The library and information community have often been at the forefront of adopting new web technologies, but generally less thought is given to measuring how these technologies are being used.
An annual report may mention the number of followers the library's Twitter account has accumulated, or the number of article downloads from its institutional repository, but such a light-touch approach to web metrics neither recognizes its full potential nor acknowledges its limitations.
'Web metrics' is a broad term used to refer to the quantitative measurement of the creation and use of web content. It incorporates both 'webometrics', the quantitative study of web content and use for research purposes, and 'web analytics', the quantitative study of web content and use for the evaluation and improvement of a service.
The potential of webometrics is seen in studies that have found Google search data to be useful in estimating things such as flu activity or that Twitter mood can predict stock market activity, but it's for web analytic purposes that most librarians will come into contact with web metrics:
- Does the web site receive as many page views as expected?
- Is Twitter being used effectively to engage with users?
- Are digitized collections being used as expected?
The librarian trying to answer such simple questions can quickly find themselves overwhelmed by the variety of tools and metrics available. As well as analytics software (e.g. Google Analytics) offering to detail a web site's users in a multitude of different ways, there are services that provide metrics for comparisons between web sites (e.g. Alexa, Quancast, Compete). Social network sites often prominently display certain metrics associated with an account, with additional tools such as NodeXL and Webometric Analyst enabling the downloading of data for deeper network or sentiment analysis. Whilst general search tools such as Google and Google Trends may be used to gain insights into how an organization or its content is being linked to or mentioned.
With so many metrics available, here are five tips for the librarian to keep in mind when selecting and using web metrics.
1. Don't try to track everything
With so many metrics available it's important that the librarian only focuses on those things that matter. Just because Google Analytics lets you see how many people in Addis Ababa are viewing a web site using Internet Explorer and Windows XP, it doesn't mean it should necessarily be given undue attention when analyzing the impact of a local history collection in rural England. The librarian should have a clear idea of why they are interested in a particular metric, and how it aligns with the wider aims and goals of the organization.
2. Don't just reach for the closest metric
When Twitter prominently displays the number of followers an account has, it is not necessarily because it is in the Twitter users' interests to have as many followers as possible, but because it is in Twitter's interest to encourage users to try to build as many followers as possible. Of greater relevance to the librarian may be how many messages they are receiving from their library's users, or where they fit in a network of peers. Such metrics, however, are likely to require greater work.
3. Every tool has limitations
Although it might be possible for a librarian to count the comments on a blog for themselves, in most cases they are reliant on the information gathered by third-party tools and services, and it is important to recognize that they will all have limitations: Google doesn't index the whole of the web, and Twitter doesn't say how many times an update has been read. Some tools and services will give limited information, and some will give inaccurate information. It doesn't mean they can't be used, but unless you understand the limitations of the tool you can't determine what it can tell you and what it can't.
4. Accept the fact that tools will change
The web is a fast changing environment, and the information or functionality offered by a site or service one day will not necessarily be offered the next. At such times it may be necessary to replace one metric with another. Within the webometrics community this was most keenly felt when the major search engines removed most of their link functionality, which webometricians had been using to understand how and why certain sites were linked to. No metric or tool is irreplaceable, and alternative methods and tools were adopted.
5. No metric ever tells the whole story
One library's Facebook page having 1,000 likes and another library's page having 100 likes doesn't necessarily mean one library is engaging more successfully than the other. It could be that the more 'successful' library has had a page longer, has a larger number of real world users, or merely posts updates more often. Nevertheless, it would seem to indicate a need for further investigation. It is only through the accumulation of multiple metrics that we begin to really understand the impact of a web site's increasingly diverse content, and unfortunately the answers are rarely as simple as line managers would like them to be.
What have your experiences been in using web metrics? What tools would you recommend?
Let us know in the comments below.
Related knowledge and skills