Information Architects always want to know how easy our site is to use and how often it is used by the audiences we target. For the most part, few people ask for our data on usability or user experience but we regularly receive this kind of request from other parts of the business, "I need to know how many ‘hits' the website has received this year and how that compares to our competitors?"
It's not that these people don't care about how easy the site is to use but, generally, reporting on how the site performed in usability tests or how it was received during a focus group is fairly squishy data. More powerful and comprehensible is reporting that visits to the site are up x% over last year: hard numbers are just more tangible and they measure up better compared to other metrics shared around the business.
Web analytics offer quantitative insight into user behaviour. They can be used to benchmark site performance and report to management. But, because web analytics data tells what people do on a web site, analytics data should be used to inform and direct more qualitative user research methods such as focus groups or usability tests that tell us why they do what they do. More importantly, you can use the natural interest people have in web metrics to introduce the more qualitative measures into the business overall.
In my work, I am always looking for ways to get clients more interested in what people are really doing on their website and to encourage them to ‘think like a user'. In a perfect world, I would have an ongoing engagement with site visitors - they would be involved in focus groups and usability tests so that every move we made could be vetted by the people for whom the site was built. Wishful thinking, I know.... With web analytics, you can see what people are doing on your site 24 hours-a-day, 7 days-a-week. Even without in-person contact, an organization can systematically stay on top of the needs and interests of its website visitors.
An information architect should care about web analytics because it allows you to broaden the scope of your research. For example, if you want to improve a particular section of your site, ideally you would call in your target audience for some interviews, build a prototype, and then test it with users. Even with a tremendous research budget, you have only looked at a portion of the site and reached a specific part of the visitor population. Web analytics is research on your entire site with your entire user population. With analytics, there are ways to identify problems or issues that you might not have known or cared about otherwise.
I look at Web analytics as one more tool in my research toolbox. While I might use focus groups to get opinions on issues to explore or usability tests to look at the performance of specific areas of the site, I use analytics to find areas that might be otherwise overlooked. The findings from my analytics reports influence the information architecture and the content strategy of a site just like any other piece of research and also can provide a way to pave the way for future research.
Many designers balk at using web analytics because they are intimidated by the numbers and the potential complexity of the analysis. I will share a little secret with you - you do not need to be a statistician to use or interpret web analytics. What you do need to be is a good interpreter - something that most IAs already are - because what analytics tell you is the story of your website over time.
Storylines include how people got to your site, what they searched for when they were there, what they looked at and what they did not look at, and how all of this has changed over time. As Information Architects we need to translate and interpret the story that the analytics tell us to communicate what is going on to the people who need to know. Not having a person available to tell the story of the data is one reason web analytics are under-utilised. The numbers themselves are not that interesting, but once you add the context, the story begins to unfold.
Translating the data into a story also draws the attention away from the inevitable inconsistencies of web statistics. In my case, I am not able to use cookies or page tags to track visitors, so my data is always going to be less accurate than it could be because of the way the log files are interpreted. Even with this limited data, I can still tell a story that looks at the big picture and the trends over time and that helps the site owners to make informed decisions on site architecture, content strategy and other priorities.
Just as important as knowing what analytics can tell you, is knowing what they cannot tell you. First, statistics are fairly useless at allowing you to compare one site to another. For the most part, there is no consistency among analytics programs and no way to guarantee that they are looking at the logs or the traffic the same way. Second, web analytics data tells what people do on a web site, but not why they do it, which is why it is important to use analytics data to inform and direct more qualitative user research methods.
Often, clients extrapolate a whole scenario as to why certain traffic patterns emerged. While this excitement is a terrific indication that the person is engaged with the site and the data, it is dangerous to leave statistics data open to subjective interpretation. You must be the voice of reason and remind the client that, in order to find out why people did what they did, you have to conduct additional research into the thoughts and motivations of site visitors.
What I have laid out here is a guide on how to starting using web statistics data to inform your information architecture and site design.
Step 1: Check out the site
Like with any project, the first thing you should do is check out the website you are analysing. Your purpose here is to absorb enough about the site that when you look at the analytics report you will have some context in which to operate. The questions you should ask yourself are probably the same ones you use already to get oriented:
Who is the site's audience?
What are the key pages on the site from a user's perspective?
What are the key pages from the business's perspective?
What is the mission of the site?
What are the goals of the site?
From a technical perspective, what is the structure of the URLs? How are the pages generated?
Step 2: Check the accuracy of the stats report
If the site is using a hosted analytics solution, most of these things are taken care of for you. However, in the case where you cannot use a hosted analytics solution, you have to make sure that your software is set up correctly. While you can do this yourself, you must have some technical understanding of server logs and the configuration of your site.
Variables, such as non-human visitors, should be excluded from the report to ensure that the numbers are as accurate as possible. If you have the luxury of setting up an analytics report or contacting the person who sets the report up, you can control these things. If not, you have to check out report yourself to make sure. The numbers you are looking at are anywhere from 10 to 70% wrong depending upon how "clean" your report is.
Things that should be excluded from reports:
Searchbots, spiders, and other non-human visitors - if you see that the top visitor to your site is "crawl-66-249-65-83.googlebot.com_Mozilla/5.0 (compatible;Googlebot/2.1; http://www.google.com/bot.html)" then you know that these are not being excluded.
