Earlier this week, the New York Times published an article with the headline: "Google Made $4.7 Billion From the News Industry in 2018, Study Says". It was based on a report from the News Media Alliance as part of their lobbying campaign against the tech industry, where they are hoping to convince politicians to give them the power to force Google to pay publishers for the links that they use.
Not surprisingly, this very catchy headline and massive number got a lot of attention within media circles, with journalists and editors promoting it uncritically as part of their annoyance with Google.
But then, shortly after, analysts like me started looking at it, and we quickly spotted a number of very big problems with it. Mostly that the article contained no real explanation for any of the statements provided, and that it lacked any form of journalistic checks that we would otherwise put on such coverage if the report had been published by any other industry. And then, of course, came the realization that the basis for this $4.7 billion figure was extrapolated using some very sketchy data from 2008.
The News Media Alliance based its new report partly on a study done by the economics consulting firm Keystone Strategy. Keystone Strategy relies on a statistic that was made public in 2008, when a Google executive estimated that Google News brought in $100 million.
And when we looked into this further (which the NYT failed to do), we discovered that it was the then vice-president of Google, Marissa Mayer who threw this out during 'lunch session' at a Fortune conference.
Google News is free and has zero ads. So what's it worth to Google? About $100 million.
That's the figure Google vice president Marissa Mayer, who heads search products and user experience, threw out during a Tuesday lunch session at Fortune's Brainstorm Tech conference in Half Moon Bay, Calif. How does she put a value on a product that doesn't directly make money? The online giant figures that Google News funnels readers over to the main Google search engine, where they do searches that produce ads. And that's a nice business. Think of Google News as a $100 million search referral machine.
So this data that the News Media Alliance is using to tell politicians how much Google is earning from news links is really sketchy, and is not actually a measure of revenue but rather an estimated value for a 'funnel' towards Google Search as a whole.
Mind you, I'm not just annoyed at the News Media Alliance about this. It's really poor data, but my bigger annoyance is with the New York Times and how they published what was very clearly part of a media lobby campaign which they tried to disguise as journalism.
This is exactly the kind of thing that people will use to say that the NYT isn't trustworthy.
As Stasys Bielinis from Techmeme tweeted.
An NYT story about MSM lobbying body's weird "study" with super clickbaity $4.7B number based on another questionable study by some little known consulting firm based on an off-hand remark by Marissa Mayer in 2008, 11 years ago ... ?
Released just ahead of the House hearings where it will serve as talking points for US House elected representatives to beat Google up and look good on TV and mainstream papers and news sites for the next few news cycles?
NYT asked no Qs about the issues with this study? Just the "Study: Google Made $4.7B Billion From the News Industry in 2018" headline? Conflict of interest, anyone? How does this lousy reported and edited thing help counter the narrative prevalent on the right of MSM's bias?
PS: If you are wondering where that original "Google earns $100M from GNews" number came from:
And that number is so easily debunked/challenged when it is used for the purpose of lobbying to levy tolls on Google:
Others, like Dave Winer said that this story should have been posted as an ad.
As a paying subscriber to the NYT (not to Google) I am offended by this. They should explain how the number was arrived at in more detail, up front. And then they shouldn't run it as news. Make it an ad, cause that's what it is.
Both of them are right. This hurts all of us. It impacts our overall trustworthiness, it impacts the role people feel we have in the world, it undermines our arguments ... and worst of all, it might even end up causing more harm because it will be so easy for Google to make publishers look like amateurs when it comes to explaining their case.
It's just frustrating.
But what about the number itself? The $4.7 billion? How (in)accurate is it?
Well, I asked the News Alliance for a copy of the full report(which they sent to me, so thank you for that). And here we can read about the full calculation.
As you can see, the calculation starts out by comparing Google's ad revenue from their own properties (a specific listing in their annual report, so that's accurate, although it's rounded to the nearest billion here).
Then they take the $100 million Marissa Mayer mentioned during that lunch session in 2008, and calculate from that the share of the revenue attributed to Google News is 0.7%.
That's not all bad. But remember, Marissa didn't specifically say it was Google's News' share of the revenue. She talked about it as the referral value. But... I mean, I can go with this.
But then they have the Google News to Google Search scaling factor, which is set to be a factor of six, leading to the estimate that news' impact on Google Search is 4.2% of the total activity.
So how did they get to this factor? Well, they got it from looking at their own internal referrals.
The "Google News to Google Search Scaling Factor" in column [E] is estimated based on referral traffic from internal referral data from major publishers; based on comparing six full calendar years of traffic from two publishers, the median factor differential between referrals from Google News and Google Search is 8.2 and the minimum is 6.2. From these, 6.0 is selected as a conservative estimate in the difference in traffic.
