Dissemination should be the other 50% of what authors do: being read and having impact will not happen by itself.
Authors can influence discovery and readership through owned media – i.e. their own communication activities.
Earned media – i.e. when influencers write about your work – is key to reaching larger and more diverse audiences.
There is plenty of data for tracking engagement and use of articles, but it is scattered across multiple tools and providers and can be misleading or even incorrect.
Listservs can have higher engagement than modern, ‘cool’, social networking tools.
INTRODUCTION
It takes much time and effort to write a paper – but how much time and effort do authors put in to finding readers? In this case study, I explain why I decided to devote an equivalent amount of time and effort into finding and then engaging with my audience. Drawing on available data for three papers I published in 2017, 2018, and 2019, I describe how I promoted them, what happened, and what I learned. You will learn about the Conversion Funnel and how tools like Kudos and Altmetric can help drive and track your audience through its four layers: awareness, interest, desire, and action (downloading and reading). You will learn the difference between owned and earned media and why finding influencers and riding waves can be so important. I also identify areas inside the funnel where an author is dependent on others, lacks control, or where data is missing, each of which makes influencing the click‐through rate more difficult. The case study ends with a set of 10 lessons learned.
WHY ACTION IS NEEDED
The urban legend that many academic papers go unread beyond their authors'‘collegiate bubbles’ (Meho, 2007) was seemingly validated in 2014 when the World Bank reported that a third of its own papers were never downloaded (Doemeland & Trevio, 2014). However, as with most urban legends, the data tells another story. The World Bank's authors drew on data from a defunct repository and so missed data from a new one which showed that all reports were downloaded (C. Rossel, personal communication, May 2014). Ironically, the fuss that greeted the World Bank paper certainly drove its readership beyond its authors' bubble: it has been downloaded more than 8,000 times and, as of 19 April 2019, has an Altmetric score that tops 200. However, an essential question remains: how can authors boost their audience beyond their immediate peer group? Whilst a paywall might be a commonly cited barrier to being read (e.g. O'Brien, 2016), others exist, such as arcane and foreign language, discoverability, and even the comparative difficulty in using journals compared with other media (Waller & Knight, 2012). Plainly, you can only download what you know exists, so discoverability must be a primary barrier, especially because paywalls are now relatively easy to skirt with tools like Unpaywall (https://unpaywall.org/) able to find free versions of many paywalled articles, and as a last resort, there is what I like to refer to as the ‘Scottish Service’ (Note: According to theatrical superstition, speaking the name of Shakespeare's play Macbeth invites disaster, so thespians refer to it as the Scottish Play and the lead character as the Scottish King. Rather than invite disaster on our house, perhaps it would be safer to refer to SciHub as the Scottish Service (https://en.wikipedia.org/wiki/The_Scottish_Play).) aka, SciHub. Writing a paper is a significant investment in time (e.g. Margaryan, 2011), and authors, their employers, and funders will want a return on this effort. Of authors in the USA, 70% say they want readers beyond their sub‐discipline, and just under half say they want to be read by policymakers (Blankstein & Wolff‐Eisenberg, 2019). So perhaps it is no surprise that some 300,000 scholars – around 10% of researchers in developed economies (Organisation for Economic Cooperation and Development[ OECD], 2019) – have created accounts on a tool, Kudos (www.growkudos.com), which aims to help researchers communicate more effectively about their work (C. Rapple, personal communication, May 2019). Since January 2018, 39% of those who have registered with Kudos have used its tools to promote their articles, encouraged perhaps because using Kudos to promote scholarly papers leads to more attention, and there is evidence that downloads grow at a rate that is 23% faster than when it is not used (Erdt, Aung, Aw, Rapple, & Theng, 2017). However, these are early days, a recent review of using social media to drive downloads and citations seems to show little effect on these metrics (Davis, 2019) – but perhaps there are other objectives than simply boosting readership. Take for example the University of Manchester. The university implements a protocol that uses Kudos, The Conversation (http://theconversation.com/global), and Altmetric (www.altmetric.com), in addition to other University of Manchester services, to boost access to its authors' articles and connect its researchers with policymakers and influencers (UoMLibResearch, 2019). I work at the OECD, an institution that helps governments develop policies to improve the lives of their citizens. The help primarily comes in the form of advice based on the research conducted by the OECD at the behest of its funders. Plainly, if the OECD's research findings and knowledge are left unread, the OECD would be failing in its mission. This is why Angel Gurría, the OECD's Secretary‐General, often reminds staff that ‘dissemination is the other 50% of what we do’. For him, simply doing the research and sharing the results with governments and funders is not enough; to fulfil its mission, the OECD needs to win its fair share (or more!) of an ever‐larger audience's attention. This is captured by the OECD's Publishing Policy in the form of a simple objective: to ‘maximize dissemination’. So, when I turned my hand to being an author, I thought I would follow our Secretary‐General's injunction and spend as much time promoting my articles as I had spent researching and writing them: to see if I could maximize dissemination. This is my story of the ‘other 50% of being an author’, my story on trying to find readers beyond my bubble.
