Image representing DeepDyve as depicted in Cru...
Image via CrunchBase

Yesterday, the online article intermediary DeepDyve announced a rental model for scientific papers, prompting one news outlet to call this new business, “Netflix for Researchers.”

DeepDyve is a search engine specializing in indexing and providing direct access to scientific articles.  But it does so in a novel way — by renting, not selling, access.

Their motto is, “Research. Rent. Read.” It forgoes the ownership model of access.  Like DVDs that need to be mailed back to Netflix, DeepDyve is based on a short-term rental period of journal article access.

Users can sign up for three types of membership:

  1. Basic. Pay as you go at $0.99 per article for a 24-hour rental
  2. Silver. A monthly subscription plan at $9.99 , which allows 20 articles at any one time with a 7-day rental, and
  3. Gold. A subscription plan of $19.99 allows unlimited rentals with no rental period

The list of publishers who have teamed up with DeepDyve include:  Oxford University Press, Sage, Taylor & Francis, Wiley-Blackwell, and PLoS, along with many of the society publishers hosted by HighWire and BioOne. DeepDyve is also indexing open access repositories like the arXiv and PubMed Central. Conspicuously missing from list are some of the biggest publishers that charge the highest pay-per-view prices, Elsevier and Springer.

One wonders whether some publishers feel that DeepDyve will cannibalize revenues by introducing a pricing model that focuses on article access, rather than journal or database access.

But DeepDyve sees their service as reaching to a unique potential user groups that have generally been underserved by academic publishers including individual knowledge workers and small businesses.  Indeed, the recent study of small and medium UK enterprises on their uses and desires for the professional and academic literature revealed that the price per article charged by many publishers was deemed excessive, considering that users can’t preview the full-text before purchase and that abstracts were often “uninformative or misleading,” requiring potential readers to “purchase blind.”  The rental model reduces the economic risk to the paying reader.

High prices have been one of the chief arguments promoting article deposit in public access archives.  A $30-50 article price tag for a layperson interested in the medical literature for personal or family health concerns may be excessive for many.  But $0.99?  One has to wonder whether DeepDyve’s existence a few years ago would have taken some of the steam out of the Federal Research Public Access Act of 2009 (S.1373).

The prices set per article take advantage of the psychological concept of anchoring.  With iTunes, we’ve been trained to see $0.99 as a price point, a reasonable price to pay for a unit of information, even though a two-minute pop song and a 10-page academic paper may share very little in common.  Still, others like Geoff Bilder of CrossRef have been promoting the idea of an iPubs service for some time as a way of dismantling the information silos built by publishers and libraries.

Outsiders to publishing and academic libraries may not view the development of services like DeepDyve to be significant and noteworthy, but I think this product may indicate that our views of information-as-property are finally changing.  What the Big Deal did to the library’s relationship with publishers — and supplanting our notion of ownership with licensing — DeepDyve may reorient the reader’s relationship with libraries and publishers.

Google doesn’t want our content.  DeepDyve does.

Reblog this post [with Zemanta]
Phil Davis

Phil Davis

Phil Davis is a publishing consultant specializing in the statistical analysis of citation, readership, publication and survey data. He has a Ph.D. in science communication from Cornell University (2010), extensive experience as a science librarian (1995-2006) and was trained as a life scientist.


33 Thoughts on "DeepDyve — iTunes comes to Science Publishing"

By happy coincidence, I just blogged about the high costs of purchasing PDFs for those beyond the paywalls. Perhaps this is one possible solution.

As someone at a small museum with very limited library resources but a very active research program, I’ll be looking into DeepDyve.

I don’t understand how they’re going to make us “give the PDF back” after 24 hours.

PDFs are not downloaded to your computer but viewed through DeepDyve’s software.

What does “renting” mean?

“Rented” articles can only be viewed at DeepDyve and cannot be downloaded, printed or shared. To download or print an article, DeepDyve conveniently includes a link alongside the “rental” button so that you can go directly to the publisher site to purchase or download the the article.

