page. How can you use that information to improve the
user experience? You might want to re-examine your finding aid Web
display from a usability perspective that takes "novice" finding aid
users into account, and ask yourself the following:
Does the opening screen of the finding aid provide links or
* what this document is?
* what institutions this page is related to?
* where materials are physically located?
* how a user would request to view the actual materials described?
* who the user would contact to find out more?
Finding aid web displays often assume that users approaching finding
aids did so purposefully, when in fact, statistics like yours show
that a great percentage of them are arriving from Google searches,
possibly with no prior knowledge of finding aid or archives. I'm
seeing an increasing trend in the addition of a "what is this
document?" statement to finding aids that seeks to answer many of the
questions I listed above. In a usability study I conducted last year,
I found that the addition of such a statement greatly aided "novice"
finding aid users. For an example statement, see any finding aid at
UNC-CH, for example:
On Tue, Jun 22, 2010 at 2:34 PM, Jennie Levine Knies <[log in to unmask]> wrote:
> Dear Nathan,
> We have implemented Google Analytics on our finding aids. The code itself is
> easy enough to add to a page. We added it to our Footer include file, so it
> shows up on every finding aid and auxiliary page within the
> http://www.lib.umd.edu/archivesum site.
> I check it periodically, however, have not had the time to really analyze
> the findings. All I do is occasionally look and see what people are
> searching on and what our most popular finding aids are in terms of number
> of hits. I have three years worth of data, but so far, have done nothing
> with it. I initially set it up because I saw a great presentation about it
> by Chris Prom at SAA a few years ago. I was all fired up and ready to start
> working on it, but by the time I had collected enough data, I lost the
> I would be curious to hear if people use specific aspects of Google
> Analytics, and what they think the information really tells us. I also have
> not set up any "Goals" because I never had time to sit down and figure out
> exactly how they all work (I sort of feel like I need to take a statistics
> class before embarking).
> I keep trying to find a graduate student who would be willing to take my
> data and make some sense out of it!
> p.s. A quick look tells me that in the last month, 58% of visitors to our
> finding aids website came via Google... but what does it MEAN?
> Jennie Levine Knies
> Manager, Digital Collections
> 2216 Hornbake Library
> University of Maryland
> College Park, MD 20742
> (301)314-2558 TEL (301)314-2709 FAX
> [log in to unmask] E-MAIL
> Nathan Tallman wrote, On 6/22/2010 2:01 PM:
>> Happy Tuesday Everyone,
>> Just wondering if anyone has tried using Google Analytics with their
>> finding aids. I was thinking it would be easy enough to insert the
>> necessary code in the stylesheet template and then you'd would be able to
>> get some good data on finding aid use, as well as user
>> browser/network/hardware capabilities.
>> When I started my current position, I implemented Analytics on several
>> pages of our website. Our public finding aids are all HTML at the moment
>> and I didn't want to hand edit each individual document, so chose not to
>> implement Analytics in finding aids for the time being. We've recently
>> started an EAD project thought and this idea came back to me.
>> If you have implemented Google Analytics for your finding aids, I'd love
>> to hear about it!
Metadata and Cataloging/
Digital Library Initiatives