![\"2017-05-25](\"https:\/\/blog.oclc.org\/next\/wp-content\/uploads\/2017\/05\/2017-05-25-Trouble-With-Data.jpg\")
<\/p>\n<\/p>\n
We\u2019re being inundated with data. That\u2019s what we\u2019re told, right? We hear all the time how many exabytes of new data are being created every day. There\u2019s just one problem: maybe none of it is the data we actually need.<\/p>\n
I recently had the opportunity, along with several of my OCLC colleagues, to attend the Electronic Resources and Libraries (ER&L<\/a>) Conference. I\u2019ve been going\u00a0to this great conference for the last two years, and each year it offers a really valuable look into how libraries manage e-resources. This year, several topics across multiple presentations led me to the conclusion that actionable data is actually pretty hard to find and even harder to wrangle successfully.<\/p>\n <\/p>\n An important element of any successful data analysis strategy is to first realize that you\u2019ll never have all of the data you need. At some point, you have to take what you\u2019ve got, come to some conclusions and move on. But we should also be looking for ways to start by collecting data that is more actionable from the get-go.<\/p>\n As I thought through the ER&L presentations and listened to librarians talk about their \u201cwish lists\u201d for data, I found that three themes seemed to resonate around this idea of \u201cgood data\u201d as opposed to \u201cmore data.\u201d<\/p>\n If you\u2019re not planning on making changes based on the data you\u2019re collecting and analyzing, just stop. I heard from folks who\u2019d been collecting all kinds of statistics that, at the end of the day, they weren\u2019t useful in terms of supporting actual strategy or tactics.<\/p>\nWhat changes are you willing to make based on how you analyze data? <\/a><\/span>Share on X<\/a><\/span>\n The exact opposite of that is the demand-driven acquisition (DDA) and evidence-based acquisition (EBA) models, which we heard quite a lot about at ER&L. In these programs, libraries partner with publishers to make e-resources available for purchase or access after crossing some usage threshold. It\u2019s a great example of using just-in-time data to make specific collection development decisions.<\/p>\n In the absence of readily available (or useful) data, it seems that more libraries are investing in user surveys and usability tests, especially involving students. Librarians from CUNY<\/a> spoke about<\/a> watching their student testers locate an electronic article on a specific topic. The staff watched where the students navigated online, what they typed into search boxes and how they sorted results.<\/p>\n Some libraries, like Montana State University<\/a>, are working to optimize<\/a> their e-resources for search engines and social media. Their Open SESMO<\/a> project incorporates linked data terms and social media-friendly images into e-resource records to help online information seekers find their resources. The \u201cdata\u201d in this case is based on observations of students in order for the library to \u201cbe in the right place at the right time.\u201d<\/p>\n Smart, simple glimpses into the \u201clife of the user\u201d can inform your strategy as much as pages of gate stats.<\/p>\n Most librarians I talked to are trying to manage whatever data they can collect in Excel, which has limitations. Library staff who want to engage in data-driven collection development are still sometimes forced to make guesses about what their users want or compare \u201capples and oranges\u201d when data comes in different formats from different sources.<\/p>\nYou will never \u201chave it all\u201d<\/h2>\n
\n
1. What are you willing to change?<\/h2>\n
2. Be an anthropologist, not a number cruncher<\/h2>\n
3. All together now<\/h2>\n