Magazine article Information Today

Discovery Channels

Magazine article Information Today

Discovery Channels

Article excerpt

Getting as much information as you can is usually a good plan no matter what you are aiming to do. Sometimes, it's about experimenting with how end users prefer to receive or look for information. Other times, it's about uncovering ways to offer a deeper layer of information. Articles from the September and September/October issues, respectively, of EContent and Online Searcher examine how discovery tools help accomplish both types of tasks.

It's Variable

In "The ABCs of A/B Testing," Erik J. Martin offers this explanation: "A/B testing is the term used for randomly experimenting with a control variable (A) and an experimental variable (B) for the purpose of statistically testing a hypothesis." In the realm of online site design, A/B testing is used to compare two (or more), usually similar, versions of webpages or mobile pages to determine which version, A or B, users prefer. Doing so, in theory, should lead to a better conversion rate, which may include clickthroughs, hits, leads, and sales. A more sophisticated form of A/B testing, multivariate testing, lets end users try out separate elements within a piece of content or design to determine how each element impacts their interaction with the site.

According to one digital analytics expert, epublishers and content providers need to test every component of their site-including content length, search algorithms, and content positioning-to meet "an objective that is easy to understand yet difficult to execute: How do I make it simple for visitors to find and consume the content that interests them the most?" Another expert notes that just tracking visitor statistics will not relay how to improve a site's conversion rates.

Martin suggests that the novice should begin implementing A/B testing on a small scale, gradually working up to more complex testing. He breaks down the basic formula into five steps. First, pick the metric you want to test and improve, then form a hypothesis, such as "Bigger but shorter headlines will increase a visitor's stay on a page." Next, pick two designs to test (the original page versus the new page). Make sure your test results are statistically significant (which requires a 95% or higher confidence level). Use a tool or service that specializes in A/B testing. Finally, test for at least 7 days to obtain a large sample size of time and visitors. One expert recommends starting with your site's high-impact sections, such as a homepage or landing page.

While conducting A/B testing is neither easy nor foolproof, and users may initially resist any changes to the site, the procedure should increase the time a visitor spends on a webpage, ultimately leading to improved customer satisfaction and conversion rates.

Linked to a Feeling

Since I'm not a librarian, most of the terms used in David Stern's feature, "Making Search More Meaningful: Action Values, Linked Data, and Semantic Relationships"-such as "RDA" (resource description and access), "triplet structure" (the defining of an object along with its potential properties and values), and linked data ("a set of best practices for publishing and connecting structured data on the Web")- were new to me and a little hard to digest at first. However, the further I got into the article, the more interesting it became. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.