Yesterday’s news that Facebook has launched its Instant Article initiative, whereby it hosts selected news and magazine content rather than linking to another website, is a good reminder about the changing power dynamics that shifts in discovery practices are introducing. The New York Times’ coverage of this launch focuses extensively on its implications for The Times itself, which is one of less than a dozen publishers that is experimenting with what is for the news industry a new content distribution paradigm.
Facebook’s Instant Article initiative appeals to news publishers, a group looking to find additional avenues to monetize their online content as print continues to erode. One statistic that especially caught my eye: “For The Times, Facebook represents from 14 to 16 percent of its web traffic — a figure that has doubled in recent months.” Trends like these raise several issues for any publisher, the most important of which is that whatever traffic a discovery service gives a publisher today, it can take away tomorrow. Facebook has regularly updated its newsfeed algorithm, which is possibly the single most powerful recommender system in use today, to highlight news over memes one month, personal interest over the public good another.
For this reason, the news brands that are internationally powerful enough, such as The New York Times, The Wall Street Journal, and the Financial Times, have sought to strengthen their direct relationships with readers rather than rely simply on traffic sent to them by discovery services like Facebook, Twitter, and Google News. Often these have taken the form of online subscriptions which are required for access after a reader exceeds a metered threshold of article consumption over a given time period. Certainly, online subscriptions are a preferred source of revenue as compared with advertising, not least because the former can erode only so fast while the latter can disappear overnight when an important enough third-party updates its algorithm. But while the revenue from subscriptions has helped news publishers’ bottom lines, there is another dimension to their interests. Ultimately, they are looking for mechanisms to strengthen their relationships with readers – and potential readers – as a mechanism over time to discourage or at least limit “grab and go” behavior.
News publishers share with scholarly publishers the challenge of “grab and go” behavior. From the publisher’s perspective, an individual who lands on an article (or ebook or other material type), perhaps downloads it, and then disappears back to the discovery service, represents a lost opportunity, as compared with an individual who visits The Times, or ScienceDirect, or Oxford Journals Online, due to mindful interest in the materials there. Savvy content platforms are analyzing the behavioral differences among researchers who arrive via an internet search engine, through Google Scholar, through Scopus or Web of Science, through an A&I service, through a content alert, or through a bookmark, a LibGuide, or other direct mechanism. They are finding in some cases material differences in users’ behaviors that provide them with strong incentives to try to convert “grab and go” behavior into a stronger relationship. Strategies for doing so are only beginning to be designed and implemented.
Why then would news publishers even consider joining Facebook’s Instant Article initiative? For one thing, there is money to be made given Facebook’s willingness to drive traffic and advertising revenues, while there are promises of analytics no less sophisticated than those the publisher has on its own platform. But while these explanations may offset some of the disadvantages and perhaps provide marginal revenues, they impede the core objective of strengthening reader relationships.
Instead, publishers are battling against the fact that in many cases, readers have a much stronger relationship with their discovery services than they do with content providers. This is no less true for scholarly publishers than it is with news brands. One has shared far more personal information with Facebook, Twitter, or Mendeley, or through a user account with Google or any of a variety of other services, than with The Times, or ScienceDirect, or Oxford Journals Online. While scholarly authors may in some cases have strong relationships with their preferred publishers or journals, on the reader side workflows and affinities increasingly favor discovery over content. For Facebook, the long time lag between a user clicking an article in their news feed and it loading for them – 8 seconds! – was an enormous stumbling block impeding its users’ satisfaction, and since it manages the discovery algorithm, it makes choices about what content to preference for its users. Scrolling through one’s Facebook feed, one is perhaps interested in the latest story that a friend has shared but far less frequently a loyal reader of The Times’ coverage of that particular issue.
Some may laugh that 8 seconds should make a difference one way or another, but I recently wrote (here and in greater detail here) about some of the much more serious stumbling blocks that impede researcher access to scholarly materials online. What is the equivalent of Facebook’s Instant Articles initiative? If the indexed discovery services drive increasing amounts of researcher traffic to scholarly content, it is easy to see advantages to putting increasing amounts of content “inside” the discovery services, or at least on their affiliated content platforms. It is possible to interpret aspects of EBSCO’s and ProQuest’s strategy over the last few years as consistent, at least, with this type of outcome. Other scenarios are possible if Google were to wish to act on the high share of discovery that it drives through its Search and Scholar products, or if Elsevier through Scopus and Mendeley do so, or if Twitter or a variety of innovative startups play that role. Even if discovery and access were to take place in a highly decentralized environment, the need to address the access stumbling blocks that researchers face are no less urgent.
But discovery, at least in the consumer sector, seems to possess a certain type of centripetal force, perhaps like gravity one that is a function of its mass. Some libraries and publishers may yet have an ability to influence which discovery services for scholarship will achieve this level of “critical mass.”
Discussion
20 Thoughts on "Grab and Go and the Gravitational Pull of Discovery"
If it’s all about readers’ relationship discovery, then publishers can’t hope to compete head-on because we’re in a different business. So, the question becomes this – how to ‘infect’ discovery services with our content so we win more eyeballs and having won those eyeballs, how do we persuade enough eyeball owners to contribute to our bottom line (either via their institution’s subscription or by getting out their credit card). We’re trying to increase the number of ‘infections’ by allowing anyone to share and embed our content – you could say we’re using our readers as the transmission agent (our ‘mosquitoes’?) to infect third party websites with our full text content. We had an example this week when we released a book on alcohol usage and it ended up being embedded in a number of news websites (as well as shared via Facebook), boosting discoverability significantly. We reached a large number of eyeballs in the process (on the day of launch, we recorded more than 50,000 views on the embeddable edition). The key question is this – how many of these eyeballs’ owners clicked through to our premium services (which allow downloading, local printing, copy-pasting, access to underlying data etc) and thus contributed to the bottom line? This week it was nearly 8%. As they say, it you can’t beat ’em, join ’em!