One librarian decided to look into this service a bit more closely and found the following results after a cursory exploration:

1. Many of the records are marked free – what you get on many of them is an abstract from PubMed, sometimes with a link to pay for the article from the publisher. Many free records are patents even though I marked “journal article” on that search.
2. Lots of records are Preview Only meaning you can’t “rent” them for $.99. Most of these do have a link to buy them from publisher.
3. On their records with a preview available, and with the $.99 rental available – many of the previews are not the first page of the article, but instead the abstract or in at least one case the cover page that some online journals have – not even the abstract.
4. I went ahead and rented an article for $.99, 24-hour access. This gives you the article in a flash player. You can never see more than ½ page at a time. In a fit of stubbornness I wanted to see if there was any way I could end up with the whole article saved – I could. To do it I had to do screen shots of the article, 2 screen shots per page and the article I rented was a 31 page review . Needless to say I didn’t bother to do the whole article, but you could…
5. Searching is very clunky, does mysterious things with what seems to be a web search.

The end evaluation: Not much of value here.

Obviously this service will need to improve vastly to fully compete with libraries at this point.

Fair comments, Jill. DeepDyve is still in beta with many new improvements in the pipeline both in features as well as additional content and we recognize we have much work still to do. One noteworthy point is that DeepDyve’s mission is not to compete with libraries but rather to provide a service to those ‘non-institutional’ users who do not have access to this content.

I serve on Highwire’s committee of publishers who look at these third parties who want in on our content, and believe me, there are hordes of them. There are a lot of folks out there looking to build a framework/store around journal content, basically piggyback on to the journals and siphon off a portion of revenue. I’m not convinced that services like DeepDyve offer much in return for that revenue.

First, many publishers see a significant revenue stream from “pay-per-view” access to their material. I’m not sure why they would want to share that stream with a third party who don’t seem to offer much more than a search engine and an e-commerce storefront. It’s unclear to me how large the market is that they’re going after (knowledge workers outside of subscribing institutions) and why those workers can’t use normal means like Google or PubMed to find articles.

Second, as you note, the 99 cent price seems completely arbitrary, based only on the previous use of that price point by iTunes. Why 99 cents? How does that equate with costs of production, revenue and usage of the material? The subscription plan seems even more troublesome. If I’m a librarian and I can get unlimited access to a journal’s content for $240 a year, I’m going to drop the much more expensive real subscription to the journal immediately.

Also, as noted by commenters above, reading an article for one day on a screen is probably not going to satisfy most customers. They want to buy a downloaded pdf that they can keep forever, or at least have a permanent printout of the article to refer to. What’s being offered here is more of a rental, a business model that has so far failed miserably for music, and is even less appealing here. Of course, the tech savvy reader will just make screen grabs of the article and get a permanent copy that way.

Even if you do buy into the rental concept, can anyone explain why it would be in a publisher’s interest to split revenue for this with a third party instead of just setting up a similar system themselves?

Excellent questions, David. In response:

1. We believe our service will complement, not replace PPV revenues. Our target user are the non-institutional users who find articles via Google and often comprise a majority of a publisher’s traffic. In our discussions with publishers, the average conversion rate of these visitors purchasing PPV is something less than 0.2%. In our interviews with focus groups, as well as the research report cited above (Small/Med UK Enterprises), the common reason for ‘abandonment’ of transaction is the friction of the overall purchase process and the cost of the article with only an abstract as the teaser. Based on our focus groups, the feedback is that users will in fact be more inclined to purchase once they can ‘rent-to-own’ or sample the article first. If you assume 2% of these visitor rent, you would only need 10% of these renters to later purchase (rent to own) to double your PPV business (10% of 2% = 0.2%).

Another reason cited for high abandonment is that aggregators such as Google and PubMed point users to the publishers but from the user perspective, there is considerable inconvenience as they must take their credit card out each time and navigate and transact separately. Imagine users going thru Google directly to book publishers if there were no Amazon. Part of iTunes success is the simplicity of their service where users can search for everything and with one-click impulsively purchase a song.

2. We believe the $0.99 has both symbolic value as well as proven transaction value. The $0.99 conveys to the user the service model. But iTunes has also shown that at that price, not only will users make impulse purchases, it is now affordable such that they will no longer pirate content. We believe that many of the 99.8% of users that do not PPV end up acquiring the content thru other means. iTunes has shown that $0.99 will make it morally convenient for users to respect copyrights and stop piracy.