NOTE: even if an admin has excluded bots and spiders once from your report does not mean that you are in the clear forever. Definitions need to be updated at regular intervals to keep your report clean.
Internal visitors - the IP addresses of internal web visitors should be excluded. For example, if you are looking at ENERGY STAR stats, then all visitors from ENERGY STAR staff and their contractors should be excluded from the report.
Step 3: Read through the reports quickly
Now that you are familiar with the site and know that the analytics are being reported correctly, read through the analytics report to get a high-level overview of what is being reported on and the overall trends. Based on what you know about the site, are the top visited pages and other metrics what you expected to see?
Step 4: Compare the trends over time
Next, look at a number of months' reports side by side to get a sense of the trends. Some traffic trends are cyclical, e.g. there is more traffic during the week than the weekend or the summer months are slower than non-summer months. Are the same pages appearing over and over again? Are there some pages that do really well and then fall off the radar?
Step 5: Make the Communications/PR/outreach team your new best friends
If you have not connected yet with the people who do communications or outreach for your site, an analytics report is a good reason to do so now. These people can tell you about promotions and campaigns that might impact traffic. They can also keep you in the loop about what the hot issues are in the news and the information ecosphere that might get people to your site organically.
Also, marketing and communications people will become your champions because your analytics report shows them and others how well they are doing.
Step 6: Look at the Key Metrics
Up until this point, I have talked all around the analytics report, but not about the actual data that is contained in the report. A statistics reports contains a million different pieces of data, but how do you know what the most important pieces are to look at? These are the metrics that I use most frequently in my reports:
Visits and page views Here, don't just pay attention to what appears on the list, but look at what doesn't appear. Are there things that from your knowledge of the site should be popular but are not? IA is never done in a vacuum - we always have to consider the user needs and the business needs. Take page views, for example. Sometimes more page views is a good thing because people are very interested in your site content, but perhaps people are viewing more pages because they cannot find what they are looking for.
Site navigation preferences Look to see if visitors going where you want them. I was able to justify a site redesign when I found that none of the most visited pages were available from the homepage or any other top level pages.
Entry and Exit pages See if visitors are leaving at expected places such as from an information-filled (rather than link-filled page), or at unexpected places such as in the middle of a multi-page process.
Keywords From a statistics report you can get the keywords that visitors used on the site search engine and on external search engines, when those searches resulted in visits to you site. It is important to look at the terms that show up on the statistics reports as well as the words that do not show up. Internal search engine terms helps you identify what people are not finding in your site through browsing. Both types of terms provide a window into your visitors' vocabulary and into the way they look at the content of your site.
Referrers The sites that refer visitors to you site are important to look at for a number of reasons. Referrers might identify potential partners for your business, locate previously unidentified audiences for your content, and, as use of user-contributed sites increases, find other sites that your site owners can participate in to reach a wider audience.
Conversion Any multi-page process within a website can be measured to determine the percentage of people that make it from beginning to end.
Step 7: Present the data clearly and provide next steps
The last, and possibly the most important step, is to present your analysis clearly and to make it actionable. As the analyst, the deeper you get into the data, the more you have to be careful that you are not confusing people with the minutiae which can do more harm than good. A summary at the beginning of the report that lists items of interest and any recommendations for improvements suits those who are not interested in reading the charts and graphs that make up the meat of the report.
I try to use my summaries to tell the story of the web site and the visitors that came to it. What initiatives or activities caused traffic to the site? Where did people come to the site from? What did people do once they arrived to the site? What did they search for? These are the same key metrics that we teased out of the report earlier, but when you put these metrics into a narrative, you will reach more people.
Once you have summarized the key findings, make some recommendations for what should happen next. Sample recommendations could be to watch a particular area to see if a new finding is the beginning of a trend or a fluke or to conduct in-person user research to explore a part of a site in more depth. Even if your recommendations are small, it is important to let people know what they can do to improve or change performance, or how to get more information. More than anything you can do, this will begin the process of making analytics an invaluable part of your user-centred design process.
Hallie Wilfert is a Senior Information Architect at SRA International, a technology and strategic consulting firm in Arlington, VA, where she works mainly with government clients. Previously, Hallie worked as the Director of New Media at the Institute of Medicine of the U.S. National Academies. As an information architect, she is committed to advocating for the user while satisfying the needs of the client.
Hallie received her B.A. in American Studies, studying at the University of California, Berkeley and the University of Maryland, College Park. Hallie is a member of the Information Architecture Institute and the Usability Professionals Association (UPA). She has presented on integrating web analytics into user centred design at the 2008 Information Architecture Summit and the 2008 International UPA Conference.
The FreePint Family is a family of resources to help information workers be more effective, raise the value of information in their organisations and contribute to success.
'FreePint... provides most of my professional development because it won't come through work and [other resources] just don't cut it.'
FUMSI Forum: Do you have a research question? Post it to the FUMSI Forum, where professionals share Q&A and useful tips on how to Find, Use, Manage and Share Information. It's free.
![[Reuse options]](http://license.icopyright.net/images/icopy-w.png){kind=small}
Information Architects always want to know how easy our site is to use and how often it is used by the audiences we target. For the most part, few people ask for our data on usability or user experience but we regularly receive this kind of request from other parts of the business, "I need to know how many ‘hits' the website has received this year and how that compares to our competitors?"
FreePint Shop:
VIP LiveWire:
