The Way We Measure Online Ads is Ruining the Internet

The most used metric for the reach of online ads, CPM (Cost Per Thousand Impressions) is completely and utterly broken and sits at the foundation of possibly one of the biggest con jobs in the history of modern economics.

A journey into the world of Digital Marketing Metrics

It’s July 2014 and I’m shown into a board room on the top floor of One Coca-Cola Plaza in downtown Atlanta. I make an awkward joke about the ‘secret formula’ while I prepare to discuss the results of the four College Football Fancams we produced for Coke Zero the previous year.

The digital innovations team is excited. They took a chance on this guy with the weird South African accent and it turned out to be a massive success.

Thousands of fans spent an inordinate amount of time on the Coke Zero branded Fancams, they shared the content through their social networks, entered sweepstakes, watched the videos and did everything you want them to do in a successful digital campaign.  This is what you could call "deep engagement". We were thrilled! 

So, now it’s time to go big and the innovations team has organised a number of meetings with “different stakeholders to discuss our results and see how it could help them achieve their KPI’s”.

As a bit of background: My company, FanCam, captures super-high-res 360 degree images  of cities and stadiums that are so detailed that people can go and find a picture of themselves and tag it on Facebook. This is a screen-capture of a SuperBowl FanCam.

As a bit of background: My company, FanCam, captures super-high-res 360 degree images  of cities and stadiums that are so detailed that people can go and find a picture of themselves and tag it on Facebook. This is a screen-capture of a SuperBowl FanCam.

After showing them the cool pictures, I dive straight into the numbers:

“For our Auburn Fancam we generated 120 000 pageviews, 90% of which were unique users, 68% on their mobile devices and an average time on site of 5 minutes per user. We captured 5 000 email addresses and identified the top 20 influencers at the school. Our...”

“That’s not particularly good,” interrupts one of the smartest guys on the social team.

I’m caught completely off guard and there is palpable discomfort in the brief silence that follows.

“One tweet from @CocaCola delivers more impressions than that and it doesn’t cost us a cent.”

He goes on to explain that they recently ran a Twitter campaign that delivered a billion impressions.

I’m completely stumped and I suddenly feel very, very far away from home. I thought our results were good.  No, I know our results are good, but I have no comeback. A billion is a hell of a big number!

I deflect with a half decent, “I’d love to schedule a separate meeting so that the two of us can dive into the numbers a bit more,” and quickly put our Super Bowl Fancam on the projector to bring balance to the force, but my internal ‘non compute’ signs are flashing like crazy. I needed to discover the value of an impression to find out if I still had a business. 

What I learned was so surprising but yet made complete sense. I'll start with the basics and we'll go from there. 

Your Time Online is not Accounted For

To state the obvious, that free content you’re consuming every day, is paid for by advertisers. You’re not paying to see those holiday pics on Facebook because Facebook is getting paid by advertisers to display ads alongside them. Your news is free for the same reason and so is Pinterest and Twitter and Gmail and Bing and almost everything else. This is how the internet works.

Your attention is worth money and in exchange for it, you get subsidised content. It’s a great deal! 

Where it gets interesting however is how your attention is measured and paid for:

The metric of the day is the CPM (Cost Per Mille) or simply put, “cost per a thousand impressions served”. Google, Twitter, Tumblr, Youtube and almost every other company providing you with free content is getting paid on a CPM basis.  It’s the metric that makes the internet free.

It’s also completely and utterly broken and sits at the foundation of possibly one of the biggest con jobs in the history of modern economics.

The CPM Delusion

In 2014, advertisers bought $50B worth of impressions in the United States alone. That’s enough money to pay the full tuition fees of half the country's college students.

There is a problem however, in late 2014 Google reported that 56% of the ads bought in this way were never displayed in front of human eyes. You need to read that again.: Never displayed in front of human eyes. Not “never seen”. NEVER SHOWN. 56%. 

How is this possible?

Well, the part about that CPM definition you (and many advertisers) missed, is the little word at the end: ‘served’. While it would be reasonable to assume that when you pay $20 CPM, you would reach a thousand people, that is not what ‘impressions served’ means. ‘Impressions served’ means, a server ran a program displaying your ad somewhere - not necessarily in front of a human.

