Does Netflix Have a Better, or Worse Business Model than Comcast?


Download article:


This report compares the business model of an Over the top (OTT) internet television service provider with that of a cable TV service provider.

The analysis is focussed on Netflix (OTT TV) and Comcast (cable TV) which are analysed below on a like-for-like basis.

Financial data published by Netflix and Comcast, supplemented by our own analysis of that data show that three costs account for around 80% of total operating expenses:

  • Content acquisition: costs incurred in licensing third-party content or in self-producing content;
  • Content delivery: delivering the content to the subscriber’s home;
  • Marketing: the cost of building and maintaining demand for the service and content.

General & administrative expenses and allocations for depreciation & amortisation then have to be added, but these costs are proportionally similar for both companies, so they can be ignored for this analysis.

We will first look at the network architectures for both Comcast and Netflix:

Comcast: Cable TV Network Architecture

The diagram below summarises Comcast’s network architecture.

We have illustrated Comcast’s ‘vertically-integrated’ model where the company’s services are delivered using its own network, which terminates inside the customer premises, at the network connection socket.

There are three key points:

The first point is that although originally used solely for the delivery of Pay-TV service, Comcast’s hybrid fibre coax (HFC) network today supports the delivery of broadband internet service and voice service. 

Indeed, if we look at the financial performance of Comcast’s ‘Cable Segment’ over the three-year period from 2010 to 2012 then we see that over 70% of the revenue growth has come from selling broadband service and voice service to residential users and business customers. Video accounted for just 18% of revenue growth with advertising and ‘other’ making up the remaining 12%.

This story – one of non-video services accounting for the bulk of revenue growth – is common when we look at mature Pay-TV markets around the globe.

The significance of this is that when we compare Comcast with Netflix, we have to remove the portion of Comcast’s operating costs that are attributable to the delivery of broadband and voice services. What we are then left with is an approximate cost for providing the pay-TV service only.

Because Netflix does not offer broadband or voice services this adjustment will allow the cost comparison between Comcast and Netflix to be made on a like-for-like basis.

The second point relates to the electronics needed at the customer premises:

  • Set-top Box (STB): needed to provide Pay-TV services;
  • Broadband Modem: needed to provide broadband and voice service;

We will first consider the STB and then the broadband modem.

The set top box (STB) allows Comcast’s network to be connected to the user’s television set. We have decided to ignore the STB partly because Comcast does not make any money from the sale of the STB itself, which is viewed as an enabler of sales of PayTV, broadband and voice services and not as a revenue item in its own right (there is another reason for ignoring the cost of the STB which we will explain in the next section.)

We have also ignored the cost of any broadband modem that Comcast will have to provide if the customer chooses to use Comcast’s broadband service.  The reason for this is that we have already removed the portion of network costs that relate to Comcast’s broadband service and so we should also remove the equivalent costs at the customer premises.

The third point relates to how Pay-TV operators, including Comcast, are today offering what we term ‘Internet-enhanced Pay-TV’ services. Features such as allowing customers to use an app on their smartphone to control the TV set remotely, rating shows and  interacting with other viewers during the show all require the user to have an internet connection.

OTT internet television service providers such as Netflix also offer similar features which are delivered using the same internet connection that is used to stream the television content.

But because we have removed the cost of broadband from Comcast’s operating cost structure, we have to assume that the Comcast user is obtaining their broadband internet service from a third-party provider (i.e. not Comcast).

Title:Network Architecture: Cable TV Service Provider (Triple Play Example Shown) e.g. Comcast
Markets:Television, Internet, OTT Internet TV

So, on this basis, Comcast has ‘outsourced’  some of its operating costs to the user – which is an argument that some make about Netflix who require customers to have their own broadband service.

But we have just seen that by providing internet-enabled value added features Comcast is effectively doing the same thing.

As we explain below, we are not including the cost that Netflix’s customers have to pay for their broadband service in Netflix’s cost structure and and so we are not including the equivalent cost for Comcast – to be consistent and keep our comparison on a like-for-like basis.