HOW TO ENGAGE YOUR AUDIENCE
The other 50%: Preparation
Just as there are tools and techniques to make writing easier, so are there tools (e.g. Kudos, Altmetric, Plum Analytics (https://plumanalytics.com/), and various social media channels) and techniques (e.g. a conversion funnel) for boosting readership. Kudos was developed to help researchers promote their publications and track their efforts in doing so. It invites authors to create a shareable summary ‘publications page’, where the work's core messages are presented in a non‐technical way along with a link to the original work on the publisher's website. (Kudos calls this a ‘publications page’, but I will refer to it as the article summary page to distinguish it from the article landing page on the publisher's website that hosts the actual publication.) It offers tools that enable authors to create tagged links that can be embedded in ‘owned media’ messages and content (see Box 1 for definition). A dashboard gives daily reports on the number of times the tagged links send traffic to the Kudos‐hosted article summary page, so the success of each owned media effort can be assessed. The dashboard also displays the publication's Altmetric ‘doughnut’, which leads to its detailed Altmetric dashboard.
BOX 1. Owned, earned, and paid media
Owned media is when you post content on communication channels that are under your control. These could be websites, blogs, social media, or email. Earned media is when other people (often known as ‘influencers’) are talking about your work on their websites, social media, and other channels. It includes traditional or mainstream media and things like letters to the editor, book reviews, and citations, as well as word of mouth (e.g. mentions at conferences). Paid media is when you pay to have your work advertised in both traditional and online media. This would include promoted items on social media channels and in search results.
Altmetric is a tool that enables researchers to track the reach and influence of their publications in ‘earned media’ (see Box 1), specifically mainstream media, on social media, blogs and websites, in reference management tools, Wikipedia, and in policy documents. Those with access to its premium features can track citations in journals and dig back through a publication's impact history. It works in almost real time, so it gives an author the chance to join in online conversations that they might otherwise have missed, for example, Twitter threads and blog postings. Using both tools together meant that I could track the impact of my own promotion efforts and see where one of my papers was being talked about by my audience. The latter was important in helping me to engage with my readers, to discover and join in conversations with people who had actually read my papers.
Funnelling conversions to drive readership
The Purchase Funnel (https://en.wikipedia.org/wiki/Purchase_funnel) is a long‐established marketing principle. Originally coined in 1898 by E. St. Elmo Lewis, it comprises four steps to making a sale: awareness, interest, desire, and action. Being an analogue process, its lack of reliable, affordable, data points brought about that oft‐quoted marketer's quip about not knowing which half of one's advertising spend is wasted (Bullmore, 2013). The arrival of e‐commerce, with its extensive digital exhaust, meant that the number of people clicking from one step to the next could be tracked cheaply and easily, and the ‘conversion rate’ (the percentage of people clicking from one funnel step to the next) could be calculated. Access to these data has revolutionized marketing, and any good marketer will now know the return on every advertising dollar spent. However, as an online marketers' objective is not always a sale (they might be after your vote), the digital funnel is known as the Conversion Funnel (https://en.wikipedia.org/wiki/Conversion_funnel). For an author in scholcom, the objective is to be read (and, probably, cited), so this forms the lower, ‘action’, part of the funnel. To get there, a reader must first be made aware that an article exists through search, owned, and earned media and have their interest stimulated by a summary page or abstract such that they desire to seek the full text and act by downloading and reading the work (Fig. 1). Not everyone will pass right through the funnel; a percentage will be lost at every step. So, maximizing dissemination requires boosting the number being introduced at the top and, using traffic data, removing frictions to reduce the drop‐out percentages at each step through the funnel.
Conversion funnel – formally published journal article with summary page.
Source: Author's illustration.
Understanding and exploiting the Conversion Funnel is fundamental to any promotion strategy designed to boost readership:
Search is not just how findable your content is to general search engines; it is dependent on the way your publisher, and partners such as Kudos or ResearchGate, prepare and present your content on and to academic discovery services (where a majority of scholarly searches take place; Blankstein & Wolff‐Eisenberg, 2019). The best way authors can boost their chances of appearing in search results is to post a variety of outputs (e.g. video, slides, blog posts) with simple, engaging titles in addition to the formal work. Your article's title should be clear and to the point, and all relevant keywords should be woven into its abstract.
Owned media is in the hands of the author (and potentially their employer, funder, or other partner) and should be used in the long run to drive awareness among the author's own network. This is where the author has the most leeway to act in the pursuit of readers.
Earned media can be leveraged by asking colleagues to send messages via their social media accounts but comes into its own when ‘influencers’ choose to review, comment, react, mention, and cite a published work and/or its author. Authors can seek out influencers, especially those beyond their own bubble, for example by sending out a press release.
A web page summarizing the article (ideally presented in an accessible and non‐technical way) could be hosted on the author's personal, departmental, and/or institution website; on the author's page on collaboration networks like ResearchGate or LinkedIn; and on the summary page of tools such as Kudos. To maximize discoverability, an author would use all of these places. The objective is to pique the interest of the visitor, to create the desire to go to the publication's landing page on the publisher website.
The publication landing page will display the title and abstract of the work, but it could also show key illustrations and other elements of the work that encourage and stimulate the visitor to act – to hit the download button. Authors have little ability to act here because this page is usually under the control of the publisher or repository owner.
The final, ‘action’, step is to download and read the work, which could lead to further acts such as saving a publication's details on reference and citation management tools such as Mendeley and, looping back to ‘earned media’, citation and sharing among colleagues.
MY STORY
Round one: ‘We've failed’
My first article drew on data to show that the proportion of born‐open journal articles had stalled at around 20%, leading me to conclude that the dominant open access models, Gold and Green, had failed, and therefore, a new approach was needed (Green, 2017a). It was published in time for 2017s early autumn event season that comprised ALPSP Conference, COASP, STM Annual Conference, and Open Access Week. The timing was important because I wanted to use these events not only to promote the article but to engage with its intended audience: publishers and librarians. On publication, I posted announcements on my Facebook page (where, at the time, I had about 150 friends), Twitter (~600 followers), and LinkedIn (~400 connections). The result, 72 click‐throughs. I also posted an announcement to two Listservs, generating 1,422 click‐throughs. Over the next 15 days, a period that included both the ALPSP and the COASP Conferences, I made 12 more ‘owned media’ promotional efforts: 10 using Twitter and 1 each on Facebook and LinkedIn. Most of the earned media was on Twitter, where more than 500 different people tweeted about the paper (see Fig. 2). This was an impressive volume, and many were researchers who exist outside my bubble (publishers, their suppliers, and librarians).
Number of tweets per day 7–22nd September 2017. Weekends shaded in grey.
Source: Altmetric.
The launch day (L+0) spike tailed off on L+1, a Friday, but picked up over the weekend and was sustained on L+5 and L+6 but then tailed off as soon as the ALPSP Conference started. Was my audience otherwise engaged and too busy to tweet? Or had the discussion exhausted itself already? All I can report is that a majority of ALPSP attendees that I spoke with had not heard of the paper, illustrating how hard it is to gain the attention of one's target audience even with the help of social media. To my surprise, it all kicked off again on L+9 (Saturday 16th) with more than 100 tweets because an influencer, ‘mathgenius’, posted the article title and a link on Hacker News (HN, https://news.ycombinator.com/item?id=15265507), which in turn was featured on HN's front page, triggering an automatic tweet to HN's 906 followers. This was re‐tweeted 30 times, including by various other HN bots, one of which had >20,000 followers. Midway through L+9, a tweet first posted on L+2 by Jon Tennent got a second wind and, together with the HN audience, drove the ‘conversation’ through the weekend. A fair proportion of the tweets contained comments or snippets from the article showing that the paper was being read, and it was not just a bunch of bots chatting to each other. The spike on L+13, midway through the COASP meeting, was the result of my re‐tweeting an image that I found circulating that day on Twitter (Fig. 3). I linked the image to my article and, in an attempt to reach the COASP audience (I did not attend the meeting), added the meeting's hashtag. This tweet had 17,892 impressions, 42 re‐tweets, and 84 likes, and the trackable link to the Kudos publication page was clicked 166 times.
Image I found on Twitter and re‐tweeted with a link to my article and hashtagged to COASP Conference.
In addition to all the action on Twitter, Altmetric logged one blog post (Retraction Watch's Weekend Reads), seven mentions on Facebook, 19 Google+ posts, and three Reddit posts – none of which, apart from one on Facebook, were initiated by me. Incidentally, I am only able to piece together this story of what happened thanks to the earned media history captured and stored by Altmetric. Over the next 8 months, I continued to promote the article, mainly using Twitter, each time using a trackable link from Kudos – each effort is shown with an ‘A’ in Fig. 4. After the launch month's high click‐through rate (CTR) (Table 1), the CTR fluctuated, with the next highest being in January, 5 months after publication. I also uploaded the Kudos‐created summary page in PDF form onto ResearchGate, where it has been viewed 753 times.
Altmetric score since publication of ‘We've Failed’ article (Green 2017a).
Source: Kudos and Altmetric.
Table 1. Efforts and click‐throughs
Month
Efforts
Click‐throughs
Click‐throughs per effort
September 2017
22
2,278
104
October 2017
9
222
25
November 2017
2
64
32
December 2017
2
4
2
January 2018
1
275
275
February 2018
4
220
55
March 2018
1
7
7
April 2018
3
33
11
May 2018
1
0
0
Source: Kudos.
It is all very well being able to see who has been tweeting about my article and to get anecdotal feedback at conferences and from the occasional personal email, but what I really wanted was to know how often my article was being downloaded and by whom (or at least know at which institutions my readers work or study). Knowing where and by whom my article was being read would give me insight into where I might be having an impact and, crucially, where I was not being heard. As I had learned at the OECD, knowing this would help me learn more about my actual readership and help me target future promotion efforts to greater effect. At OECD, we share download data with authors, and it usually confirms prejudices and produces surprises in equal measure. For example, I will not be breaking any confidences by revealing that European Union and United Nations institutions have a healthy appetite for OECD publications and datasets. But who would have thought that one country's army officer training school cannot get enough of OECD's works on education policy and the comparative performances of 15‐year‐olds at school? This latter data point prompted our education department to find out why resulting in an unknown unknown becoming a new connection. So, download data are invaluable yet, as I was to discover, hard to get. Even though Kudos is set up to integrate download data, few publishers are able to export per‐article, per‐day usage data, and unfortunately Wiley, Learned Publishing's publisher, was not one of them. I had to request the data from the editor who obtained it from the publisher to discover that, by the end of September 2017, the article had been downloaded an astonishing 69,148 times. (This counter‐compliant data point was double‐checked to ensure it had not been distorted by bots.) In October, it was downloaded 1,834 times, in November 967 times, and at an average of 315 times a month from then on to the end of 2018. All I could get was the totals; I was unable to get any data on which institutions or even which countries were reading my article, and I had to wait until the middle of the next month to get last month's data – hardly real time and no help when it came to planning future promotion efforts.
Round two: ‘We're still failing’
A year later, I began to wonder if there had been any progress to overcome the failure to deliver open access. A cursory glance showed that nothing had changed: the needle showing the proportion of born‐open articles had not moved, so I reached again for my keyboard. This time, thinking on why the needle was stuck led me to conclude that scholarly publishing was unaffordable whether done on an open access or subscription basis. I suggested that lessons from digital transformation be drawn upon to reduce costs and proposed a two‐step process whereby scholars would first publish a preprint, and then, providing the preprint gained attention, the author would be invited to submit a paper for formal publication. In order to be faithful to this proposition, I posted the paper as a preprint on the Zenodo platform on 6th September 2018, once again aiming for the autumn event season (Green, 2018). In order to help readers funnel back to the original paper (and in addition to the usual citation link in the references), I added a tagged link in the preprint's abstract that would take readers to the Kudos‐hosted summary page of the 2017 paper. By the end of April 2019, this link had been clicked 429 times, which is 8% of all visitors to the preprint landing page. Unfortunately, Zenodo's DOIs could not be integrated with the Kudos platform, so I could not use it to promote the preprint. However, Zenodo did integrate with Altmetric, so I can report on the preprint's owned and earned Twitter coverage (Fig. 5).
Tweets per day for the preprint recorded by Altmetric.
Source: Altmetric (Note that the x‐axis scale is very different to Fig. 2.).
This time, I had to work harder to gain attention: 36 of the 136 tweets (26%) over the launch period were mine (compared to 14 of 565 – 2.5% – the year before). My persistence was rewarded: for example, my three tweets during the COASP meeting triggered 20 re‐tweets. However, at an average of 9 tweets per day, attention was markedly down compared with the 35 tweets per day for the paper published a year earlier: the influencer ‘mathgenius’ did not come to my aid this time. I did not keep a monthly record of the downloads (displayed in real time on the Zenodo platform), but at the end of April 2019, the preprint had had 5,426 views and 1,796 downloads, and recently, the count has been growing at about 300 and 150 per month, respectively. However, as before, the download data have no detail: my readers, their institutions, and their whereabouts remain unknown to me. However, one of my objectives during this launch period was to ask for comment and feedback on the preprint, so I could improve the final paper. Within a month, I received substantive input from a dozen individuals, including two who corrected errors: this I considered to be a success.
Round three: Is open access affordable?
When I was writing the preprint, I was in contact with the editor of Learned Publishing, Pippa Smart, where the first paper was published. As she was not put off by the reaction to the preprint, I submitted a revised version to the journal in October 2018. It went through the usual peer review and acceptance process and was published on 25th January 2019 as part of a special issue ‘Bring the Facts, Bust the Myths’ (Green, 2019a). As with the preprint, I had to work hard to win attention on Twitter, creating 29 out of the 146 tweets that mentioned the paper (Fig. 6), but with the launch period falling between two of the winter conferences (APE 2019 was in mid‐January and R2R was in late‐February), I was unable to generate much momentum after L+9 (2nd February).
Number of tweets per day for 25th January to 9th February 2019. Weekends shaded in grey.
Source: Altmetric.
Between January and May 2019, I promoted the paper on two Listservs, generating 497 click‐throughs, LinkedIn (48) and ResearchGate (14); tweeted 35 times (846); wrote two blog posts (88); and commented on two other blog posts (48). Downloads of my article for January to April totalled 4,015 (Fig. 7, Article A). It is interesting to note that the ‘half‐life’ of my paper seems a little longer than the other two most‐popular papers, but Article D is unusual in building audience month by month.
Downloads (January to April 2019) per article for the first issue of Learned Publishing in 2019. My article is A. Note: The entire issue is free to download by anyone throughout 2019.
Source: Wiley/Learned Publishing.
Riding waves
One of the techniques I used to promote my articles is called ‘Riding the wave’. Essentially, one keeps an eye open for events, industry discussions, public statements, and social media conversations with which one can engage and draw attention to a paper. For example, in early 2017, Elsevier published a suggestion about how to work toward open access (Hersh, 2017), which triggered a fair degree of comment on Listservs and the Twittersphere. I posted a reply in the form of a blog post on Medium in which I included a tagged link to my paper (Green, 2017b). I then drew attention to the blog post using Twitter and LinkedIn, attracting 1,400 reads from which there were 310 click‐throughs to the Kudos‐hosted summary page – a click‐through response rate of 22%. Another example was the invitation for formal responses to Plan S. I posted my response as a blog post (Green, 2019b) and included tagged links to both papers' Kudos‐hosted summary pages. I drew attention to the post through Twitter and LinkedIn, and this effort resulted in 54 readers clicking through to the first paper's summary page and 72 to the latter. Most of my wave riding has been on Twitter where I use one of two techniques: attract the attention of conference delegates by using conference hashtags or join conversations by replying to suitable tweets, in both cases using tagged links so I can track the result. Five wave‐riding efforts that involved more than just ad hoc use of Twitter are summarized in Table 2. Each effort contained messages from the paper, so even if readers did not click through, a message was transmitted. It is interesting to note that it is still possible to generate a worthwhile click‐through and response rate many months post‐publication.
Table 2. Summary of efforts (excluding individual tweets)
Timing
Context
Effort
Channel
Result
CT
RR
Paper 1
L+20
Elsevier proposition ‘working toward OA’
Reply to Elsevier
Medium
1,400 reads
310
22%
L+47
Invitation
Pushmi‐Pullyu
LSE Impact Blog
‘Most‐read listing’
67
n/a
L+143
J of Infomatics Board ‘mutinies’
Are mutinies effective?
Medium
610 reads
13
2%
L+153
Plan S Response deadline
My response to Plan S
Medium
611 reads
54
9%
Preprint
L+46
Invitation
Fail Fast
LSE Impact Blog
‘Most‐read listing’
?
n/a
Paper 2
L+2
J of Infomatics Board ‘mutinies’
Are mutinies effective?
Medium
610 reads
55
9%
L+15
Plan S Response deadline
My response to Plan S
Medium
611 reads
72
12%
L+96
BBC Radio 4 Programme on OA
Replies to 6 Tweets
Twitter
334 impressions
28
8%
Source: Kudos, Medium, and LSE Impact Blog.
Timing is days post‐launch. CT, click‐throughs to Kudos publication page; RR, response rate (CT/result).
DISCUSSION
Data everywhere but not a drop to drink
We know that our digital environment generates a firehose of data. Yet, for authors in scholarly communications, data are hard to come by. Unlike e‐commerce, where marketers create effective funnels with vertically integrated digital platforms, a scholarly author has to try and construct a Conversion Funnel from poorly‐ or unconnected platforms and tools, many of which will not or cannot share their data (see Fig. 8).
Conversion funnel showing data sources and availability.
Source: Author's illustration.
For my two formally published articles, I was able to access data from my owned social media accounts and, thanks to Altmetric, some earned media channels (e.g. Tweets written by other people). Kudos could give me data about click‐through rates on my tagged messages, traffic volumes to the summary page they host, and click‐throughs to the publisher page. For example, for the first paper, as I write this, Kudos has logged 3,694 clicks from the 64 tagged promotion efforts I have made via owned media channels, 6,299 views of the summary page hosted by Kudos, and 878 clicks on the button that leads from that page to the article's landing page on the publisher website. That latter step from summary page to article landing page is a 14% click‐through rate – or to put it another way, only 14% of summary page viewers were sufficiently interested to have the desire to click through to the article. However, this is where the data chain breaks: I have no way of knowing how many of those who arrived on the article landing page were actioned to download the paper. All I know is that more than 70,000 downloads have been recorded, but I am none the wiser about the share that came from search and my own efforts or from earned media, nor do I know anything about them, not even where they are located. That the number of visitors to the Kudos‐hosted summary page (6,299) exceeds the number of clicks on tagged links (3,694) shows that the summary page is getting traffic from search and earned media – but I do not know how much from either nor have access to any logfile data that could help me understand more. When it comes to citations, I get conflicting data. As I write, Kudos tells me the first paper has nine ‘CrossRef citations’ yet confusingly invites me to view them on Google Scholar, where I find a list of 10 citing works above which is the metadata for my article and the message ‘cited by 14’. Meanwhile, Altmetric shows eight citations (sourcing the data from its sister company, Dimensions). The article homepage on the publisher site shows seven citations. Confused? You will be. De‐duping these records to arrive at a clean, comprehensive list of where my paper has been cited would not be easy – none of the sites offers a data feed or downloadable file. Nor do any of these tools offer alerts when new citations are found: for this, I have to rely on services like ResearchGate (which, incidentally, reports 14 citations). A simple data feed from Kudos and Altmetric would have made it easier to create the charts in this paper – I had to type the data into a spreadsheet. Altmetric's premium customers can download the data for their publications, but you have to learn where the link is – something I only discovered when doing a final edit for this article! The data from Wiley arrived as a table in a word‐processing document and I had to spend time copy–pasting into a spreadsheet before I could chart it.