Wait... what? You can't do statistics like this.
We know that the way people interact with news is massively different from how they interact with other forms of content. For instance, the referral patterns for a webshop that sells mountain bikes will look massively different from how people are referred to the New York Times. So you cannot extrapolate data like this.
Also, this referral data is based on only two publishers.
I can understand what they are trying to do here. They looked at the volume of referrals coming from Google News and those from Google Search, and then they compared the two. This told them that Google Search refers 6.2 to 8.2 times as many clicks.
They then conclude that, if Google News was worth $100 million in 2008, Google Search, which drives 6+ times more clicks for publishers, must be worth six times that.
They argue for this comparison in the report.
While the number of unique visitors to Google News grew at an average rate of 23% per year between 2009 and 2018, over that same time period Google revenue from Google properties has grown on average 22% per year.
This similar growth in Google News and Google properties suggests that the 2008 estimate of the value of Google News relative to Google properties can be applied to recent years.
But again, this is a nonsense number. You cannot make this comparison.
As a media analyst, I would call this a null result. It's a number that is so uncertain that I would not consider it usable.
What is the right number? Well, I don't know. None of us do, and that's part of the problem. And I'm as frustrated by this as the rest of you. I would love for Google to just tell us.
But it actually gets worse.
If we look at the last part of the calculation, we can see that they estimate that Google Search share of news is 4.2%, or $4 billion ... which they then say combined with Google News (0.7%) is $4.7 billion.
Wait-a-minute. No no no!!!
This type of calculation would fail any statistics student.
What they are doing here is adding Google News on top of Google Search, despite the fact that Marissa said the value of Google News was the referral value to Google Search.
In other words, it's not two separate things. Google News estimated value is a share of the total revenue from Google Search.
What the News Media Alliance did was this:
What they should have done, of course, was to list Google News as a share of Google Search, like this:
This is a very basic flaw in their calculations.
But you also have to remember that news content on Google Search doesn't drive any revenue either. The report even points this out:
Google makes most of its revenue from its own properties; in 2018, they brought in $96.3 billion of Google's $136.8 billion total annual revenue (70%). While a large component of this is Google Search, most content served on Google Search does not generate ads. According to one study, approximately 3.4% of distinct Google search queries (2.6% of all Google search queries) resulted in a click on a paid result. Yet the content not generating ads still contributes towards Google's mission to organize the world's information and make it universally accessible and useful.
So, all of Google's news usage is basically just a referral to get people to use Google again another day for something else. This still has value for Google, but it makes it really hard to quantify the actual revenue this generates.
The biggest problem I have with the report, however, is how it tries to make it sound like Google isn't paying publishers anything in return for all the money it makes ... which is obviously not true.
When Mathew Ingram pressed them about this, they replied:
When asked for comment, the NMA tells CJR the purpose of the report was to look at how much Google benefits from news, not the opposite.
So they are fully aware that this study was designed to drive a specific narrative by deliberately leaving out what is a very critical element, while at the same time claiming that Google doesn't pay publishers anything. Can you imagine if any other industry had done that?
This is not just lobbyism. This is deception. It's the kind of thing we usually only see from some of the worst companies and political parties when they try to drive a certain result. And for an association dedicated to journalism and newspapers, this is offensive!
But how much do publishers earn from Google?
Well, granted publishers aren't paid directly by Google, but everything Google does is driving traffic and exposure. And this traffic has just as much monetary value for publishers as the News Media Alliance says it has for Google.
So, let me turn this around for a second and apply the same arguments and calculations to that.
We know, according to Parse.ly, that search referrals (of which Google absolutely dominates) represent about 22% of the total traffic to news sites.
So that means that Google is driving 22% of the total digital newspaper revenue. Right?
That is the argument that this report makes. It is saying that Google News is driving direct value to Google Search as a form of referral funnel, so clearly the same must apply when Google also drives that traffic to publishers.
So, just in terms of advertising income for publishers, we know, according to PEW, that digital US newspaper advertising generates $5.1 billion in revenue per year, which means that Google must drive (via the traffic it refers) $1.1 billion of that.
But that's just advertising. Then we also have to include all forms of digital revenue, which for some newspapers is now higher than the advertising income.
This is money that Google 'gives' to publishers (indirectly of course).
Or look at the New York Times, it reported $709 million in digital revenue in 2018, of which Google would then be responsible for generating $156 million in direct referral revenue.
