Oligopoly in the Media

We have antitrust laws because, usually, oligopoly is bad for the industry and society in general. During the last financial crisis, so many banks were too big to fail. Now, most likely, Boeing will be saved by tax money, even though we have not seen any real innovation from Boeing or Airbus for a very long time.

I’d like to take a look at the media concentration for the two biggest forms of media: TV and Digital, and OTT, which is booming at this moment.

Top 100 Websites in the US by Monthly Traffic
Top 100 TV Networks in the US by Viewership

It’s no surprise that digital media has a much higher concentration toward a very small number of players compared to TV networks. Most of the top websites use Google or Facebook to monetize their web traffic. In terms of the share of U.S. digital advertising revenue by company, Google + Facebook accounts for 60% of the total market ($130Bn in 2019).

Despite recent M&A activities, TV networks remain very fragmented. This is affecting their ability to unite and fight against digital media. OpenAP is a good initiative to bring together all TV networks. However, lots of uncertainty surround it.

OTT is a fairly new market. With all of the new apps coming to the market, I thought that it was very fragmented. I was shocked to learn that it is actually much more consolidated already.

Top 100 OTT Apps in the US by User Time Spent

The top 10 OTT apps account for about 80% of total time spent. This is almost two times higher than that of TV networks.

Share of Top 10 destinations by Media

Why is the OTT landscape not more fragmented? Early on, most of Netflix’s shows were syndicated. As a result, there should not have been a barrier to entry from a content point of view. I guess there are some barriers in terms of technology. However, I don’t think there is a noticeable difference in terms of the user experience among all major OTT apps.

One thing to notice is that most of the top OTT players (Netflix, YouTube, and Amazon) have other cash-flow-positive businesses (e.g., DVD and Search Ads) to support their tech and content development. Other TV networks had a good portfolio of content but treated Netflix like a distribution partner. Their investment in their own OTT platform was too little and too late.

I don’t need to emphasize the many advantages that monopoly companies enjoy. All digital advertising technology companies suffer from the duopoly of digital media. I predict that the media concentration will continue to increase in the next 5-10 years unless there are new regulations. This isn’t going to help a third-party measurement company.

Current antitrust law doesn’t count the global presence of these powerful companies. Many of them also form a strong product portfolio (e.g., Google) or build vertical integration (e.g., AT&T and Comcast). As long as their market share is not very high in one category, they are not going to be prosecuted for violating antitrust laws. 

I hope my prediction is wrong. However, the industry might need a paradigm shift (e.g., AR and VR) to regain its competitiveness.

Should Time Spent Equal Ad Spending by Media?

Queen of the Internet Mary Meeker publishes a Time Spent vs Ad Spending chart every year.

I believe that it’s a very insightful analysis and says a lot about how the industry is moving ahead.

In my opinion, three factors contribute to the gap between Time Spent and Ad Spending.

  1. Ad effectiveness: Clearly, the effectiveness of advertising is different by media. It might be different according to the type of advertiser as well. Digital media clearly took advantage of the recent ad-tech boom and significantly increased ad effectiveness via better measurement, targeting, and attribution solutions.
  2. Media owner concentration: If the number of media owners is small, they tend to control the price. Google and Facebook have much higher market shares in the digital media space than TV networks do in traditional media. I will write about this topic in my next blog post.
  3. Rapid change in time spent in recent years: Typically, ad spending changes after time spent shifts. There is some time lag, as the entire industry has to catch up to the trend. If the change in time spent happens very quickly, it creates a bigger gap between time spent and ad spending.
% of Time Spent vs. % of Ad Spending (2008-2020)

I have carried out the same analysis using eMarketer’s data. I’m not sure why the ad spending number is very different from Mary Meeker’s analysis. Based on my analysis, both TV and digital had very low ad spending back in 2008. In the last 10 years, both of them were able to reach their fair amount of ad spending. Other media outlets (e.g., newspaper, OOH, magazine) are quickly losing their shares.

If we focus just on TV, the trend looks quite interesting. Until 2015, while time spent decreased quite a bit, ad spending increased significantly. My hypothesis is that this was driven mainly by the improvement of the ad effectiveness of TV commercials. However, after 2016, the learning curve started to saturate. Ad spending decreased at the same pace as time spent. Based on this analysis, most likely TV’s ad spending % will continue to shrink at the same pace as the time spent %.

TV Time Spent vs. Ad Spending (2008-2020)

Although digital’s ad spending and time spent are pretty much the same, when we look within digital, we see a very different picture. Mobile’s ad spending % is actually much higher than time spent %. The time spent of other connected devices (e.g., Roku, Chromecast) is much higher than ad spending. Logically, mobile ad spending growth should slow down quite a bit. The ad spending of other connected devices should grow much faster. The time spent of desktop will gradually shrink to the level that matches the ad spending of desktop.

2020 Digital Media Ad spending vs Time Spent

Why is the gap for mobile so big? The peak of mobile time spent growth happened around 2013. After that, the growth rate of mobile slowed down quite a bit, though ad spending has been growing at a very high pace. In addition, mobile ads are controlled by a few tech giants that are capable of artificially increasing the price point.

Mobile Time Spent per Day and Growth Rate

Other connected devices are in the completely opposite situation. The growth of time spent has been accelerating. However, most of the brands are still only doing the pilots. Ad spending is significantly lagging behind time spent. The supply-side market is much more fragmented as well.

