path: root/transcripts/startup-science-2012/pete-binfield-peerj.mdwn

PeerJ Pete Binfield

I we can set a goal, to sequence the human genome for $99, then why not the
same for academic publishing? It just launched this week and we are fortunate
to hear the inside scoop of a very ambitious new model.

We launched on Tuesday. The people behind it are myself, the one journal to
rule the world was justm entioned and I used to run PLOS One, and it actually
is the one journal to rule the world. It's going to publish about 3% this year,
about 25,000 articles. I left four weeks ago to do this. My co-founder is Jason
from Mendeley. He left back in December. We've been working on this for about 6
months. It was originally his concept.

So uh. I thought I'd give you a quick overview about what academic publishing.
Not everyone is knee-deep. This is our original pitch deck to VCs. Academic
researchers have an obligation to publish in order to advance their careers.
The scholarly journal market is dominated by subscription publishers that
charge thousands of dollars for an annual subscription, and you have to hand
over the copyright and content to the publishers. So it's artificially
restricted. That's a $10 billion/year industry and highly profitable. It's
subsidized by the free labor of the academics that write for free, they review
for free and so on.

The interesting development in thelast 10 years has been open access whih is
about 10 years old now. It wants to distribute content freely without
restrictions, like with Creative Commons. The model is that the author pays and
the author pays the processing charge. Thsi is typically in the range of
$1300-$5000 per article they publish. This pays publication costs.

What is open access? In the journal world it's a distribution model, not a
business model. The author pays thing is not the concept of open access. Open
access is usually licensed under CC. The broadest definitino is CC-BY 3.0.
Although some publishers try to add a NC term to it. This is the gold standard
of open access licensing to appyl. If you apply that license, your content can
be reused, remixed and redistributed for free provided that you acknowledge the
original service.

Commercial products and applications can be built on top of OA content,
provided that you have not applied the NC subterm. Anyone can repackage and
sell it off. We want to encourage more of that because it hasn't been happening
in the OA world. OA is rapidly disrupting the established subscription model.

OA honestly improves accelerates and facilitates academic progress. It's a good
thing in the world and the world deserves its academic content to be published
like that, how can you build a business model like this and how d oyou persuade
the world? OA is disruptive. This is a recentp aper that is trying to predict
the disruption curve of OA. It's about 15% of the world' scontent right now.
This study predicts that by 2017 that about 50% of all scholarly articles will
be OA. By 2019 it might be 90%. They did this by extrapolating the Christenson
disruption curve. The speed of which.. has been growing.. PLOS One has been
dubling in size, 3% this month but next month might be 5%. If you roll that out
over OA in general, I wouldn't be surprised that within 5 or 8 years that we
will have complete disruption of the publishing industry.

This is the curve for PLOS One through the end of 2011. It's publishing a
pheneomonal amount, about 2k articles/month and that's growing each month.

So what is PeerJ? We are a new OA publisher, we are publisher of an academic
journal that looks a lot like PLOS ONe. It's PeerJ. We have a preprint server
called PeerJ Preprint, and teh two work together. Authors can submit to the
preprint server, or others can submit directly to the jounral. The two are
tightly integrated. The most interesting thing is that.. I mentioned that the
standard business model is that authors pay for publishing each time, and that
payment of $1300-$5000.. we are shifting that model so that you don't pay to
publicize, you pay to become a member of PeerJ, you subscribe and by being a
member you get lifetime rights to publish future papers with us for free. It's
a very different psychological motivator.

How do you incentivze peer review? There are some innovations there as well. In
order to remain a member in good standing, you must review once per year, it
can be a peer review, it can be a comment on an article. We want to encourage
the community to contribute to the community. The end result is that everything
is OA under a CC license.

I thought that this would be interesting, this is straight off of our website.
These are the guiding principles that we are building PeerJ from. We are an
innovative company that serves the academic community that attetmps to drive
costs down. Literally we believe that we are going to take the best features of
traditional publishing worlds, all the traditional benefits around archiving,
peer review, versions of records, we are going to marry that up to modern
internet technology to improve that process. We are a service industry for
academia. WE serve academia and not the other way around. THis is what's needed
to get the company off the ground. We want the prices to get down and lower.