Netflix: Over the top (OTT) Internet TV Network Architecture

The diagram below shows a simplified version of Netflix’s network architecture.

There are three key points:

The first point is that Netflix only offers internet TV service: the company does not offer voice or broadband service.

This means that all of the company’s network operating costs relate to delivery of the core internet TV service. Therefore, we do have to adjust Netflix’s network operating costs as we have done for Comcast.

The second point relates to the customer’s broadband connection.

In order to provide its service, Netflix requires the customer to have their own broadband connection, which let’s say costs $30 a month.

The key question is whether this cost should be included in Netflix’s model  - either by adding, say, $30 to Netflix’s $7.99 monthly service charge, or by adding the network-related cost of providing that service to Netflix’s operating costs.

We think that that the $30 per month should be ignored.

The reason is that the Netflix customer does not just use their broadband service for Netflix.

Indeed, it is quite likely that the customer decided to pay $30 a month for broadband internet long before they chose to subscribe to  Netflix.

On this basis, the customer would already have valued  the non-Netflix services that are made possible by a broadband connection (e.g. YouTube, email, social networking, picture sharing, online shopping, education, Skype etc.) at $30 per month, and so the incremental price being paid by the user for Netflix is zero.

Of course, this is not quite right because with the explosion in video traffic over the internet, broadband ISPs shave begin introducing usage caps and tiered price plans that are based on that amount of traffic consumed, so that, in fact, Netflix users will be paying to use Netflix.

But, even in this case, Netflix users are unlikely to just use Netflix – what about usage of Hulu? Or, in the UK, BBC iPlayer?

So to get a more accurate handle on the effective cost that the user is paying just to use Netflix we would need to assign a value to all the main internet service categories that the customers uses and then, just for the portion that is internet TV, we would then have to apportion that across the different internet television services the customer uses, based on the hours per month actually viewed.

Title:Network Architecture Over the Top (OTT) Internet TV Service Provider e.g Netflix
Markets:Television, Internet, OTT Internet TV

We think that we would then be left with, not $30 per month, but maybe something like $3 per month.

Be we do not even think that this much smaller cost should be included.

The reason is that, as we explained above, Comcast requires the user to have a broadband connection so the customer can benefit from a range of internet-enabled, value-added features. Our methodology so far has been to remove all broadband-related costs from Comcast’s operating cost structure.

If the Comcast customer was also paying, say, $30 per month for a third-party broadband service then we would have to decide what proportion of that cost should be attributable to the internet-enabled features that Comcast has integrated into its Pay-TV service.

Again, like Netflix, we think that this will be another modest sum in the region of a few $ – and so we are ignoring both.

Thirdly, we should talk about the STB which one might expect to see in the customers premises  – but which we have removed (see diagram below).

This is because we see the future being one where all TV sets sold in developed internet markets will come equipped with a connector that allows the user to connect their TV set directly to the internet. This means that Netflix can provide its service directly to the user without any special set-top box (e.g. Roku, Boxee, Apple TV) being required.

This means that, in the long term, most OTT internet television service users will not require a STB.

For those who cannot still quite let go of the idea that the $30 that a Netflix customer might pay for a broadband connection should be added to Netflix’s monthly service fee – to allow a fairer comparison to me made with Comcast, then we would counter that we should really have to add back the cost of the STB that Comcast requires the user to have – and which the customer has to pay for (even if Comcast does not make any profit from the STB sale).

The only way to put things back on a level playing field would be for Comcast’s STB functionality to be integrated into the customer’s TV set – but that is never going to happen.

The reason for this is that while the electronics and software needed to add a connection to a TV set is standard, every Pay-TV network – of which there are hundreds – is based on a proprietary set of technical specifications which would mean that each TV model made by each TV set maker would maybe require 50 different technical implementations and a range of physical connectors. This simply is not going to happen.

So in the long term OTT services will not require a dedicated STB, but all Pay-TV services will continue to require a dedicated STB.

Because we have ignored the cost of this for both Comcast and Netflix,  we are being somewhat kind to Comcast.