Other Connected Devices Time Spent per Day and Growth Rate

Based on this analysis, I predict the growth of mobile ad spent will plateau very soon.  It also gives me reason to believe the golden age of OTT/CTV is coming.

Will Addressble TV really save the TV industry

Since I started TVision, every year I hear from people in the industry telling me, “Addressable TV is going to save the TV industry. It will grow exponentially starting this year.”

eMarketer has been lowering its estimates on US Addressable TV Ad Spending every year since 2015. Clearly, we start every year being too bullish about addressable TV, then get disappointed around Christmas.

Let’s take a look at the value of addressable TV from both the buy and sell points of view.

For brands, the value is clear: It will improve targeting capability. Brands will be able to broadcast ads to only their target audience, which is very attractive. The question really is: How much of a premium should brands pay?

Let’s take a look at the following four brands. They have very different addressable audience percentages. Brand A is a large CPG brand that sells soap. Everyone is a potential customer. Brand D is a luxury auto brand. If we assume that the average linear TV CPM is about $10, Brand D’s target audience based effective CPM will be around $100.

Addressable TV today is sold at a much higher price. Usually, its targeting data is based on household-level data. As we all know, a TV set is a passive multi-viewer device. Even if the targeting data is 100% accurate, there is at least a 50% chance that ad exposure is directed at the wrong individuals. If we assume the price of addressable TV CPM is $50, the actual CPM might be $100 or above.

This is a very rough analysis. However, it does show that if a brand has a very board target audience, most likely it doesn’t make sense to leverage addressable TV at all.

I think that even for Brand D, the cost benefit is questionable. First of all, at $100 CPM, Brand D can purchase premium display or video ads via a digital channel. A digital channel provides a much better capability to target each individual and not the entire household. In addition, the potential customer base might be much larger than 10% if the sales cycle is very long. A poor college student may eventually buy a BMW in his 40s. However, it’s impossible to target the high-potential college student in an effective manner.

To sum up, unless it’s a super-niche brand, most likely addressable TV doesn’t make sense, at least at the current CPM level. 

BrandAddressable Audience %Linear CPMEffective CPM (Linear)Addressable TV CPMAddressable TV CPM (Actual)
Brand A70%$10$14.29$50$100
Brand B50%$10$20.00$50$100
Brand C30%$10$33.33$50$100
Brand D10%$10$100.00$50$100

TV networks have been selling the best inventories during Upfront for the last few decades. By selling all the premium inventory at once, networks were able to artificially create scarcity and charge a premium. However, addressable TV inventories are much harder to optimize for the yield, even if it is part of a private marketplace. To offset the potential risk, networks must charge a very high CPM for addressable TV ads. Given that the Upfront market is getting stronger every year, it’s very unlikely that networks will make addressable TV part of their main offering.

Please don’t get me wrong. I do agree that current Nielsen age/gender targeting is outdated. However, this doesn’t mean that digital media like one-on-one targeting is the best answer for the TV industry. Most of the brands use TV to quickly get a massive reach and grow their brand. Data-driven linear might be a happy middle ground. I think that the automation of the TV planning/buying process makes a lot of sense but, most likely, the answer is not addressable TV.

What you can measure isn’t the same thing as what you should measure

I was in London the other day for an event (pre COVID-19, now feels like it’s ages ago). Later, I had lunch with Orlando Wood at System 1, who gave a great presentation at the event. This is the link to his book: https://www.amazon.com/Lemon-advertising-brain-turned-sour/dp/0852941471

If we block the left or right side of a person’s brain, it will completely change the way this person thinks and behaves. The left brain is able to draw only a very symbolic and conceptual flower. Meanwhile, the right brain can draw the outline of the flower, but it lacks details. To be fully functional, we need both sides of the brain.

He also presented his research to show that, in the last 10 years, more creatives in our industry have been driven by the left brain (i.e., flatness and abstraction). As a result, ads are becoming less effective.

Orlando told me that one of the key symptoms of left-brain-driven thinking is that it is 100% data-driven. Everything needs to be explained by data … and data is the only source of truth.

Lots of companies are starting to offer a TV attribution solution. Traditionally, it hasn’t been very easy to understand the ROI of TV campaigns, though now it’s getting lots of traction. I do believe that TV attribution creates lots of value for marketers and that it’s better for the industry to keep measuring it. However, this doesn’t mean that TV attribution is the only thing you should measure.

I have a degree in operations research. It’s tempting to make everything data-driven. However, we are far from able to fully understand the marketing effectiveness in an accurate manner. This means that all data products available in our space provide only part of the answer. There are many important questions to which we have not found good answers, such as the long-term value of TV and other long-form content.

We cannot blindly choose what we can measure as our KPI to track. In my opinion, TV attribution cannot really measure the strength and impact of brand equity. It also doesn’t factor in the long-term effect of TV, especially around brand-building. If TV starts to use CPA as the only KPI to compare against digital media, then TV budgets will quickly shrink to zero. Brand marketers must believe in what they think is really important and continue pushing it forward, even if it cannot be fully backed by data. I know that this is not easy in this day and age. However, the strong conviction and courage to execute are key success factors in transforming the industry.