What are the prices? We launched on Tuesday, if you become a basic member,
that's $99 and that's a lifetime right to publish one paper a year with the
journal. That remains to be seen. I think, we're pretty confident. Enhanced is
$199 and that's 2 papers/year. The $300 is the right to publish an unlimited
number of papers per year. You can upgrade at any point, the catch is that all
contributory authors have to be a member. Each of the authors on the paper have
to be a member in good standing. So all their future papers are free. There is
also this requirement to do a review every year per member.

Some of the interesting things is that all of our content is CC-BY. We will
make a sfotware platform for peer review, we have Public APIs, we are going to
open source a bunch of our stuff, and it will be a cultur eof innovation.
Anyone that wants to build tools to interoperate with ourselves, we'd love to
talk with you. We launched on Tuesday, and this is the great reception we got.
We got great media coverage, we lined up interviews in Science Magazine, Times
Magazine, Nature, we got started with an overwhelmigly positive response. There
are some doubters coming out of the woodworks but in general everyone has been
very positive.

The award for the best headline is that: New OA Journal Lets Scientists Publish
Until They Perish. That's PeerJ. Thank you.

So about oA in general? What can we do to promote it? When I talk about the
startup science community, everyone loves OA. But when I talk with people in
the academic community, I get shoulder shrugging because they don't understand
what the world is missing out on. We have to engage the academic community and
they are the oens doing editing and writing. How can we actually get the
academic community to understand the importance of this movement?

There are many reasons for why the academics aren't embracing OA. One of the
ways that this community in this room could help to accelerate this movement.
The OA community believes that OA content is better, that it is going to
improve the dissemination of the content, that it will facilitate faster
discoveries.. the problem so far is that we haven't proven that point, we
haven't built killer apps or groundbreaking discoveries based on OA content.
There have been some smallish examples but nothing earth-shattering. Perhaps
that is because we are only getting 15% of the content. But some killer app
that took OA content and made it much more valuable to the world and much
better, then academics would see why they should publish in OA in the first
place.

I have a bunch of questions, but to save time, I'll pick one. One thing that I
am curious about is do you have a ranking system, a way to filter out, I'm not
goign to say that there's going to be drivvel, I think the accountability
system will deincentivize that.. clearly I have a certain set of keywords I
want. Is there a filtering system to bring the relevant and most credible
papers, the papers with the hgihest reviews post publication?

Yes, there is. We won't necessarily have a proprietary secret sauce algorithm,
but we will use something else. The metrics movement is attempting to solve
this problem for us. They are coming up with alternative metrics around who
tweeted about it, who wrote about it, etc. We are going to be building that
into the system, expect to see the metrics, to see what filters out good and
bad or what's relevant, etc.

First of all, thank you. This is a really great model. I'm not sure what it
looks like as a business, but as a service it seems really awesome. I
appreciate it. I like how it incentives longevity, so that's pretty great. So
what does it look like as a business? How many new grad students every year?

There are 10M publishing academics. The way it breaks up is that there's 25k
journals that publish 1.5M articles per year in academia. STM is about 1M
articles per year. They are being published by about 10M practicing academics
worldwide. So the audience is possibly 10M and lots of grad students come in
and out, and that's hard to find out. I would say about 500k come in each year,
and 500k drop out each year, it's about that order of magnitude.

There is some secondary other alternatives around this. Down the line there
would be advertising on the site, yes. But the primary business model is the
subscription or membership.

Is there any worry or thought about making sure that the people are who they
say they are on their system? Is that a problem? Do you think it matters? Like
author identifiation, not having more than one account or something.

Every member who is going to do anything significant will be a publishing
author of some sort. There are industry requirements, like the Requirements for
Authorship, you have to contribute to the paper, you have to be cited, yo uhave
to verify that you are happy to be a coauthor on the paper. We will also be
using unique author identifiers, as an industry initiative called Orchid, they
will have their beta in September, and that's an industry-wide adopted
initiative to give every academic a unique identifier and to track their
progress centrally.

