TrendMD: A Google AdWords for scholarly content
Scholarly content represents the leading edge of humanities knowledge. With over 5,000 new papers published across 27,000 online journals daily, many scholarly articles simply never reach their intended audience and consequently fail to generate the impact they deserve.
TrendMD was founded by physician-scientist Dr. Paul Kudlow, and leading open access publisher of JMIR Publications, Dr. Gunther Eysenbach, with the simple goal of connecting the right content to the right audience. The two joined forces with Alan Rutledge, a talented software engineer, to create a personalized recommendation engine that allows us to distribute scholarly content across online journals and blogs and push content directly to its intended readers.
When a publisher installs our free recommendations widget on their journals and blogs, we start promoting their content across our network and send them new and interested visitors. The widget keeps their readers reading by surfacing content that the publisher has published in the past. Publishers also earn revenue from clicks on sponsored, third-party content presented on their sites.
Here’s a concrete example from the reader perspective:
1) The British Medical Journal publishes an article about heart disease and would like to promote it. It sets a budget, and the article is disseminated through TrendMD’s network (over 400 journals and growing).
2) Meanwhile, you’re reading a different article on heart disease in JAMA. Since JAMA has the Trend MD widget installed, it tells you that you might also be interested in the first article in the British Medical Journal.
3) You click the link. Your click “costs” one dollar, which is deducted from the British Medical Journal’s tab. JAMA and Trend MD split the proceeds—and the British Medical Journal enjoys the increased exposure and site traffic.
Traditional recommender systems in academia generate recommended links via semantic enrichment (i.e. relatedness).
TrendMD utilizes a variety of factors to determine the best links for each reader, including relatedness, audience click behaviour (“people that read X, also clicked on Y”), article popularity, and personalization (i.e. what you have read in the past). Our algorithms are optimized to connect the right content to the right reader, at the right time.
As of February 2015, the average click-through rate across the TrendMD network (420 journals and scholarly blogs) is 6.4%. The use of audience click behaviour and personalization typically increases click-through rate by as much as 40%, according to our A/B tests. While TrendMD has yet to conduct formal reader surveys, our high click-through rates as well as positive feedback from our publishing partners indicate that readers find our personalized recommendations useful and engaging.
But we didn’t just create TrendMD for publishers and readers. Content providers of all sorts, including individual researchers/authors, and pharma can promote their research through TrendMD. As buyers, they select a budget and pay only for the actual click-throughs they receive. While Open Access was a first step towards authors and publishers enhancing the visibility of scholarly content, TrendMD is a technology that addresses an unmet need of authors and publishers to actively distribute their best work across journals and publishers.
We tested TrendMD’s effects on article visibility. In September 2014, we completed a parallel-group randomized controlled trial on 100 articles, all published on the same day, randomly selected from 75 medical journals indexed on PubMed. 50 articles were distributed and 50 articles remained undistributed. Articles that were distributed across the network received an 87% average increase in article views at 6 weeks versus control (191 views versus 102 views).
Moving forward, we are running a series of pilot tests to determine how distribution over the TrendMD network affects other metrics such as altmetrics and citations.
With a five-percent week-to-week growth rate since launch in May 2014, TrendMD is quickly revolutionizing the way scholarly content and knowledge is disseminated across the world. TrendMD is already working with many of worlds top scholarly publishers such as BMJ Group, Taylor & Francis, de Gruyter, Nature Publishing Group, and many others. Before a publisher can become part of the TrendMD network, TrendMD staff completes an audit of their content to ensure that it meets our standards of quality. For more information on our content guidelines, please see: http://www.trendmd.com/content-guidelines
In terms of our traction to
date, TrendMD has gone from one journal to being installed on greater than 420
journals and blogs owned by 16 premier scholarly publishers. We serve up to 12
million recommendations to approximately three million unique scholarly readers
per month, and growing exponentially. TrendMD also has a 100% retention rate of
the widget on journals owned by our early publishing clients. Our vision is to
create the infrastructure necessary for translation of research into innovation
in the years to come.