A note on how wages might be affected as computers do more and more journalism jobs

For readers interested in the larger implications of automating jobs that journalists and many other office workers do, I recommend this brief post by economist Paul Krugman. 

It’s a bit technical, but if you work through it the point he is making should be clear. 

If we are at the beginning of a widespread change that results in more and more jobs being done by networks of high-speed computers, the wages earned by good old-fashioned human beings will start to decline.

Why it makes economic sense for Google to build computer programs that do the jobs that journalists do

A quick note.  Google appears to be expanding its computer programs that do the work that journalists do.

These programs are called digital journalism tools, and the keywords here are digital and tool.

And once again, journalists and journalism organizations are helping Google take away their jobs.

Why do I suspect this - Google is offering journalists and journalism students a 10-week fellowship for $7,500 and $1,000 in travel funds. The price someone is willing to pay journalists for their labor is a key economic signal - it tells the journalists exactly how much they are worth to the company that is paying them. 

That means at this moment Google and the foundations are willing to pay journalists about $8,500 for 10 weeks of work.

Google only makes $0.02 a search, so the journalism foundations listed in Google's press release are subsidizing the fellowships. (I'm not linking because the last thing I want to do is help Google).

The subsidy might be needed because at $0.02 a search, each journalist has to generate 425,000 searches before the work they do is worth any money.

The small amount of revenue generated by each search is the reason that Google must develop computer programs that are more efficient and capable than any human journalists can be.  The programs at Google must compete with computer programs at other new media companies that are also learning to do the same jobs that journalists do.

What are these companies competing for?  To drive traffic to websites where digital advertising tools are used to sell and deliver advertising. The keywords here are digital and advertising. 

I wonder if the foundations financing this venture have offered to do it again next summer. Google will has every reason to keep working with the foundations so long as the fellowships help Google refine its digital journalism tools, making the tools better and better and faster and faster at doing jobs that used to be done by people.

Meanwhile, journalists at all the newspapers that recently abandoned their print editions are about to find themselves competing against computers that drive traffic to websites and sell advertising. And when people compete against machines, machines always win. 

Newspapers will put themselves out of business if they take advice from commentators who are telling them to become “more like” new media

Subheads in this post:

1. Each newspaper should charge its readers because prices are a critical economic signal that newspapers need to survive

2. Why new media companies have not reached the point when they can began charging for access to their products

3.New media companies have a business model designed to automate journalism jobs

I started this series of posts because newspapers are finally starting to charge online readers for access to news and advertising.

Each newspaper should charge its readers because prices are a critical economic signal that newspapers need to survive

The prices that readers pay tell each newspaper how much readers value that particular newspaper’s news and advertising.  There is no other way that newspapers can figure this out.

Each newspaper must know before it creates a news story if there will be enough revenue to cover the costs of producing that story. If the newspaper does not know how much revenue the story will generate, it also does not know if it can pay the journalist who creates the story.

So executives watch how prices change to figure out if they should hire journalists to produce news stories. If prices are going up, that’s a signal to produce more news stories and hire more journalists.  If prices are going down, that’s a signal to stop producing news stories and stop hiring journalists.

If the newspaper doesn’t charge readers anything at all, the executives don’t have any way to tell how many journalists they should hire to produce news stories. (It’s actually more complicated, but this is the basic idea).

Of course, we are far past the point of fine adjustments in the number of journalists a newspaper hires. Newspaper executives have been forced to lay off thousands of journalists and other employees because they are being devastated by competition from new media companies.

The fact that newspapers were not charging for online access probably made the problem worse.  Executives had no way to tell how many journalists they would need tomorrow to produce the news. All they knew was that prices for advertising – a distant second best for detecting the value of a news story – kept going down.

So charging readers is a critical tool for executives who are trying to make their business model work and keep their newspapers alive. Those executives need every tool they can find because the competition is so lopsided.

The newspaper executives are competing with new media executives. As I explained in my last post, the new media business model is completely different from the older media business model.

 

Why new media companies have not reached the point when they can began charging for access to their products

If new media companies charge for access they will hide the signals that they are looking for. The signals new media companies are finding are patterns in the jobs that journalists do each day.

 

New media companies give all journalists free access because they are trying to record everything that all journalists do all day long. Once the new media company has recorded enough actions by enough journalists, it uses computers to find patterns in the jobs journalists do. Patterns are required so the company can automate journalism jobs.

 

New media companies have a business model designed to automate journalism jobs
 

Once a job is automated, it becomes almost impossible for people to compete.  A high speed computer can almost always do the job faster, cheaper and better no matter how hard the person tries.

Here are some jobs done by people who work at newspapers that new media have already automated – advertising salesperson, advertising creator, advertising marketing researcher.

Many journalism jobs have also been automated. One example is the police reporter’s job. The job has 5 big steps:

 

1. go to police station. 2 find all police reports. 3. sort through all police reports to see where and when different kinds of crimes were committed. 4. go back to paper. 5. write story that tells readers about a handful of crimes that happened in a handful of neighborhoods.

 

Automation eliminates two steps – 1. go to police station, and 4. go back to paper. 

 

Automation is faster and better at the remaining steps. Automation can tell readers about every crime that happened in every neighborhood. All readers have to do is ask.  Then the computer goes to work:

 

3. sort through all police reports to find the crimes committed in the neighborhood the reader asked about. 5. write a story that tells the reader exactly what they want to know.

 

You can try it yourself. Here is just one example of a computer hard at work, doing its job as a police reporter.

Newspapers are competing for their economic life.  Commentators who tell newspapers that they should learn to be “more like” new media are talking through their hats.

If newspapers become more like new media that will ensure that new media win the competition.  That will also ensure there are even fewer jobs for the people that newspapers employ.

When two products are similar, consumers usually select the one that costs less. Consumers always select that one that costs less if it is actually a better product.  And the best way to make a better product that costs less is to use automation.

So if newspapers want to become more like new media they will also have to become experts at automation. Newspapers will have to learn to automate all of the jobs that they currently hire people to do.

 

At that point, the competition might look like something like this:

The best computer programmers in the world vs. journalists trying to learn how to program computers

Who do you think will win?

Why new media can’t replace all the news and information created by old media

My last post mentioned that new media companies would collapse if they had to produce original news and other information. This point is often lost in the debate about newspapers and other pre-internet media companies charging for access to news and information.

Older media companies should and do use many of the cost-saving tools and techniques developed by new media companies. But this is not a two-way street. New media companies cannot afford to begin producing original news, advertising and other information.

All companies use the same business model: Revenue – Cost = Profit. Another version of the model is useful for understanding the differences between older and new media companies: 

Average Revenue – Average Cost = Average Profit

Google and Facebook are new media companies, and A.H. Belo is an older media company.  All three companies are in the advertising and information business, but in very different ways.

Google primarily produces free search results for internet users around the globe. Google generates most of its revenue selling advertising. Facebook offers free membership in a social media website for internet users around the globe. Facebook generates revenue selling advertising and virtual and digital products. Belo’s primary business publishes 4 daily newspapers in U.S. metropolitan markets. Belo generates most of its revenue selling subscriptions and advertising.

Each company illustrates the business model for the industry where it operates.

		Google	Facebook	A.H. Belo
Total production		1,722,071,000,000	845,000,000	686,468
Unit		(searches)	(members)	(circulation)*
Total revenue		$36,531,000,000	$3,711,000,000	$422,513,000
Revenue/unit		$0.02	$4.39	$615.49

^*print and digital
Based on company financial reports for 2011-12. Google searches from statisticsbrain.com.

Total revenue and production at the new media companies overshadow the newspaper company.  But the averages show the new media companies need very large production to generate their billions in revenue.

If the new media companies were the same size as Belo, Google would have just $14,562 in annual revenue. Facebook’s annual revenue would be about $3 million.

So the much higher average revenue at the newspaper company appears to be an advantage.  But this is misleading.

Low average revenue at the new media companies means they must also keep production costs low. New media companies solve this problem with automation. They are experts at using computers and high speed internet connections to produce and deliver search results and social media pages. Employees create or maintain the automatic processes (computer programs and hardware) that do most of the work.

The production and delivery of original news and information is much less amenable to automation. Each news story must be created by employees who gather and format original information. Employees must work continuously to create a steady stream of news. Production and delivery of printed copies of a newspaper is also labor intensive.

		Google	Facebook	A.H. Belo
Employees		32,467	3,539	2,100
Total revenue		$36,531,000,000	$3,711,000,000	$422,513,000
Revenue/employee		$1,125,173	$1,048,601	$201,196

The second table shows the advantages of automation. Revenue per employee at the new media companies is more than four times larger than the newspaper company.