One of these problems is that it's a walled-garden where you create an
artificial garden, a wall, the price point to enter the system. Walled gardens
are very similar to the way in which previous journals and their peer reviewed
networks have become stalled in their time. The advantage of one that doesn't
have a walled garden which doesn't have such an initial requirement, like the
pervious gentleman's discussion, is that it allows for more chaos, but which is
why you need stronger means of authenticating individuals, like the previous
gentleman at the mic suggested. How do you avoid with your walled garden
approach at PeerJ, running into the same problem with peer review where it
calcifies again where it becomes non-transparent because the incentives to
maintain the walled garden, where it becomes difficult to penetrate the .. for
judging the way they do?

So potentially, I might have misunderstood.. allr eviewers on the paper do not
need to be members of the company as it were. Peer review will be sourced from
anywhere. It will not be members only reviewing on members. The peer review
will be independent. They can find peer reviewers from anywhere in the world.

It's still a walled garden because the people who invite the peer reviewers can
have an agreement with thsoe one. Here's the same issue.. it doesn't matter if
it's interactive or this what you're trying to build.. there's an incentivize
or revenue incentive to make it less transparent because embarrassment to
really cover it up.

We are going to be encouraging open peer review, and we want people to
publically provide their names. The authors will give the option to publish the
whole peer review history. So they will be able to publish their peer review
comments as well. Those will be published along side the paper to make it more
transparent. But it's an option.

With an open system, you have to have all of the leemnts open to have it work.
If you have hidden systems, then that selectively biases the outcome or the
appearance of each.. I'm trying to illuminate this with the difference between
closed/open systems.. .t his is a delimma in how we maintain the authentic
nature of feedback mechanism. That is how the scientists trusts the nature o
their quality o their peer review. It gets back to the problem. 

I think we are more open than the typical journal. Most of those peer reviews
are never published. It's a problem with academia in general.

Did you say you had a proprietary model of rating the journals? Yes, we have a
custom software system for peer review. Why is that internal proprietary
system? We want to open source it, we will open source a lot of our software,
but we're not sure how much will be open soruce. We have to buidl it for our
own purposes, so that we can control the process to build it out the doors,
down the line it's potentially open source sure.

Pretend that this is the last century.. is the mechanism that you are building
is also going to be a publication for Thomas Edison, Gregor Mendel or citizen
scientists? Perhaps not Edison... but sure, you don't have to be credentialed
to publish papers. PLOS One will publish high school student papers and
undergrads, as long as it deserves to be published, so yeah. But what about the
peer reviewers? We will definitely be enforcing some standards.

Based upon your expeeriences iwth PLOS One, there are many textual searches
instead of human curration to determine biological entwroks/ Are there any kind
of comparison study that has been done on using only open source or open access
literature, as the basis for definition versus the traditional sources? It
would be interesting to see teh quality metric.

There is an expert on this. Hedler in the corner is an expert in data mining
and published papers. I think part of the issue is that the proportion of OA
literature is always including some closed access stuff. Hedler will know
whether there are some studies like that.

So, as a scientist, I know that, frequently people pubilsh and cross the line
of very low impact work because they can't publish it anywhere else. How are
you going to balance the reputation, like.. is it your goal to make this a high
impact journal?

No, it's meant to be very PLOS One like. It will accept anything that is
acceptable and deserves to join the literature. It can be papers about
failures. PLOS One is not a dumping ground for the bad stuff.. it's about
everything. It's not ust about the impact factor. But the average impact factor
is greatr than 90% of the journals in the world already. On average the content
is not awful. It's actually just fine. They found that 75% of all authors went
to PLOS One as their first choice. It wasn't down the chain of their
rejections. We want the world to decide what was good enough, what was useful,
etc.

What about cross genetics, higher impact session, and have tiered version or
something?

It's just the one journal to contain everything. Within that one journal, maybe
we can say here's the bestof in genetics, in comp bio, we'll be able to do that
with just a display mechanism. We don't intend to have separate branded
journals that papers get dumped into. No.