Sorry, we could not find the combination you entered »
Please enter your email and we will send you an email where you can pick a new password.
Reset password:


By Thomas Baekdal - September 2017

Data that looks like it means something, but doesn't

Over the past many months, I have seen tons of articles in the media about how Facebook and Google aren't entirely accurate when it comes to selling audiences to advertisers. Articles like how Facebook claims to reach more than the total population of the US, how Google is now refunding advertisers because of fake traffic, and many other articles like them (which is not unique to them).

As a media analyst, I obviously agree that these things aren't acceptable but from a technical perspective, I can understand why they happen. For instance, the reason Facebook has more US accounts than there are US people isn't entirely surprising. And the same with Google, who had allowed sites to be monetized by advertising without first verifying they were providing value to brands.

This is obviously a problem and it needs to be fixed.

But, every time I read an article about these things, I'm constantly reminded that we, in the media industry, aren't really doing a good job either. So when journalists complain about the metrics on Google and Facebook, it's kind of a fake outrage, because our own metrics are often even worse.

So, in this article, I'm going to do two things.

First, I'm going to illustrate a number of very common problems with advertising metrics that I see in the media industry every day. Metrics that, in many ways, are worse than those we see on the tech platforms.

My second point is that, when I look at the trends, I see that media is losing the advertising market. And the main reason this is happening is because what we tell advertisers isn't really that good. I have some suggestions for improvements!

My hope is that this article will make you think, and encourage you to change the way you, as publishers, sell your advertising. Because if you want to beat Google and Facebook, you either have to offer brands something that the tech platforms can't do, or give advertisers better metrics so that they have a better understanding of exactly what value you bring.

Today, publishers are terrible at this, so let's change that!

Common metrics that are completely bogus

Let's look at five very common problems with metrics that we see in the media industry every day.

Giving them nothing

The first metric is kind of a weird one, because it's the 'no metric'.

One thing I often see with publishers, is that many still believe that they somehow control the market, and because of that think they don't have to prove their worth.

One example of this is what we see with the New York Times. If you go to their site, scroll down to the bottom of the page and click on the tiny 'advertise' link, the page that you end up on looks like this:

You will notice this page has no metrics of any kind. Instead, there is just this tiny text explaining that they have won more Pulitzer prizes than anyone else.

For an insightful view of the world, there's no paper like The New York Times. For more than 150 years, Times readers have expected their newspaper to provide the most thorough and uncompromising coverage in the world. The Times has won more Pulitzer prizes than any other news organization and remains No. 1 in overall reach of U.S. opinion leaders.

While this is certainly an admirable achievement, telling people this has very little to do with advertisers.

And the rest is just practical information, such as what the ad specs are, where to submit the ad files, and what the editorial calendar looks like.

I see this often in the media industry. Many publishers still think they live in a world where they don't have to put in any effort to make money, at least in the initial contact (obviously NYT has a sales team).

Compare this to Google or Facebook.

If you go to Google, you can use their ad planner to design your ad, and they will use their algorithms to tell you how much money you should spend to achieve the best impact, and also tell you what they think this impact will be in terms of potential reach and clicks.

Google's numbers might not be completely accurate, which every advertiser knows, but at least they are giving you something, whereas if I go to the NYT I have no idea.

This is why Google and Facebook are winning. Even with all their flaws, they still offer advertisers a platform that provides better overall insights than what most publishers offer.

The biggest number syndrome

The second problem is what I will call the 'biggest number syndrome', and it's just as bad.

You see this if you go over to newspapers like the Guardian.

If you visit their advertising page, you are presented with this metric at the very top of the page:

First of all, 'browsers' is a completely meaningless metric because it tells you nothing about how your ads are actually going to perform. If are a company selling beds, how many people is 152 million unique browsers?

You don't know. This is a terrible metric.

But what's really weird about this is that the Guardian knows that it's a useless metric. When you scroll down their advertising page, you are presented with a really good video (which unfortunately I cannot embed here).

In this video, they talk about how they are able to deliver actionable results to brands. They say that this is because they are able to combine technology with 'unique audience data'.

The Guardian also says it's not the volume that matters, but the impact. As they say: "The world doesn't need more ad inventory; it needs better inventory"... which they are completely right about.