As I pointed out above, there are obvious flaws with this type of calculation, but as publishers, you cannot claim that Google isn't paying you anything. They do ... a lot!
They pay you in the same way that Google News is paying Google.
At this point, of course, many would tell me that none of this is true, because what Google has done is to 'steal' this market from publishers, and now they are only giving you scraps in return. They say that Google does provide publishers with clicks, but that they took those clicks from us first.
The idea here is that, if Google were to go away, all this traffic would somehow go back to the newspapers as a form of direct traffic. And it's because of this, they argue, that Google should now pay for links.
This is an argument I have heard many times, and there is just no data to back that up.
In fact, we have data that illustrates the opposite of that.
The simplest example is to look at Spain, and what happened there.
In Spain, the publishers made many of the same arguments that we now see in the US (remember, this discussion has been going on in Europe for close to 15 years). And they managed to convince their politicians to back them in the hope to force Google to pay up.
What happened instead was that Google looked at the numbers and simply decided it wasn't cost effective for them... and thus closed down Google News entirely.
The impact was quite chaotic, and the effect was a drop in reach for publishers. As they said at the time:
The negative impact on the online press sector is also very clear, since a very important channel to attract readers disappears, resulting in lower revenues from advertising. In addition, the new fee is also a barrier to the expansion of small publications with little-known brands, and an entry barrier for new competitors, since they will be unable to count on these platforms to increase their readers' base.
There are obvious discussions about the actual long-term impact, with some publishers (mostly the big ones) today saying that it didn't make a difference over time, but it clearly didn't help.
In general, from what I have seen (and sorry, I don't have good precise data to point to), my conclusion is that it's a myth that Google is taking more than they give back.
In fact, I estimate it's the opposite, that Google, via the many ways they focus on news, actually give publishers proportionally far more exposure and traffic than they could ever hope to gain organically.
One example of this is the news carousel on top of Google Search.
The News Media Alliance talks about this in a weird way. In their segment where they are trying to explain just how much Google is using news content, they write this.
Google Search incorporates news content in search results, including featured snippets such as "Top Stories" carousels. In 2014, Google launched "In the News" as a new type of search result display, with the goal to make it easier for users to get the latest news and information. In 2016, Google replaced "In the News" with "Top Stories" carousels to make the desktop user interface match the mobile version. This update was both an aesthetic change to Google Search and also reflected Google's efforts to visually distinguish high-quality news content from rest of the search results in response to consumers' need for accurate news stories.
The News Media Alliance is trying to explain how much Google is using news content "without paying for it" ... but there is another more important thing going on here. This news carousel doesn't just help Google, it also massively helps publishers gain more traffic than they could get by themselves. Google is boosting publishers up to a much higher level.
So the reality is that, if Google were not to do this, publishers would likely get far less traffic than they do today. This was what we saw in Spain. Your 'organic reach' without Google is nowhere near what you get from being featured by Google.
Ask any brand if they would like to have their products featured in this way. They would fall over themselves to get it.
So, from my perspective, there is a very fundamental flaw in the way publishers think about this. You think Google has taken more than they give back, but when I look at the patterns, I don't see that.
This leads to my final point.
One of the fundamental things people need to understand about Google is that they are not in the same market as publishers. It's a separate market with very different market mechanics.
We see this in many ways. For instance, with advertising, we see how Google is actually monetized by what I call non-media exposure.
The reason brands are not spending so much money on Google is not because of links to the media, it's because of non-media. I have talked about this in relation to brand safety.
Publishers keep thinking that Google isn't brand safe and that brands should advertise in newspapers instead, but this level of thinking makes no sense from a brand perspective.
As I tweeted a while back:
Once again, the news media industry still doesn't understand what 'brand safe' means.
Brand safe = A person on Google searching for 'mountain bikes', which you can then show an ad for.
Not brand safe = A person reading an article in a newspaper about hate speech, with an ad for hair gel.
The media today is defined by moments. And the best, the most relevant, and the most brand safe place to advertise is on platforms where people have a moment that aligns with the intent to buy.
So, if I go to Google and search for 'mountain bike', the entire page is filled with ads, because brands know that this is the best place to reach people. It's the perfect moment.
So, Google didn't steal the advertising market or the exposure from newspapers, because they are not doing anything like what you do as a publisher.
They have created an entirely different market, a non-media market, that matches people's moment with the specific intent that brands are looking for.
This has nothing to do with news or publishers. In fact, if Google were to stop including news on Google Search, I'm not sure it would have that big an effect on their revenue. Why? Just look at Amazon.