Toby Green
OECD Publishing
“Instead, publishers are battling against the fact that in many cases, readers have a much stronger relationship with their discovery services than they do with content providers.”
Interesting stuff, but is this a new problem or just an old problem in new guise. If you replace ‘discovery services’ with ‘bookshop’ (or even ‘physical library’) then can publishers say they didn’t have this problem in the analogue world? Publishers’ brands have always been of low value to readers, perhaps because, as bizarre as it may sound, we’ve never actually serviced them.
We’ve serviced authors wishing to share their thoughts, ideas, results with the wider world, and people (bookshops, libraries) who then buy the resultant products to sell on to the ultimate users (ie readers). But readers? Not so much perhaps as there’s rarely been a direct, eyeball-to-eyeball relationship.
The failure in this analysis is in not forging positive mutually expansive relationships with discovery services in the early days, in the way publishers did with bookshops and libraries.
Your comment on the role of publisher brands in discovery is an important point Martin. It is true generally that publisher brands have not been of value to readers. It is worthwhile looking at the exceptions to this, however. Journal brands have long been important to readers (as well as authors) and some publishers have built their brands around these: Nature Publishing Group, the Lancet, Cell Press, JAMA Network, and so on. Society brands have been important to readers. And on the book side, university press brands matter.
A brand remains a strong signifier in the context of a discovery system. If presented with a news article from the New York Times, or a scholarly article from Science, in the context of my Facebook feed, I am more likely to view those articles than ones by publications I have never heard of (and I suspect there is a bit of a retreat to strong brands going on give the volume of click-bait out there at present).
A strong brand gives a publisher more options — and more leverage when working with discovery services.
Certainly it is no coincidence that Facebook has chosen some very strong brands, such as The Times, for the launch of this initiative.
Perhaps so, but haven’t there been challenges engaging these intermediaries in recent years? Has something more structural changed about the benefits of having a direct relationship with the reader?
Just out of curiosity, how recent is this use of the word “discovery”? In science discovery means finding something new, not something already written. Until I wandered into the scholarly publishing community a few years ago, I had never heard discovery used this way. Is it new or old jargon? Is a bookstore a discovery service or must it be digital?
That’s a good question. I’ve been using discovery in this sense for more than a decade and can’t remember a time in my career when it wasn’t used in this way. I’d be interested in others’ answers to this question about jargon.
Here is a 2012 article indicating that the term discovery services was emerging and needed to be explained.
http://www.infotoday.com/cilmag/apr12/Hoeppner-Web-Scale-Discovery-Services.shtml
The so called indexed or web-scale discovery services that many libraries now use as the front-end to their collections date back about that far. But the use of the term discovery to describe the researcher’s act of seeking information resources dates back further.
No doubt you are correct. Here is my problem. I have preached that the logic of search is one of the great scientific and technological challenges of our time. If I call it the logic of discovery people will think I mean scientific discovery, not bibliographic discovery. Hence my concern with the use of the term discovery to refer to literature what amounts to literature search.
Search is also a flawed term, though, in that the discovery I am discussing includes anticipatory and contextual methods and includes everything from basic browsing to advanced alerting.I am comfortable referring to information discovery (which is what I mean) as compared with scientific discovery (which is I think what you mean).
“Discovery learning” is a long-standing educational theory and I think contemporary use of the word “discovery” in the information/publishing professions is in some ways more akin to that perspective than the concepts in information seeking behavior and search that some might assume discovery is aligned with. Here’s an interesting article that reflects discovery learning rather than search per se:
Marcum, J. W. (2001). From Information Center to Discovery System: Next Step for Libraries?. Journal Of Academic Librarianship, 27(2), 97. http://www.sciencedirect.com/science/article/pii/S0099133300001816
It seems to me that given audiences know where to go but the mass is to small to be critical.
Interesting thoughts indeed. I wonder however why, if there is much to gain by hosting content on discovery platforms, have the providers of these platforms not aggregated CC-BY Open Access content? I know of a handful of smaller companies doing this aggragation, but the big players have not pushed in this direction. Maybe there are differences between the news and scientific content markets that could explain this. Mendeley and ResearchGate are special cases that leave uploading to the authors.
This is a really excellent question. I don’t know if Ebsco’s and Proquest’s content platforms include any open access content. The considerations associated with whether and how to do so would presumably go beyond just the discovery dynamics I discussed in this post.
I first heard this discovery term maybe a couple of years ago?
“Some may laugh that 8 seconds should make a difference one way or another”
I would suggest that an 8 second response time should be cause for tears, not laughter. Data from 2009 (aka the digital Middle Ages) indicated 40% abandonment at >3 seconds. And the world has moved on quite considerably since then.
Response times of <3 seconds are a commercial imperative for any player in the digital space.
Agree entirely.
Wonderful and insightful piece, as always, Roger!
One of my big takeaways from this is the increasing importance for publishers to diversify their discovery channel participation and investments. Our “promiscuous” metadata must “infect” (to call back to Toby’s great term) many, many indexing services and discovery tools — both in order to capture diverse types of users and safeguard against usage impacts of changes in those channels.
Do you see any role for differential distribution in any cases? For example, providing better indexing, or full text, or more real-time updates, for some discovery services than for others?