It could be on a part of the page nobody ever visited or it could be on a screen in a dark room somewhere, but the assumption that you reached a 1 000 potential customers with your 20 bucks is very, very wrong.

But wait, it gets worse… The 44% of ‘good display ads’ in the aforementioned Google stat are described as “viewable ads”, but that doesn't mean they were actually viewed.  In fact, the IAB definition of  ‘viewable’ is that only half of that ad needed to be displayed on a screen for one second to count as an impression. That’s right: half of a banner ad appearing at the bottom of your screen for one second is deemed enough for you to have viewed that banner ad, read its message and pay attention to it. All while you’re actually busy paying attention to the content on the site.

So what does Google say about how many ads were displayed in full for 3 to 5 seconds? Well, Google doesn’t say anything. Convenient? If you’re selling banner ads at CPM, sure. So let’s do some napkin math here:

I spent $20 CPM and hoped to have 1 000 potential consumers pay attention to my ad. In my case, I’d like them to see my whole ad and have it in their field of view for 5 seconds, at least. Let’s look at the numbers

I pay for 1000 impressions (this is the 'M' in CPM).  Google says 56% never reached a screen. That leaves us with 44% of 1 000. That's 440 impressions left. Considering that an ad needed half of its pixels displayed for 1 second to get into this ‘good bucket’, let’s be conservative and assume 10% of these ‘viewable’ ads had the whole ad in front of a human for 5 seconds. That gives us 44 impressions. Let’s say you pay attention to 10% of the banner ads actually displayed on your screen for 5 seconds or longer (do you??). We all know it’s less, but we can give CPM the benefit of the doubt here... That leaves us with 4.4 actual meaningful "impressions". Okay, so rounded up that brings us to 5 effective impressions for $20.

That’s $4 per viewed banner ad or in other words $4 000 effective CPM. And here’s the kicker, you're paying $4 000 CPM for something that consumers actively hate (see ad blockers) and will almost always find itself alongside another banner ad, diminishing what value there was even further.

Measuring consumer attention on a CPM basis make absolutely no sense to anyone - except those selling it.

“If this is true,” I hear an imaginary executive in Madison Avenue ask, “why have display ad budgets been increasing year over year and why aren’t they showing any signs of slowing down, smart ass?”

Dear “soon to be unemployed executive,”, two reasons:

One -    Nobody is incentivised to rock the boat:

Brands? Nope. As long as that guy in the board room in Atlanta can tell his boss he delivered a BILLION impressions, he sounds like it’s worth having him around - and as long as Pepsi is reporting to deliver 2 Billion he has no incentive to report in any other way.

Agencies? Hell no! They earn commission on the $50B. Media sellers like Google, Facebook, Twitter? Not now, they’re busy printing CPM money.

Two -    There has been no alternative mainstream metric to measure attention in the digital context.

So what has changed?

The Accountants are coming.

Despite all the hype, big brand digital marketers have had difficulty linking views and impressions to actual sales. Upper funnel awareness and target market lift was used to justify buying banner ads on ESPN’s home page, without having to report on the ROI delivered. In fact, Pepsi and Diet Pepsi lost 5% market share the year they went fully digital in their advertising. (Long read on thisHow the Mad Men Lost the Plot).

CFO’s had to stand back because they weren’t paid to understand the dark art of advertising, they were there to make the numbers work. Problem is, advertising has become a numbers game and the accountants are smarting up. The most asked question in marketing is becoming: what's the ROI of that? 

Against all wisdom, these accountants have been allowed to have Facebook and Twitter accounts and you can bet your bottom audited dollar there’s at least one CFO out there who has looked skeptically at the thousands of impressions delivered by that picture he shared off his calculator. Add their traditional distrust of CMO’s and the fact that Google admits 56% of the banner ads bought were never seen and you can be sure they’ll soon be asking some serious ROI questions in those budget allocation meetings.

Short version: If numbers people start asking serious questions about CPM, marketing people are going to struggle to justify their spend and will need to find more common sense metrics to do so.