LICENCES AND REUSE: A CAUTIONARY TALE
As a favour, OECD once published a book for a resource‐strapped fellow IGO. They insisted the work be published using a CC‐BY licence. Six months post‐publication, the authors and IGO asked OECD to issue a commercial distributor with a take‐down notice not because the distributor was offering a version for sale but because it was a crudely produced e‐book that, in their opinion, could damage their reputation. The distributor had found the e‐book online and had probably used some sort of automated process to strip the (copyrighted) artist images from the cover and inside pages and re‐cast the work in a new format: the result was anything but professional (a dog's breakfast came to mind). To the frustration of the authors and IGO, I had to explain that there was nothing to be done; the distributor had not contravened any of the rules of a CC‐BY licence. I tell this story because CC licences cut two ways when it comes to boosting dissemination and impact. Yes, others may well expose your work to audiences beyond your reach, but there are two issues to consider. First, there is the issue of reputation risk described above. This can be mitigated by adding ND (non‐derivative) to a CC licence, requiring disseminators to stick with your version of the work. Second, and this is harder to overcome, unless you work closely with your disseminators, you will have no idea who is re‐posting your work, if your work has reached a larger audience, or – indeed – if you are losing traffic and citations to alternative versions. In a world where funders are demanding impact reports from their fundees, getting access to all the download and citation data and knowing where your work has made a mark is going to be more and more important. At OECD, we encourage disseminators to use our shareable and embeddable editions because they are trackable: we can see when they have been embedded in websites and blogs and can monitor how often they are viewed there, and we can offer users a route to the fully downloadable and actionable editions on our website. Working with partners to reach a broader audience is important, but keep an eye on your reputation and get the usage data.
LESSONS LEARNED
Be strategic. Find and use a toolkit that will create a Conversion Funnel to build awareness and draw users through the interest, desire, and action steps. If possible, aim to publish just ahead of a series of events at which the paper can be promoted. (Note: I know that this will be a major challenge for most journals because they have such long and unpredictable production times and lack tools to plan releases. This is a major issue in journal publishing and one that publishers should be working to fix!) Choose your redistributors with care.
Be data‐driven. Log, measure, and track your audience's progress through the Conversion Funnel. Measure your owned media promotion efforts, so you can find out what works and what does not.
Be reactive. Use tools that report results of owned and earned media in real‐ or near‐real time. This will enable you to shape future promotion efforts around what is working and to engage with online conversations when they are happening. This is particularly important on Twitter and other social media sites where discussions and threads have short half‐lives.
Listservs rock. They might predate the internet, but postings to Listservs had a higher response rate than any other channel.
Reaching your target audience is hard: be active, be persistent. Even if your target audience is well‐defined and easy to target, winning their attention is hard because everyone is inundated with new information every day. So, do not be afraid to keep on going on. To avoid boredom and stimulate reaction, vary your message and tone. Use illustrations. Be opportunistic: if you suddenly discover there is an event going on, use the conference hashtag to follow it and jump in if you get the chance; if there is a new industry debate catching your target audience's attention, write a blog post (complete with tagged links to the article) and draw attention to that. Be active: do not be like one of the 80% on ResearchGate who just lurk (Khvatova & Dushina, 2019).