Doesn’t higher productivity mean new media companies can hire employees to produce original news and information if the companies that produce information go out of business?

No.

First, these figures don’t account for the costs of automation.  New media companies have enormous banks of computers scattered around the globe so they can deliver search results and web pages in the blink of an eye.

Second, companies only hire workers who can help increase profits. So new media companies would only produce original information if those employees matched the productivity of computer programmers and technicians.

But employees who produce original news and information have much lower productivity than employees who create or maintain automated production processes. The last table shows what would happen if productivity at the new media companies was as low as productivity at the newspaper company.

		Google	Facebook	A.H. Belo
Production per employee		53,040,657	238,768	327
unit		(searches)	(members)	(circulation)
Employees needed if each produces 327 units		5,266,272,171	2,584,098	2,100

If new media companies had the same productivity as the newspaper, the search engine would need more than 5 billion employees and the social media site would need more than 2.5 million employees. Obviously, the new media companies cannot afford to produce all of the news and information that they need to be successful.

New media companies are well aware of their dependence on free access to an enormous variety of information. That is why they support free expression and work to maintain free access to information created or posted on the web sites that they own or index.

But even if the new media companies wanted to produce the information they require, they could not afford it.  They could not even come close.

It would be helpful if more participants in the older vs. new media debate understood the actual business models. Perhaps they could move beyond the current unproductive discussion.  Perhaps they could focus on the best ways to ensure consumers have continued access to the useful products that all of these firms provide.

(Employee totals at Google and Belo include some who work in secondary businesses at these firms.)

 

Newspaper paywalls are a rational response to competition - what took so long?

Sometimes it takes a long time, but in the end rationality will have its say.  This appears to be the case for some prominent members of the newspaper industry, which has struggled to adapt to competition from the internet.

The steady stream of reports that more and more newspaper companies like the New York Times, Gannett, and E.W. Scripps are beginning to charge some customers to access news and advertising shows the executives making these decisions are starting to figure out the economics of the internet.

The tactics may not work for every newspaper.  When the tactics do work, newspapers are unlikely to ever regain dominance in their respective markets. But that is not the point. The markets where newspapers operate changed and became more competitive, and newspapers have to figure out how to compete.

Papers are starting to divide readers into groups based on each group’s willingness to pay for access to news.  Some readers only read a few articles from time to time, so they are unlikely to pay anything for access.  Newspapers can satisfy these readers by giving everyone free access to a limited number of articles – 10 or 15 seems to be a common number – each month.

But other people want access more often and will pay if they have no other choice.  Some people want to regularly read specific kinds of news and advertising, or perhaps they will pay for the convenience of access in multiple delivery channels – print, the internet, mobile devices.  Each group will be willing to pay a different amount depending on the strength and characteristics of their demand.  This means newspapers can charge each of these groups different amounts for access to news.  The different prices for various combinations of apps, mobile access, web access and print show that newspapers are starting to do just that.

Economists call this price discrimination, and I understand why that’s not a great way to market one of these new pricing plans.  But the industry has come up with an even worse way to describe them – paywalls.

The industry’s inelegant term does, however, remind us that part of making price discrimination work requires that newspapers first stop allowing everyone to access all of their content for free. Scholars such as Steve Wildman, Danel H. Simon & Vrinday Kadiyali and Florian Stahl(et al), have pointed out that newspapers are competing with themselves if they charge for access to print editions and also give away news on the internet.  Once readers figured out what was happening, they began to abandon the paid option in print.

There are many non-trivial reasons the newspapers that are changing might have stayed with this irrational strategy for so long.  It takes time to react to and understand new forms of competition. Just acquiring and learning to use the technology that allows price discrimination probably required substantial time and effort.

Another reason might have been that newspapers were told – and are still being told – they should not charge readers for access.  Newspapers are told not to charge for access because internet-based media companies make money without charging for access to news and other information.  But search engines and social media have a very different business model, and that model would collapse if those companies had to pay for the creation of content as newspaper companies do.

So it was good to see this week that newspaper executives are talking about things like subscriber retention and creating different bundles of digital and print subscriptions (another kind of price discrimination) that focus on high quality news. It will still be a long, hard slog, but at least they are giving themselves a fighting chance.