And they have even framed this concept around a simple principle called "Fewer. Better".

And then they go on to say that they can prove it. By, among other things, telling people that for every pound people spend, they see a return of £3.84.

Every single thing about this video is just awesome. The Guardian is giving advertisers a very clear and very powerful reason to advertise with them. And this is incredibly important if you want to stand out and compete with much bigger platforms like Google and Facebook.

So... let me ask you this.

If you say that you have "unique audience data" that allows you to generate £3.84/pound in ROI, through a more premium ad experience targeting a "fewer but better" audience... why the heck is this on the top of your page?

Why are you doing this, Guardian?

You just told me that you are awesome because you have something that Google and Facebook don't... and yet, your main selling point is this completely useless metric of 152 million browsers... according to Google.

What the heck?

It's things like this that, as a media analyst, just make me want to scream. And I see this not just with the Guardian, but with so many publishers.

You still think you can compete with Google and Facebook by using the biggest number that you can find... even when you know that this number doesn't give you a competitive edge.

If a brand merely wants volume, reaching 152 million browsers on the Guardian is nothing compared to the billions of browsers you can reach with Google and Facebook... and especially not if you use a metric like 'browsers' which means nothing.

Please, publishers, stop doing this. The only thing this accomplishes is to make you look smaller and less valuable than Google and Facebook.

How many readers?

Another, very similar problem, is how publishers define a reader in ways that simply aren't very good, neither from a data or sales perspective.

There are many variations of this.

In the simplest form we have the AAM statements (the Alliance for Audited Media). The idea of an AAM statement is good, because it provides you with a ton of data about the circulation numbers for a magazine.

Like this:

From the perspective of a media analyst, I love this, because it gives me very detailed data, which I can then compare with every other publisher.

But from a brand perspective, this is meaningless... because how many people is this? What's the difference between paid, verified, and single-copy... and why should I ever have to care about that?

Obviously, you can look up the definitions for each, but the problem is that we are not speaking the language of brands. Instead, we are focusing on internal metrics.

These metrics are very important for publishers because they tell us something about our businesses, but it doesn't help brands make smarter decisions.

What brands are looking for is the metric of "what impact can I have?", and this is something that both Google and Facebook are very good at explaining. Again, if I go to Google and define a target audience, Google will suggest keywords, what budget to set, and tell me how many people within this segment I'm likely to reach, and how many of them are likely to click on my ad.

This is a far better metric.

We see the same thing when we look at more detailed audience breakdowns. For instance, many publishers have something like this:

So, imagine that you are a fashion company aiming to influence women in their 20s, how many is that?

Well, we have the 18-24, but... hmmm... maybe, 18-34? No... wait... 21-29 is the right age group but that is for both men and women... so maybe "women & age 18-34"... and then we just subtract a bit?

You see the problem?

You are giving me all this data, but in a way that means I have to do all types of mental calculations in order to understand what I get.

It gets even worse if you match multiple things.

Many publishers have a 'demographic card', or 'audience profile card' that they show to advertisers, giving brands a kind of big picture of what their audience is like. If you try to match multiple things and look at the audience profile card, you end up even more confused than when you started.

For instance, if you want to reach women (5,269,000), age 25-34 (1,518,000), who have kids between 6-11 (1,746,000), a household income of $75,000+ (3,902,000), in a professional/managerial role (2,473,000)... which number do you look at?

This is useless!

I can understand why we did this back in the old days before people had computers and smartphones, but today, all this does is to make you appear as a less competent ad partner than Google and Facebook.

This makes you, as a publisher, look completely out of date.

On top of this is the deception of readers, which we see every time a publisher informs brands how many 'readers per copy' they have.

First of all, we all know that a print magazine may be seen by more than just one person, but the way we measure this is absolutely absurd.

For instance, I have seen this measured simply by having a publisher multiply their circulation with the average household size... so if there are four people living in the household that subscribes to the magazine, you suddenly claim that four people are reading it.

In the above case, you will notice that this magazine claims each copy is seen by an average of 5.67 people, which is far more than the average household size.

The likely reason for this is that it has simply added every potential reader in public places (like the people in doctor's waiting rooms), as if everyone there would read just this one magazine.

