Google: The Internet Behemoth and how it profits from YouTube
There is an ongoing debate about the size of the losses at YouTube and for how much longer the parent, Google, can afford to fund its errant child's excessive lifestyle. Credit Suisse put a high price on it; Brough Turner criticised their analysis; RampRate decisively debunked it.
The debate has focused upon YouTube as a standalone service and little attention has been given to the spin-off benefits accruing to the parent. Google controls a significant, and growing, share of the means of production of the entire Internet industry. We argue that ownership of YouTube is a crucial ingredient for Google's control of the economic rent that Google extracts from the whole of the Internet value chain.
We believe that YouTube is used indirectly to drive profits at the parent, and that Google is currently incentivized to keep these profits hidden from prying eyes. The key indirect benefits accruing to Google of owning YouTube are as follows:
i) YouTube gains Google a critical slice of growing online video eyeballs, which will attract more marketing dollars to the Internet as a whole. This is much more important in the USA, where the main competitor Hulu is ad-funded than the UK, where the BBC iPlayer is taxpayer funded;
ii) YouTube gains Google yet more important meta-data which can be cross-pollinated with data from other Google services;
iii) YouTube traffic strengthens Google specifically in peering negotiations and generally in network design;
iv) YouTube is probably a small fraction of Google's overall cost base, and the spin-off benefits from lower overall unit costs;and
v) YouTube positions Google very powerfully for a key role as a gatekeeper in the copyright world.
This article explains these indirect benefits in detail and explains a strategy for telcos to adopt in the online video world.
A rising tide raises all boats
On a macro level, the more 'eyeballs' and time spent on the internet, the bigger percentage of advertising budgets that advertisers will allocate to internet marketing. Compelling content is an essential attractor of 'eyeballs'. For Google with its massive share of the internet market, it doesn't matter quite as much whether the video stream itself is monetized today as long as it picks up a share of the increase in the overall digital advertising budget.
What is more uncertain is whether online viewing would have grown regardless of Google owning YouTube. Nonetheless, Google's ownership ensures YouTube's survival and gives Google a significant foothold in this growing and potentially important area.
Source: OFCOM - UK Communications Market Review 2008.
In 2008, overall UK advertising revenue was £17.5bn with TV advertising £3.8bn and Internet advertising representing £3.3bn. The overall market declined by 3.5%, whereas Internet advertising grew by 17.5% - not bad in a recession.
We are not saying that video accounts for the whole of the increase in Internet advertising. Far from it. A significant part of the rise will be accounted by other factors: the overall increase in the Internet population, the time spent on social networking and other services, the increase in e-commerce spending, and the relative effectiveness of Internet advertising as a channel. However, video viewing is becoming significant, as the latest US Comscore Video Matrix data for April illustrates:
- 78.6 percent of the total U.S. Internet audience viewed online video.
- The average online video viewer watched 385 minutes of video, or 6.4 hours.
- 107.1 million viewers watched 6.8 billion videos on YouTube.com (63.5 videos per viewer).
- 49 million viewers watched 387 million videos on MySpace.com (7.9 videos per viewer).
- Hulu accounted for 2.4 percent of videos viewed, but 4.2 percent of all minutes spent watching online video.
- The duration of the average online video was 3.5 minutes.
What is more uncertain is whether online viewing would have grown regardless of Google owning YouTube.
Google's Share of the Advertising Pie keeps on Rising
As well as Internet advertising growing faster than overall advertising, Google's share of that advertising is growing as well.
Source: IAB/PWC Data and Google SEC filings
In 2008, the Google share of the overall US Internet advertising market had reached 45%, up from 36% in 2006 - pre YouTube acquisition. Again, all of this growth cannot be attributed to YouTube - either directly or indirectly.
In terms of direct YouTube advertising revenues, Google does not publish a figure. Credit Suisse in their infamous analyst note estimates YouTube 2008 advertising revenue at only US$200m (associated with overall losses of US$470m) and the IAB/PWC data estimates video advertising at only 3% of the overall Internet advertising market.