There’s another player who is going to drive this change though - and it’s just as unlikely a hero as the accountants. Introducing good old print media. These guys are backed up in a corner, battered and desperate. They hate the impression economy, they just don’t know it by that name... yet!

How is it possible that a listicle on Buzzfeed has more value than a New York Times article? In the impression economy, where all impressions are equal and the time and quality of the attention is not measured, the listicle is worth more than the article.

In the attention economy however, the time on site, intensity of attention and the profile of the target audience have the potential to help the NY Times sell their attention at a higher rate.

Screen Shot 2015-11-18 at 6.29.27 PM.png

The Financial Times (accountants + newspaper) have already made the switch. Since 2014, they’ve stopped selling ads on a CPM basis and have brought in CPH (cost per hour) as a more sensible replacement.

The sales pitch basically goes like this: “Who wants 500 hours of attention of CEO’s in Germany?” A number of brands signed up and the FT has reported that they’re really happy with the initial results. Why? Because it makes sense to smart advertisers.

Which brings us to an alternative metric to measure attention in the digital space.

Let me try and explain with a real world example to illustrate the difference between an impression and actual attention. 

My wife is going away for the weekend. I’m in charge of our 3 boys aged 8, 6 and a terrible 2. Just before she heads out of the door she reminds me that our son Jean needs to be at his cricket match at 9am, Tian’s medication has been measured out, and I should not ‘accidentally’ leave Petri at the park. 

“Are you paying attention?” she asks as I scroll through my Twitter feed on the phone.

In the impression economy I can answer with a resounding, confident and truthful, “Yes”. I have indeed registered her as an impression. I am definitely aware that she is there and that she is saying … something. In the real world, however, I’m about to get my ass kicked.

Humans don’t live in the impression economy, we live in the attention economy and my dearest wife knows this. She also knows (as should all advertisers) that she requires a certain type of attention for the communication to be effective.

“Have you been listening to all I have been saying?”, comes the follow up. In other words: Did you listen from start to finish (time) and did you do so without reading tweets (intensity)? This is how you measure attention in the real world. Not by registering an ‘impression’.

There is a third attention metric: knowing if you’re capturing the attention of the appropriate target audience, but I hope my wife never asks that question.

Marital bliss aside, how do we apply the fundamentals of the attention economy to digital advertising?

Well, we start by measuring attention in a way that makes sense.

Enter Dominic Good, advertising sales director at the Financial Times, explaining the sense behind their recent switch from CPM to CPH.

“While CPM values every impression the same, CPH uses time to measure value. The FT has shown through extensive testing that brand familiarity and recollection among readers increases significantly the longer an ad is in view. Adverts seen for five seconds or more on show up to 50% higher brand recall and familiarity than ads that are visible for a shorter period of time.”

He is echoed by Twitter co-founder and founder at Medium, Ev Williams: “We pay more attention to time spent reading than number of visitors at Medium, because in a world of infinite content—where there are a million shiny attention-grabbing objects a touch away and notifications coming in constantly—it’s meaningful when someone is actually spending time,”

“After all, for a currency to be valuable, it has to be scarce.”

And that is the beauty of the attention economy right there, scarcity.

See, you can copy, paste and serve as many banner ads as you like, but human attention is a limited resource. It’s a function of the number of humans in your target audience and the number of hours they have available to pay attention to your brand. It really is that simple.

The future

Imagine a world where we admit to the reality that attention is a limited, valuable and diminishing resource. This is a world where “attention holding” is more valuable than “attention grabbing”.

Where CMO’s focus on tactics that will allow them to capture as many of those minutes for as little money as possible. Where real world principles are used to judge the effectiveness of digital marketing campaigns.

No more wasteful spending on ineffective, unseen banner ads.

No more slideshow websites to generate more impressions.

Instead, more money for innovation and high quality content and serious questions about what consumers care about (pay attention to).

By the way, we had our follow-up meeting with that smart Coca Cola team and ended up doing another 20 games for Coke Zero (even more successful than the first ones) but those couple of awkward minutes on the 29th floor of One Coca-Cola Plaza, put me on a path to try and understand the world of digital marketing metrics - and it’s been a wild ride.

This is the Moneyball of marketing and everything is about to change - again.