This is not a real number. It looks at the maximum potential audience, rather than the real audience, and it doesn't account for any overlap, and so forth.

This is as bad as when Facebook says it can reach more than 100% of the population... and every brand knows this.

15 years ago I was working at a big fashion company, and back then we didn't believe this metric either. I have had magazine ad sales people tell me that each of their magazines would be read by 7 people on average.

7 people per magazine? How can I possibly believe that?... "Oh" they say, "We did a study. We can send it to you".

Heh... no!

Again, we come back to the same problem as before. Facebook and Google might not have 100% accurate numbers, but the media industry isn't really demonstrating that they can do better.

How many kids?

We also have these kind of metrics that don't mean what we think they mean.

Look at this audience profile card:

The metric I want you to look at is the one in the middle. What do you think this number means?

Or rather, imagine that you are a company selling child safety seats, would this magazine be a good one to advertise in because of the number of children the readers have?...And how many children is that?

Well, at first, this seems like a pretty straight forward metric. If this magazine has 1.1 million readers, with 2.1 children each on average, this would mean that there is a potential market for 2.4 million kids.

But this doesn't sound right. This would mean that there are more kids than adults, and unless something crazy has happened to the demographic of the western world, this can't be.

But maybe we need to look at households instead. So, if every household has 2.1 children, that gives us 464,000 potential kids... but that still doesn't sound right.

The problem here is that this metric isn't actually about children at all, because all they did was to ask people: "how many children do you have?"

For instance, if you ask someone in their 20s, they might say that they have one child.

But if you then ask someone in their 70s, they might tell you that they have two kids.

For instance, here is a mother with a son and a daughter. To her, she answered the question correctly, because they are both her 'children'... but they are also no longer children.

This is a common problem that I see all the time, and especially with the way the media is using data. Often, the questions we ask in our demographic surveys don't match their intended use.

In this case, the survey found that their readers had 2.1 children on average, but it didn't take into account whether these children were actually children.

As a result, we end up with a highly misleading metric. Because, if you are a company selling car safety seats for kids, you are led to believe that advertising in this magazine would reach a huge number of potential parents. But in reality, this magazine might not be a good place for that at all.

Again, we see why Google and Facebook are winning. On Google, you can set up a search ad so that whenever people are specifically searching for 'car safety seat', your ad will show up on the top of the page.

For a brand, this is a far more useful form of advertising, and it's much easier to understand than what we see in the media.

What index/rank?

Another big problem in the media industry is how many publishers have turned to 'indexes' and 'ranks' as a way to convince brands that they are more valuable than their competitors.

I'm going to give you one example from Marie Claire.

If you go to Marie Claire's media kit, you will find their AAM statement (like with every other publisher), but you will also find a description of their readers like this:

What they have done here is compared themselves to other magazines, like Allure, Cosmo, Elle, Glamour, Harper's Bazaar, InStyle and Vogue... and then created an index for it.

For instance, it claims to have the largest concentration of millennial readers (51%), which it defines as people aged 20-39.

Okay... but then we look at Elle, and we find that they make exactly the same claim. On Elle's advertising page, they claim to reach "the youngest fashion audience", and that they have a 'millennial' Rank #1 (index: 147) vs. their 'fashion set', which is defined as: Harper's Bazaar, Marie Claire, InStyle, Vogue, Popsugar, W.

Wait.. what? They are both number one?

What's happening is that there is very little difference between them, and they are not defining what a millennial is the same way. Elle defines it as 18-34, while Marie Claire defines it as 20-39.

They are also not comparing each other to the same competitors, so by cherry picking the data, every magazine is able to tell brands that they are number one.

If we turn back to Marie Claire, they also present us with this:

Every single thing on this page is questionable. First, let's look at the audience of 18 million.

In the small print you can see that this is defined by combining their print and digital readers... while also adding all their followers on social channels.

This, of course, is not a valid metric, because a lot of these people are going to be overlapping. The people who follow Marie Claire on Instagram are obviously also reading their website. And since Marie Claire says that 51% of their audience is millennials, one must also assume that a lot of their print readers use both their website and follow them on social as well.

Again, this is exactly the same problem as when Facebook says it can reach more than 100%. It's an inflated metric.