Find influencers. As I found with the first paper, someone influential can take your message to a wholly new audience that is way beyond your own bubble. You might only have a couple of hundred followers on Twitter, but you might know someone who has a thousand or more. If you cannot approach them directly, wait until they post something relevant and reply intelligently. If they have a blog, watch what they write and then comment, with tagged links, when you can. If your work might interest a broader public, do not hesitate to contact journalists; earned mainstream media can reach way beyond your own bubble and reach important audiences like policymakers and concerned citizens.
Be creative. Do not post ‘read my article’ messages. Post snippets that inform, pique curiosity, or contribute to debate. If possible, use illustrations that inform or entertain. Have a clear call to action, such as inviting comment and feedback or that leads to the next step in the Conversion Funnel.
Keep going. Unless your work is really out of date, keep promoting it because there are always new audiences or new contexts that make your work relevant, even months post‐publication.
Hassle your publisher for download data. Until publishers make download data publicly accessible in real time, regularly ask for it with as much detail as possible (where, when, who, etc.).
It is less work than it seems. During the 2‐week launch period, I found I was scanning Twitter and other social media channels perhaps four or five times a day (for a total of perhaps 30 min a day) and spending perhaps another 30 min creating new tweets and replying/engaging with conversations on earned media. Afterwards, I dialled back the effort to my normal scanning level with the occasional burst of effort to write a blog post when needed. I am sure it never amounted to the other 50% of my day – I'm sure I spent longer researching and writing the original papers – but the results in terms of readership and impact are, I am sure, better than if I had simply published and passively left it to search engines to find my audience.
ACKNOWLEDGEMENTS
Each of the three articles cited in this case study are free to download. I must thank ALPSP and Wiley, respectively owner and publisher of Learned Publishing, and Zenodo, funded by CERN, for publishing my articles on a free‐to‐read and download basis. I also thank Kudos' Charlie Rapple for prompting me to write this paper.
Biography
T. Green
REFERENCES
Blankstein, M., & Wolff‐Eisenberg, C. (2019, 12 April). Ithaka S+R US Faculty Survey 2018 [Web log post]. Retrieved from https://doi.org/10.18665/sr.311199
Doemeland, D.& Trevio, J. (2014, May). Which world bank reports are widely read? (World Bank Policy Research Working Paper No. 6851). World Bank OKR [Web log post]. Retrieved from http://hdl.handle.net/10986/18346
Erdt, M., Aung, H., Aw, A., Rapple, C., & Theng, Y.‐L. (2017). Analysing researchers' outreach efforts and the association with publication metrics: A case study of kudos. PLoS One, 12( 8), e0183217. https://doi.org/10.1371/journal.pone.0183217
Green, T. (2017a). We've failed: Pirate black open access is trumping green and gold and we must change our approach. Learned Publishing, 30, 325– 329. https://doi.org/10.1002/leap.1116
Green, T. (2018). We're still failing to deliver open access and solve the serials crisis: To succeed we need a digital transformation of scholarly communication using internet‐era principles. Zenodo. http://doi.org/10.5281/zenodo.1410000
Green, T. (2019a). Is open access affordable? Why current models do not work and why we need internet‐era transformation of scholarly communications. Learned Publishing, 32, 13– 25. https://doi.org/10.1002/leap.1219
Khvatova, T., & Dushina, S. (2019). Scientific online communication: The strategic landscape of ResearchGate users. Retrieved from https://doi.org/10.13140/RG.2.2.18535.09122
Organisation for Economic Cooperation and Development. (2019, 15 May). Main Science and Technology Indicators. OECD Science, Technology and R&D Statistics (database). Retrieved from https://doi.org/10.1787/data-00182-en
UoMLibResearch – University of Manchester Library Research Service. (2019). “Here at @UoMLibrary we're serious about supporting #OpenAccess to research …”. (Thread 1/26). Twitter [Web log post]. Retrieved from https://twitter.com/UoMLibResearch/status/1125770884214607872
Waller, R., & Knight, P. (2012). Overcoming barriers to the use of journal articles within the geosciences. Planet, 25( 1), 27– 32. https://doi.org/10.11120/plan.2012.00250027