Content Acquisition Costs

Our analysis is that the content acquisition costs for Netflix, Comcast and – to provide a third benchmark – Sky in the UK, are all about the same: 42% of service revenues.

We have estimated Netflix’s content acquisition costs by looking at “Streaming content obligations” which the company began reporting in 2011.

These obligations are the total liabilities that existed at the end of the reporting period. The company reports these liabilities based on the terms of the contracts it has entered into with content owners. When the company prepares its balance sheet at the end of the year it reports the total liability, but provides additional detail which shows when those liabilities will become due over the coming years. Most of the liability falls due within the following two years.

The figure charted below for Netflix for 2012 (equivalent to about USD 1.7 billion), relates to the difference in streaming content obligations reported for 31 December 2011 and 31 December 2012 – which we think is a reasonable proxy for what the company actually paid for streaming licences in 2012.

Title:OTT Internet TV vs. Pay TV
Sub Title:Content Acqusition Costs as a % of Sales
Companies:Comcast, Netflix, Sky
Markets:Television, Internet, OTT Internet TV, Pay-TV, Satellite TV, Cable TV
Years:2007, 2008, 2009, 2010, 2011, 2012

Note that this cost in 2012 will have related to programming that would have been streamed in the previous 5 years (mostly the last two years) and so this does not relate solely to content streamed in 2012.

While this is clearly inaccurate, the reporting policies of Comcast and Sky are similar and so the numbers they report for content acquisition contain equivalent inaccuracies.

Because of similar reporting policies, we think that the content acquisition costs for all three companies can be usefully compared on a like-for-like basis.

Content Delivery Costs

According to the company’s 10-K filing for the 12 months ended 31 December 2012, Netflix utilizes both its own content delivery network ("Open Connect") and third-party content delivery networks, such as Level 3 Communications, in order to “help us efficiently stream content in high volume to our subscribers over the Internet.”

Netflix has developed Open Connect  because the company  believes “it makes economic sense to have our own specialized CDN.” The company says it will continue to work with commercial CDN partners for the next few years, but eventually expects the “vast majority” of its streaming content to be served by Open Connect.

 Open Connect allows partner ISPs to  have content delivered at the desired location in their network (at no cost) or at common Internet exchanges.

The following chart shows that Netflix’s proportionate content delivery costs have been increasingly steadily for the last 5 years, while our estimation for Comcast’s content delivery costs (for video programming only) peaked in 2011 and have begun to show a slight downwards trend.

Title:OTT Internet TV vs. Pay TV
Sub Title:Content Delivery Costs as a % of Sales
Companies:Comcast, Netflix
Markets:Television, Internet, OTT Internet TV, Pay-TV, Cable TV
Years:2007, 2008, 2009, 2010, 2011, 2012
Countries:United States

The reason why Netflix’s costs have increased is because the company has been investing in Open Connect, which has required the hiring of engineering resources and the establishment of contracts with hosting and cloud storage companies, all of which increase fixed costs.

The other reason for the increase has been the company’s former dependence on content deliver network partners (like Level 3 Communications). During this phase of the company’s development overall traffic was increasing at a faster rate than per-GB CDN costs were falling, which is why absolute costs were increasing faster than revenues, hence the upward trend in proportionate content delivery costs as shown below.

Netflix’s proportionate content delivery costs appear to be on a path to exceed those of Comcast in the next few years. But we think that as the percentage of overall traffic delivered through the company’s own network increases then the rate of increase will slow down.

In the Summary section of this report we explain why we think that Netflix will eventually be forced to increase the price of its service, say by introducing a tiered pricing model and when this happens then there will be a further reduction in the rate at which Netflix’s proportional content delivery costs are increasing.

Eventually, these two effects will mean that Netflix and Comcast will have similar content delivery costs, when assessed as a percentage of revenues.

Other Costs

The largest contributor in the ‘other cost’ category is marketing which, as the following figure shows, is  roughly 15% of revenues.