There is also the problem with the 'influencing the influencer' statement, which they have derived from this survey question: "My friends/family often ask for and trust my advice on beauty or fashion or shopping."

This is not the correct definition of an influencer, because pretty much everyone will answer yes to this question.

The definition of an influencer is whether someone is influencing people outside their private spheres. In other words, are they influencing more than just their friends and family?

And, you might also need to define a threshold to this, so that someone influencing 5 random people isn't counted the same way as an Instagrammer with a million followers.

You see the problem here?

Marie Claire isn't worse than the others, and the problems I highlight here are very common for the media industry as a whole. We are absolutely terrible at presenting useful data to advertisers. We are constantly guilty of providing advertisers with audience data that is heavily inflated, and most of our data are flawed, misleading, or outright wrong.

As I mentioned earlier, I used to work for a big fashion company, and I clearly remember how annoyed I was with this... and I still don't see any improvement today.

The problem now is even worse, because without distinction, there is always someone else who will do it even better.

For instance, Marie Claire says they reach 18 million... but Facebook and Google reach billions. It says that they have the highest concentration of educated readers... except, of course, Google and Facebook who both have more. They also say they are number one in travel, outranking every other competitor... well, you know, except Facebook and Google.

The reason the old media industry isn't winning over Google and Facebook is because our data isn't as good. It's much harder (and often impossible) for advertisers to make sense of; you don't provide the same level of targeting; and we are still trying to win by competing on scale.

This simply doesn't work anymore, and you only need to look at the trends to see it.

Even though Google and Facebook are far from perfect; even when we point out discrepancies with their data; even when we go to the extreme and point out that someone posted something bad on YouTube; or that there is tons of fake news on Facebook... none of these things make any real difference.

For instance, YouTube recently announced two new features for advertisers. One is that you can now upload video elements instead of the final video, and YouTube will then automatically mix that to create personalized ads.

Custom audiences are most valuable when paired with creative that is relevant to them. But personalization at scale can be difficult-new video creative is pricey and takes time to make.

We're launching Director Mix to simplify the process of creating different versions of the same creative tailored for each audience-you give us the building blocks of your video ad, like different voiceovers, background and copy, and our system will create thousands of versions to match your various audience segments.

Campbell's Soup used Director Mix to create videos with clever copy based on the content people were about to watch. For instance if you clicked to watch clips from Orange is the New Black, you'd see a bumper asking 'does your cooking make prison food seem good? We've got a soup for that.' And it worked: Campbell's earned a 55% lift in sales and a 24% lift in ad recall with this campaign.

The second feature is that they are making it much easier for brands to create advertising momentum (which is critical to brand impact).

Similarly, we're introducing Video Ad Sequencing to help you architect an ad experience that unfolds over time. This new feature in AdWords Labs lets you string together ad creative. You can pivot, you can react-and you can take consumers down a different path depending on which ads are working for them.

For instance, you could start with a fifteen-second TrueView ad to build awareness, continue with another, longer spot that communicates product attributes, then follow with a six-second bumper ad to keep top-of-mind and drive to purchase.

Compare this to what most publishers have to offer, this is lightyears ahead. And do you think brands really care that 0.1% of their ads might be seen on bad videos if they can get a 55% lift in sales?

Google and Facebook keep winning because, overall, they still have a better product.

How other media should compete

Media has changed, and publishers now need to figure out what they can do better than Facebook and Google. The answer to this is not scale but focus and impact.

We also need to clean up our industry, because our use of data is horrendous. We need to invent better tools, so that we can provide brands with clarity about what each specific ad can be expected to achieve rather than the generalized summaries we see today.

But overall, traditional publishers need to start using data in a better way than simply something you can find in a 'media kit' PDF file ... which is something we will explore in a future report.


The Baekdal Plus Newsletter is the best way to be notified about the latest media reports, but it also comes with extra insights.

Get the newsletter

Thomas Baekdal

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


—   analytics   —


Creating a propensity model for publishers


How my focus on analytics has changed as an independent publisher


How can publishers measure trust and other editorial metrics?


A guide to analytics for independent journalists


Why producing less news leads to a boost in subscriptions


GDPR: How publishers can track things without tracking people