Show me the Meta-Data
One of the little discussed benefits of Google owning YouTube is that YouTube is more than a pure video streaming network. It is a social network in its own right. Accounts are needed to create, comment about and rate content, which is something that is not required for the core search service. It is hard to envisage that at this stage YouTube is directly monetizing this meta-data. Undoubtedly, however, it will be contributing to understanding of user-behaviour, and probably the data is already contributing to the optimization of the Google advertising engine.
Google worldwide revenues have grown by US$11.2bn to US$21.8bn in the two years since purchasing YouTube in 2006 for an all-share cost of US$1.8bn. In the same period, operating cashflow has grown by US$4.3bn to US$7.9bn.
The amount of indirect contribution of YouTube to Google revenues is uncertain, but what is more certain is that Google can afford YouTube.
Marginal Cost = Zero
One of the most important disciplines of Internet economics is to keep the marginal costs of delivering another page or another video stream as close as possible to zero. This doesn't mean that total costs are zero, but more costs are predictable and not subject to the vagaries of any explosive growth in demand.
To be successful on the Internet, then, costs must be either turned into fixed costs or else made success-based. With YouTube,the key variables are computing and bandwidth costs. Google's tactic is turning the majority of them into capital costs. Google is deploying more and more capital building out infrastructure, whether vast datacenters or its own network - effectively keeping an ever increasing proportion of its traffic on the Googlenet.
Another key variable costs for YouTube is the cost of the content itself. Google here attempts to minimize its risk by making any payout to third parties dependent upon success of advertising revenues achieved. This is the cornerstone of the Adsense program for publishers. Unfortunately for Google, it is not the traditional way that the video content industry works and therefore is a great source of friction between YouTube and video rights holders.
The Bandwidth Equation
A case can easily be made that Google could make its cost of delivery for video - zero. Every global IP transit provider would love to be the exclusive deliverer of such a significant portion of the world's Internet traffic, and the transit providers could make money by squeezing the downstream ISPs in their cost of delivery. Such an extreme network design would bear a heavy political cost for Google and would obviously be unpalatable, but it illustrates the power that Google has accumulated through the YouTube traffic.
Instead of focusing on IP transit, Google extensively uses peering, delivering its own traffic to the major peering points in the world, as we made clear in this post, which the people at RampRate used in their critique of Credit Suisse. Peering is not free - it involves buying expensive dark fibre linking the Google data centres to the peering exchanges, renting space in the peering exchanges, equipment to light up the fibre and a team of network engineers to manage the peering relationships. However, most of these costs are fixed. As important for Google is that peering will allow them to control the reliability of delivering their traffic.
A more recent tactic for Google is to build edge-caching servers within ISP networks bringing the content closer to the end-user and thereby improving the speed of delivery. The early indication seem to be that Google is building its own content delivery network in much the same way as Akamai has built one - except that the Google one is for internal use only, right now. Most media companies have to use third party content delivery networks, as they don't enjoy the Google economies of scale. For instance, the BBC uses Akamai for delivery of a significant proportion of its traffic. The Google costs are fixed whereas the BBC's are variable - they are paying a third party supplier.
The more difficult question to answer is how much is the sheer volume of YouTube traffic is helping the other Google services, especially the highly profitable search business. We would argue significantly - in terms of cost, reliability and speed. Furthermore, the overall bandwidth economics that Google enjoy are extremely difficult for competitors to replicate and represent a significant barrier to entry.
The Google strategy in distribution seems to be to keep control and only outsource the minimum. Logically, Google would only adopt this strategy if it felt it could gain a competitive edge through distribution. Scale matters in distribution and YouTube brings that scale to Google.
The Compute Equation
In a recent paper about its datacentre operations, Google argued:
"As computation continues to move into the cloud, the computing platform of interest no longer resembles a pizza box or a refrigerator, but a warehouse full of computers. These new large datacenters are quite different from traditional hosting facilities of earlier times and cannot be viewed simply as a collection of co-located servers. Large portions of the hardware and software resources in these facilities must work in concert to efficiently deliver good levels of Internet service performance, something that can only be achieved by a holistic approach to their design and deployment. In other words, we must treat the datacenter itself as one massive warehouse-scale computer (WSC)."
