This blog continues a multi-part series that suggests ideas by which our customers could deploy and monetize video applications.  Most of these ideas (as stated) are as yet un-tested in the industry, but we believe there is enough value in the ideas to encourage our more clever readers to enable or launch a service or two which might prove successful.

Last week, I talked about content adaptation.   This week, we will talk about advertisement insertion, so let's just dive right in. 

The insertion of advertising into a content stream is a key way for content providers and operators to monetize their services.   There are several ways by which advertising can be added to content:

  • 1. Text overlay
  • 2. Banner overlay
  • 3. Bug insertion
  • 4. Pre roll
  • 5. Post roll
  • 6. Logo enhancement

With text overlay, a simple text message is placed along the top or bottom of the image as it is playing.    Banner overlay is very similar, but often involves some background transparency where the 'banner' is a field which is super-imposed on the primary content image.  Bug insertion is often used by television and inserts an image or animation in the lower corner which is distracting to the user, but not overly intrusive.  Pre and post roll is the insertion of a commercial either before or after the content.   Logo enhancement is a challenging new concept wherein a company's logo could be inserted into the video stream to look like an original part of the content.      

We know two things about the value of advertisements in a web environment:

  • 1. A typical text ad or 'pop-up' brings small fractions of a dollar per view
  • 2. A directed ad can bring 100-1000 times as much revenue as a standard ad

There are three things which define a directed ad:

  • 1. The person's general demographics
  • 2. The person's location or time of day
  • 3. The person's personal preferences

An example of using Demographics is a person who watches a baseball game on their mobile handset in the middle of the afternoon.   An example of Location is a person who watches that baseball game, but happens to be located near one of the restaurants who pay for advertising space during that baseball game.    An example of Personal preferences is the person who likes steak, when one of the restaurants close to the baseball stadium is a steak eatery.  

Each time the advertiser can attain another degree of direction for their target audience, the value of their advertisement goes up and so does the revenue associated with running it.  

There are two or perhaps three places where advertisements can be added in the network: Either in the content adaptor (the subject of the previous blog), in a gateway (assuming operator intervention) or possibly in the end user device.   This last one is futuristic, but not out of the realm of possibilities, especially when we look at Set Top Boxes (STB), femtocells and media phones.

The advertisement enabling of content could easily happen during content adaptation.  The problem with this level of ad interaction is that it is limited to generic ads unless the user opts into some sort of service and reveals information about him or herself.   Therefore it is more challenging to deliver directed advertising to the mobile user.   However, this is a typical scenario for the Web service providers, who can still make good revenue even on generic ads.  

Eventually, more smart phones will be equipped with GPS and applications that could deliver that GPS information to the Web Service provider

Ad Insertion Scenario for Mobile Operator (MO)

The Mobile Operator has a problem with video content as it taxes their service infrastructure and they get no additional revenue from it.  As we stated previously, they may choose to cap or block this content or the services themselves.   This will lead to legal challenges and customer dissatisfaction.

Mobile Operators cannot insert advertisements into a user's content stream if the MO does not know about the stream to begin with.   So we can envision two methods by which the MO might discover the existence of a video stream:

  • 1. Deep Packet Inspection (DPI) of user traffic which would identify an HTTP request and corresponding RTSP stream
  • 2. Cooperation with the CDN or Web Service provider

DPI would enable the MO to control the traffic and types of traffic going through their network and there may be legal reasons why it may go on anyway.   However, it is not cheap and not easy to do in real time for the amount of traffic that will be traversing the MO networks.  However, there may be less resource-intensive techniques where the MO could use the Natural Address Translation (NAT) capabilities of its border elements to allocate service requests to certain well-known IP addresses to specific port numbers which would identify the types of video services likely to be offered by that location.   The MO could also maintain lists of popular URLs from which the bulk of content is being downloaded at any given time. This would greatly reduce the workload required by the DPI engine and reduce the scope of legal exposure in the event that people are not happy with the MO meddling with their traffic.

However, there are also ways the Web service people could combat the technique through encryption or disguising their traffic in a proprietary end to end protocol.   Ultimately, the traffic may go underground in a peer to peer relationship and the MO could not afford to let that happen.

A better model would be for the MO and Web Services Operator (WSO) to come to an agreement.   The MO would promise not to block or limit traffic in exchange for a piece of the advertising revenue.   In fact, since the MO is better equipped to target directed advertisements, a unique arrangement could be struck upon where the WSO could alert the MO that a user has made a request for content.   This could be easily done simply by looking at the IP address which comes from the MO Carrier Core network.   The WSO would immediately know which MO the request came from and alert the MO that a user with this source IP address made a request for services.  This IP is most likely translated by the border elements in the carrier core anyway, so the carrier does a look-up and identifies the user, does a database dip to find out the user's preferences and current location and runs an advertisement.  The beauty of this is that the ad is worth so much more, that the MO and WSO should be able to negotiate an ad revenue sharing model which will be advantageous to both of them.  Ultimately, the customer also wins out because if they are going to be forced to view ad content anyway, it might as well be ad content which is of interest to them.   The advertisers are also pleased because they are getting a higher utilization of their content and this may be most important because advertising revenue is what makes the world go round.

Next time we will talk about interaction which can be caused by viewing content on a mobile device.