3. We are hoping to create a win-win with our publisher partners where we can capture the 99.8% of those users who abandon the PPV transaction by offering them what would be very difficult for any individual publisher: a destination to look for everything; a convenient and “web 2.0” experience, and an affordable price point to satisfy their most fundamental need which is to read the full-text.

Thanks for filling in some of the details. Can you share the data from your research with potential publishing partners? I’m sure our review group at Highwire would appreciate some further data for analysis.

One worry this brings up is the idea that a rental is going to be good enough for the majority of viewers (if only 10% of renters end up buying as noted above). Presumably some of these potential renters in the non-institutional market are already buying via PPV, so if rentals suffice, that’s revenue that would have to be made up. Then there are your regular PPV customers in other markets, if renting satisfies most of their needs, that’s another big dent in PPV revenue that you’d need to replace.

I’m still not sure how the 99 cent price translates to this market. There’s a big difference in an impulse purchase of a song stuck in your head and a crucial piece of research you need to complete a project. There are also cultural differences in the market–we’re used to going to a book or cd reseller to buy books and music, very few go direct to the publisher/record company. But for journals, we’re already trained to go directly to the journal for access, not a third party aggregator.

As John Sack notes below, the value of a service like this is probably going to be very dependent upon the type of journal, the type of material published, and ease/pricing of the current PPV system one has in place.

The opportunity offered by DeepDyve to rent access to scholarly and technical publications has great potential for a variety of audiences, including independent medical professionals (dentists, pharmacists, doctors for example) who are not at large universities but who want to follow research in their field.

Technical writers also need access to research – and the topics they need change depending on the project. DeepDyve lets the tech writer adjust the technical field access needed.

Your article mentioned the lay public as a potential audience to benefit from the service — I agree, but how will they get to know about this new option?

We had asked DeepDyve for any data or studies they’ve done to try to quantify the size of this potential market, and they never got back to us. It’s also unclear how much of such a market would be relevant to a lot of scholarly journals, particularly ones that aren’t clinical (or aren’t about science for that matter).

Even if one accepts this as a valuable way to generate revenue, I still don’t understand why any journal publisher would want to split that revenue with a third party. Our customers are already used to coming directly to our journals for articles and for purchasing subscriptions/access. Most find our articles through things like Google, Google Scholar and PubMed. Why would a new search engine that very few have heard of supplant those widely used sources, and wouldn’t it be relatively easy for journals to implement their own rental system internally?

I did some playing with DeepDyve today (as a follow-up to my earlier comment), and must confess that the present incarnation is quite underwhelming (echoing Jill’s assessment). I focused my test search on topics relevant to my own research (dinosaurs), and mostly just found lots of links to free abstracts (nothing I couldn’t get with a Google search). Out of several hundred results, only two or three represented full articles that weren’t just links to abstracts or already available for free from the parent journals. Furthermore, most of the journals that “count” aren’t on-board with full content. I might consider subscribing if their content expands dramatically, but for now it’s just wait and see what happens.

Thank you to Jill for the review, and David for the analysis.

David: re your last comment. Surely DeepDyve must have presented *some* data regarding marketing size, potential uptake and estimated revenue generation to those publishers that they have already recruited prior to their inking any agreement? Kudos to whoever pitched the idea if not; selling fresh air is never a simple task.

IMO, researchers who cannot access the content they need at article level within their existing workflows (PPV or otherwise) are more now likely to attempt to source the information they want socially through the communities of purpose they participate in than they are to conclude that the answer is to seek for content within an article rental platform.

Cf. FriendFeed’s ‘References Wanted’ group by way of an example:

They pitched their offerings directly to Highwire, who then presented them (along with a group of other 3rd parties) to our volunteer committee of publishers who evaluate such things. We immediately questioned both the potential size of this market, and wanted to know more about the sorts of data and usage statistics they supply to partners. Highwire sent them our list of questions and received no response (the other 3rd party companies being reviewed immediately got back to us).

Another thing to note here is that these days, many journals release everything as free-access after 6 months or a year, so the number of potential articles in play for a system like this could be pretty limited.