Over the recent years, Amazon has moved into the advertising market with a tremendous level of force and growth. They are not doing this by taking articles from newspapers, they are doing it because they have created a new non-media market that is more relevant for brands to advertise on.
This is the simple reality of the trends right now. The patterns and the changes we see in the advertising market have nothing to do with news or publishers. It's not news articles that are driving the change.
But even if we forget advertising and just look at how people get news, Google is still not doing what publishers do.
One thing I'm really annoyed about with this whole Google vs Newspapers debate is that we in the media do not in any way provide the service that Google provides.
Let me give you a simple example:
Right now, I'm somewhat interested in news about Cyberpunk 2077, so if I go to Google and search for it... and I get a bazillion results (some of them mentioning news reports).
This is nice!
I can do the same if I specifically go to Google News. This too will give me a link to all the media sites that have written about this. It's not as nice as the main Google page, but still quite useful.
But if I instead go to the New York Times and search for the same (sorted by relevance), I get stories about ski jumps, tax reforms and court days ... from an article from 1962 ... I mean, what??
Some of you will point out that this is because Cyberpunk 2077 is a computer game and not a real news story, and therefore NYT didn't cover it.... but this is my whole point!
With Google, I can have any interest, and I simply ask it to show me where to find this specific type of content, and they will.
Whereas for newspapers, I apparently first have to guess whether this topic happened to be covered by this one newspaper before I search for it.
And even if you focus on stories that you know a newspaper has covered, it is still a horrible search experience. Usually, it will find the story if you search for it using the exact type of phrasing the journalist used, but if you change it just a tiny bit, it won't.
I'm sorry, how is this useful? Google has not 'stolen' your market. It has provided a feature that we in the media industry completely failed to provide .... even to this day, we are still not giving people what they very clearly need.
But then, you say, Google should pay for linking to us! ... Why?
Here is a question: Should a wedding planner be forced to pay a license to the flower shops for recommending people who want to get married, go to a specific place and pick up a particular bouquet?
No, of course not. Yes, the wedding planner makes money from 'curating' what is available and people pay them for that service. But they are doing this because, as a flower shop, you have failed to provide this service yourself. Why should they pay the flower shop a fee for that?
It's the same with news. I cannot come up with a single reason why Google should pay anyone for helping people figure out where to find the best content. They are specifically solving a problem that we, as newspapers, have failed to provide.
Mind you, I'm not saying that Google is perfect. In fact, there are quite a number of things that I think Google is doing really badly. For instance, I'm actually blocking Google from indexing my Plus articles here on this site, because the way they implemented this would mean giving my readers access to my content in a way that would be bad for my business.
In the same way, I have decided to not implement AMP, because while I like the concept, the actual implementation is terrible for publishers. It disconnects and limits the content in ways that just isn't helpful.
Similarly, I have a lot of problems with Google mobile apps, where a select few are gaining exposure at the expense of everyone else. It's the old 'the rich is getting richer' problem. In fact, Google, Facebook, and Apple all have this problem.
For instance, Apple news is great if you are one of the few featured by Apple and your content fits their focus and format. But for any other publisher, especially future publishers who want to change the way people think about news, it's absolutely terrible and limiting.
I also have a very big problem with the way the whole ad tech market works.
As Aram Zucker-Scharff, from the Washington Post, tweeted:
In order to empower this process Google has pushed for an infrastructure to encourage this behavior and not only does that allow other players to take money that would normally go to publishers. It allows ~70% of ad spend to disappear into ad tech.
This is not acceptable in any way!
I also have a problem with Google's "subscription for publishers"-thing. Again, it's not a bad idea, but it's implemented in ways that causes so many other problems for publishers.
So, I have a number of big concerns about the tech platforms, and I actually do support the notion that US publishers should get some kind of legal antitrust protection so that they are better able to organize and discuss these things.
But this is not what the News Media Alliance is doing with their report. They are just going with the "Google has taken our content and our market, so they should pay for linking to us" approach... and that's just not helpful in any way.
In fact, even if they manage to convince the politicians, I think this would cause far more harm to publishers, because you are basically trying to force Google into your old market.
It's like if Kodak said that Apple should pay them a license when people take photos with their phones, because Apple has 'stolen that market'.
It's just a completely misguided way to think about what has happened. Google is not in the same market as the New York Times.
We need a very different discussion here.
Founder, media analyst, author, and publisher. Follow on Twitter
"Thomas Baekdal is one of Scandinavia's most sought-after experts in the digitization of media companies. He has made himself known for his analysis of how digitization has changed the way we consume media."
Swedish business magazine, Resumé