While Netflix is still growing most marketing expenditure is directed towards communicating what the service is and  what the benefits are. In the case of Comcast, most of the marketing expenditure is directed towards promoting new shows and movies to its existing users, which persuades them to stay with Comcast.

Even the satellite TV provider, Sky, based in the UK has a marketing budget that has been running at 18% of sales since 2007.

Title:OTT Internet TV vs. Pay TV
Sub Title:Marketing Costs as a % of Sales
Companies:Comcast, Sky, Netflix
Markets:Television, Internet, OTT Internet TV, Pay-TV, Satellite TV, Cable TV
Years:2007, 2008, 2009, 2010, 2011, 2012
Countries:United States, United Kingdom


Our analysis is that the cost structures of Netflix and Comcast’s video businesses are quite similar:

  • Content acquisition: about 42% of revenues;
  • Content delivery: about 10% of revenues;
  • Marketing: about 15% of revenues.

It is commonly assumed that because Comcast usually offers the first release of hit shows, whereas Netflix mostly offers material at a later time, then Comcast’s content acquisition costs are higher than Netflix.

While this is true in absolute terms, when content acquisition costs are analysed as a percentage of revenues then we see that both companies are paying similar amounts for their content.

While the cost structures of both companies are proportionately similar, the two companies look very different when it comes to the top line:

Comcast’s video service revenue in 2012 was USD 20.1 billion, while the equivalent revenue earned by Netflix in 2012 was USD 3.6 billion – 82% less than Netflix.

However, somewhat bizarrely, while the two companies achieved dramatically different revenues in 2012, their subscriber numbers were far closer:

Comcast had 22 million video customers at the end of 2012, compared to Netflix who reported a total of 41.5 million subscribers at the end of  2012 – 80% of whom were streaming subscribers (92% of Netflix’s revenue in 2012 was derived from streaming services).

So Comcast earns more than 5x more video revenue than Netflix, but Netflix has 2x as many video subscribers than Comcast – which is strange, to say the least.

The reason for this anomaly is the flat rate $7.99 per month that Netflix is charging for its service, which we think is too low.

A flat rate price of USD 7.99 per month means that the monthly revenue for a Netflix customer is 90% less than the video revenue that Comcast earns from the average Comcast video customer, as shown below:

Title:Netflix vs. Comcast Overall ARPU
Companies:Netflix, Comcast
Markets:Television, Internet, OTT Internet TV
Years:2010, 2011, 2012

As we have shown in this article, the cost structures of the two companies are about the same or, if they are different, then the difference is insufficient to explain such a massive different in ARPU.

We are left with one conclusion: Netflix is chronically underpriced and so the company should introduce a tiered pricing model.

Leaving the service pricing as it is with a flat rate, while offering unlimited usage does not just mean that Netflix is leaving money on the table, it also means that Netflix’s steadily increasing content delivery costs (see prior figure) will soon start eating into operating profits.

Practically every broadband ISP around the globe began introducing tiered pricing models because they have found it economically unfeasible to continue to invest in the network infrastructure required to support the deliver of video content at broadcast scale without increasing retail prices.

Now that Netflix is hard at work developing its own content delivery network, the company will eventually succumb to the same realisation, which is that it must also increase the price of its service: not only would this allow the company to cover the steadily increasing proportionate cost of content delivery (see prior chart) but it would allow the company to invest more in content, thereby justifying a higher price.

Title:Comcast Corporation and Netflix
Sub Title:Comparison in Stock Performance
Companies:Netflix, Comcast
Markets:Television, Internet, OTT Internet TV, Pay-TV, Cable TV
Years:2010, 2011, 2012
Countries:United States

As it stands, while neither Comcast or Netflix has a fundamental advantage when it comes to the economics of their two television business models, it is Comcast that is presently winning the race when it comes to extracting value from customers.

Or more exactly, it is Netflix that has disadvantaged itself with a pricing model that is incapable of extracting the maximum revenue from subscribers.

Comparing the stock prices of the two companies it seems that Netflix’s investors might be unaware of this, or the other points made in this analysis:

Contact Andrew Sheehy at or follow him at