HighWire did some analysis of the “unknown guest turnaways” by specific journal recently. We were trying to see whether the opportunity for non-institutional readers varied significantly by journal or journal type. We presented the results at last week’s HighWire Publishers’ meeting.

The market seems to be very different depending on whether the journal is a general science or medical journal, specialist basic science journal, or specialist clinical journal (we haven’t analyzed more types of journals at this point, such as humanities and social science). We found the opportunity ranged from 5% of use to about 50% of use.

As David Crotty commented, a journal’s policy for making back content free after 6 or 12 months will have a significant effect.

Several journals HighWire works with are willing to try an experiment of putting DeepDyve “rent” links on their own web site. This somewhat addresses the discoverability problem (i.e., potential readers not knowing to look in DD, because they use Google, Google Scholar or Pubmed).

We’ll be interested in seeing rental-reader reaction to the functional limits of a ‘rental’ model: no print, no copy/paste, no save/share. These could be problems for readers whose expectations were formed in universities where they had full access. I hope DeepDyve is able to gather than kind of feedback from the renters and will be transparent about the response to the model. It will be important information for the publishing community to learn from.


Thank you all for your many comments and questions! Rather than respond to each one, I posted a blog this morning that will hopefully address most of your feedback. Feel free to visit:

If you would like to discuss further, feel free to contact me directly at:

Why would this alter the need for FRPAA? The basic idea of that bill is that tax-funded research should be available as a public good. This is not my idea of “available.” It’s just another way to cash in on research that is often conducted by public employees and funded by public dollars.

My local government used taxpayer funds to build a new stadium for our professional sports franchise. The team still charges for tickets to the games. My taxes have paid to build toll roads, bridges and tunnels, yet I am still required to pay a fee each time I drive on one of them. My taxes have bailed out banks and automobile manufacturers, yet those banks still charge me fees, and no one has offered me a free car.

There is also a great deal of research that is paid for with private funds, and that is not conducted in America by Americans.

I am reminded of the MUMPS language created for the VA back in the 80s by SAIC. The government funded it for many years to maturity. It was open sourced a few years ago and lives happily in a user-supported community.

Publishing houses’ hegonomies have been carefully cultured to fit into a model of the economy. I don’t believe it is our responsibility to maintain the environment intact for their benefit.

It’s like owning a twenty foot python.

All publishers shouldn’t be lumped into the same pile. Many are parts of not-for-profit research centers who use publishing as a means to raise funds to pay for research. Others are the lifeblood of scientific socieities who provide great services for the community. These are things that are probably worth trying to preserve.

I’m not too crazy about spending money on sports franchises or toll roads, but billions of federal dollars are spent on basic research in order to further science – not put the research behind paywalls.

FRPAA has nothing to do with privately funded research – only with research funded by US federal grants.

We’re in agreement on the stadiums. I think that access to research is a complex issue, and there are many excellent arguments in favor of it. I just think the argument that because some of it is paid for by taxpayer funds that it must automatically be freely available to all is one that doesn’t hold water, given the nature of so many other things for which our taxes are used.

Let’s look at access to research like access to healthcare.

I hope there won’t be many dissenters if I posit an increase in GDP resulting from a healthy populace.

I’m not just talking about outsiders being admitted to the community. I’m saying that information as an ecology has resources that if opened to the community, could provide insights that are unlikely to happen in the current situation. So access is a generally healthy thing, but access for all will be unpredictable in terms of productivity.

Federating all data, irrespective of scale will happen in the near future. Denying access is something I wouldn’t want to be remembered for.

It’s certainly an admirable goal, the problem is finding a way to make it work, and to make it sustainable. So far, the author-pays open access models have proven unsustainable for journals other than high-volume, low editorial oversight models. If business models are invented that are truly sustainable, I think you’ll see many publishing houses happy to embrace them (just as so many are making all articles freely accessible after a fairly short time period).

Deciding what to keep in the future publishing ecology is indeed the problem. I recall Elseivier didn’t show up at the Open Access workshop at the AAAS meeting in 2005. At least their representative didn’t speak up when the audience was asked for comment.

I don’t think any publisher, no matter how corporate, is against open access. What they’re against is not making money. Keep an eye on how many publishers are currently moving to exploit the PLoS ONE business model for open access now that it’s proving so profitable.

Comments are closed.