Is Scientific Publishing About To Be Disrupted? Michael Nielsen @michael_nielsen STM Conference, October 2009.

Is Scientific PublishingAbout To BeDisrupted?

Michael Nielsen@michael_nielsen

http://michaelnielsen.org/blogSTM Conference, October 2009

Kongo Gumi

578 CE

Prince Shotoku

Shitenno-Ji Temple

almost 1,500 years

2005:100+ employees

70 million $ revenue

Masakazu Kongo40th generation

2006:Liquidation

Purchased by Takamatsu

As an independent entity,Kongo Gumi no longer exists

How is it that large, powerfulorganizations, with access to vast

sums of money, and many talented,hardworking people, can

simply disappear?

most interesting when an entire industry is disrupted

Data General

None of these companies exist

CD Sales

Napster founded

Firstquarter2009

First part of talkWhy these disruptions happenHow we can recognize themSecond part of talk

Scientific publishing is in the (very)early days of such a disruption

Common explanations ofdisruption1. The people in charge are stupid.

Why couldn’t the record companies see things coming…

Napster

iTunes

Last.fm

pre-empt them by doingsomething similar first

2. The people in charge are malevolent.

stupidity and malevolencesometimes play a role…

…it’s a mistake to base an explanation on these factors.

underlying structural reasons cause the failure

If you look at the newspapersand record companies andsee stupid and malevolent

people…

But if disruption can destroy

then it can destroy anybody

Why are the top blogs thriving financially, while the

newspapers are dying?

news parasites

TechCrunch

Top 100 blogs in the world

Started in 2005

arguably the best reportingin the technology industry

TechCrunch is thriving

The New York Times is wilting

Operating income down 50% first quarter

TechCrunch’s operatingcosts are far lower, per word

depressing the price of advertising

Increased supply of ad space

Old supply of published material

ad space

decreased price

there’s a limited amount thenewspapers can do to makethemselves cheaper to run

~ $100s per photo

~ $10 / photo

TechCrunch isn’t being any smarterthan the newspapers

quality photography can help establisha superior newspaper brand

worth $$$

makes business senseto spend ~ $100s / photo

Organizational architecture

(Diff colours = diff skills)

Most newspapers usea very similar architecture

Compete with small variations

Good photographersare worth every cent

An opportunity for anew type of organization

New technology (internet)

radically different human skills

radically different structure

No staff photographerNo office (until recently)

No print room

no wonder it’s lower cost

What can you do?

destroy morale of all your staff

stir up the Unions

give a competitive advantageto your newspaper competitors

still be paying far more, per word, than TechCrunch

product will be no more competitive

new optimum could not have existed 20 years ago

you can’t get there without going through the valley

incremental actions wouldbe hell on a newspaper

locked in

comparable quality

higher price

~ $100s / photo

entirely sensible business decision

NOT

Standard Org.Architecture

New Org.Architecture

Radicallydifferent skills

& structure

Near-impossible to get from onearchitecture to the other

Easier to start over

Change is harder even than I’ve said

forces which suppress change

Large, complex structures can only lastwith forces that preserve that structure

Organizational immune system

good thing

When major changes are needed, the immune system will attack

those changes

attacked from within

attacked by their competitors

Only a company outside theindustry could have done it

AndrewRosenthal

Classic immune response

Deep commitment to quality journalism

values which have madethe New York Times great

The NYT can keep itsPulitzer Prizes

the last people to know an industry is dead are

the people in it

internalized the values, normsand collective knowledge

immune response is strongest

If a person inside an industry needs tofrequently explain why it’s not dead,

they’re almost certainly wrong

How can you tell if an industry is about to be disrupted?

fundamental problem is thatit’s hard to recognize the early

stages of disruption

newspapers laughed at thenotion of online news and ad sites

as competitors just a few years ago

serve a similar basic need

craigslist

craigslist

different org. architecture

People outside the newspaperindustry were willing to bettheir own money on them

Most such startups die

That’s how the new industrylearns what organizational

architectures work

But if even a few survive, theincumbents are in big trouble

lot more room for improvement

What’s this got to do withscientific publishing?

TodayEditorial

Co-ordinationDistribution

Production companies

10-20 yearsTechnology companies

Sales & marketing

Not just that they’ll use technology

Nor that they’ll have a large IT staff

technology-driven

their core competency andcore way of adding value is technological innovation

key decisions are being madeby people whose background

is technology

Didn’t people already predictsomething like this with

preprint archives?

Today, there is a flourishingecology of startups

Acquired by Royal Society of Chemistry

many of the same people as Last.fm

science blogs

easy to miss impact on research

most science blogs focus onoutreach

Tim Gowers

open sourcemathematics

entirely in the open

37 days

27 people

800 comments

170,000 words

Gowers: “this processis to normal research as

driving is to pushing a car”

4 of the 42 living Fieldsmedallists have blogs

hundreds of research blogs

Some of the world’s topscientists are spending manyhundreds of hours per year

on their blogs

open notebook science

Pioneered by chemistsJean-Claude Bradley and

Cameron Neylon

Tobias Osborne

Garrett Lisi

Steve Koch

Second sign of move from productioncompany to technology company

Changing nature of information

static entity

add value through content,production and distribution

online

cost of distribution, production and content have dropped dramatically

information is...

waking up

coming alive

people who add most value areno longer the people who doproduction and distribution

technology people

programmers

Look at where profits are migratingin other media industries

profound technical ability

How many scientific publishersare run by people who know…

…the differenceb/w INNER JOIN

and OUTER JOIN… what an A/B test is?

… how to set up a Hadoop cluster?

myriad of services to develop

many are being considered by startups

Not difficult, technically

The Amazon algorithmis public (Linden, Smith and York,IEEE Internet Computing, 2003)

PageRank

personalized PageRank

personalized search

personalized recommendations

Google doesn’t use it

too computationally intensive

… even for Google, onbillions of pages

but on the scientific literature?

Not yet done

Automatic translation

Full-text searchGood relevancy and importance ranking

RSS feeds

Spell checkingSynonyms

Alerting services

Analytics

Data miningInference of emerging areas

Factual inferenceOpen API

not yet widely adopted by scientiststrivial reasons

collaboration platforms will emerge

Timestamped

LaTeX integration

Version history

Citable

Integrate naturally with platformslike wordpress.org

Where will they be in 10, 20 years time?

Someone needs to do the same for science

Citable

Permanent

Versioned and timestamped

Searchable

Huge opportunities

Technology leads

Most ideas will fail

The people who succeedwill be those with the deepest

understanding, not the deepest pockets

Difficult for publishersto get involved

If not them, it will be someone else

Thankyouhttp://michaelnielsen.org/blog/?p=629

http://michaelnielsen.org/blog@michael_nielsen