In determining the computing cost for the YouTube service, it is impossible to look at the cost of the service in isolation or approximate it with traditional server and storage costs. The Google strategy is to build vast datacentres and to control the elements within as much as possible. Google does not use standard off-the-shelf components - Google not only uses commodity hardware and designs the whole of the software stack, but innovates on power and cooling systems.
With the capital costs of a large datacentre running as high as US$250m, the cost is more dependent upon overall utilization of the datacentre than individual units such as storage or CPU costs. In fact, as the chart below shows, power costs are as important as server costs.
Source: The Datacenter as a Computer: An Introduction to the Design of Warehouse-Scale Machines - Luiz André Barroso and Urs Hölzle - Morganclaypool
YouTube undoubtedly consumes a lot of compute resources. Most analysts have focused upon storage, but the cost of ingesting and encoding all the video is very expensive as well. But, we are doubtful YouTube represents as high a proportion of overall Google services' demand for computing as it does for bandwidth. However, we are sure that Google has a significant cost advantage over its competitors.
Google builds its own factories for the digital age. Whilst other individual datacentres are being built of a similar scale, it is doubtful that any other company has reached the same scale as Google.
YouTube and Negotiating Power
Video on the internet is currently extremely difficult to monetize - as we say in the Online Video market study, we are currently in a "Pirate world" where most content is available is available for "free" from a variety of sources. Old media are seeing their revenues cannibalized as content moves online into the "Pirate World". Eventually, either through legislation or sheer volume of eyeballs, New Players will emerge whose revenues and profits will be significant. Arguably, the effort to increase the percentage of YouTube videos that carry ads is a key indicator of this. The business model for this New World is still extremely uncertain. However, just because of the sheer volume of traffic, Google will have a key seat at the value chain negotiating table. Google will also almost certainly be the lowest cost player, in both aggregation and distribution.
What would make the negotiating position of Google even stronger is control of the method of licensing content. The recent Google Books deal shows that Google is starting to get extremely interested not only in the meta-data around copyright, but also in distributing payments directly to content creators and not through traditional aggregators. In this case, the creators are the authors and the aggregators are the publishers. It does not take a major leap in thinking to see how Google could disintermediate the traditional aggregators of video and music with YouTube.
YouTube Overall Profitability
It is irrelevant to determine the overall profitability of YouTube without considering the spin-off benefits to the parent, Google. These spin-offs are currently considerable but difficult to quantify: in attracting eyeballs to the Internet, therefore attracting more overall advertising spend; and in providing the traffic volume which improves the overall economics for the Google Cloud Computing Platform.
In the future, YouTube will be the key for Google establishing a strong position in the licensing of media content and thus not only controlling its own costs, but being a critical aggregator and distributor for content owners in its own right. Google will try to turn its current Achilles heel, copyright, into a future strength.
In the Telcos Eye
The critical message for telcos is in the scale of investment that Google is making in distribution. Despite BT claiming that the BBC and YouTube are enjoying a free ride, the exact opposite is true. Google is building out both compute and bandwidth infrastructure for delivery. Other video services, for example the BBC, are paying third parties such as Akamai for distribution.
The real battleground is around the share of the value chain. BT is really saying that it is not earning enough from its access fees and the economic table is tilted in the favour of "free-at-the-point-of consumption" video services such as YouTube and other sites, such as Hulu and the iPlayer.
But all is not lost. The growing size of the payTV market proves that advertising alone cannot fund all video services and consumers will pay for premium content. The core strategy for telcos competing is not to replicate YouTube, but by providing tools for the myriad of content owners who are unhappy with the current payback from the online world - these tools are not limited to billing & payment services, but should also include copyright protection.
The war of online value chain is not yet lost: we are still in a pirate world and Google is one of a very small club who can afford to distribute all types of content across the globe.