Multiscreen
Video Processing
Monetisation
Application Note
September 2013Multiscreen up-ends existing TV business models, but at the same time the unicast nature of OTT delivery provides opportunities to build new models. Such solutions demand end-to-end thinking, meaning this is an area in which Ericsson’s systems integration expertise means the Multiscreen Video Processing solution can solve operator’s multiscreen integration challenges.
The Ericsson Multiscreen Video Processing solution features a highly-scalable and cost-effective compression solution for delivering premium linear TV content using OTT networks. It enables the full range of connected devices to be targeted, from small-screen smartphones to large-screen connected TVs.
This document explains how the monetisation solution can be applied to enable a number of typical business models. Ericsson understands that every operator’s unique circumstances will require variation from these examples, hence this document is intended to provide a source of ideas.
1. MoNetisiNg LiNeAr MuLtiscreeN
1.1 Overview
Although some operators already have ecosystems that support advertising operations on their existing networks, for many operators OTT advertising is their first advertising deployment. The number of steps needed may therefore seem daunting.
However, deploying multiscreen advertising can be broken into a number of stages. The key dimensions are as follows:
> Pre-roll vs. Mid-roll
> In-house vs. Outsourced Ad Sales operations
> Targeting: network-wide, simple (e.g. regional-based), and in-depth (demographic-based) In other words, whilst in-house selling of demographically-targeted mid-roll adverts may be a desired end-goal, it is not necessary to launch with that level of complexity.
Hence one example roadmap could be as follows:
> Deploy pre-roll advertising with an outsourced ad network partner, without targeting
> Then either bring ad sales in-house, or move to mid-roll
> If ad sales are now in-house, move to mid-roll, and vice versa
> Integrate the subscriber database into the ad sales operation to implement targeting The following sections explain the dimensions in more detail.
1.2
Pre-rOll vs. Mid-rOll
The advertising solution supports both pre-roll ad insertion and mid-roll ad replacement. Pre-roll means the insertion of adverts before the stream has started, and hence represent additional advertising. Mid-roll replaces existing adverts with new adverts specific to multiscreen devices. However, to replace existing adverts needs both agreement from the original programmer, as well as frame-accurate knowledge of where the adverts are located in the stream. Generally this means integrating with the original playout server.
Hence pre-roll has a number of potential advantages:
> Deploying pre-roll in partnership with an external ad sales company enables OTT services to begin generating cash with very little infrastructure. Thus multiscreen can quickly be turned from being just an operational cost to being the start of a revenue source, and the experience gained can be used to justify further investments
> Inserting adverts before the user actually sees the selected TV service avoids needing to know any advert location metadata
> For an operator just starting with advertising, pre-roll insertion represents a smaller change to existing carriage agreements. Some agreements may already allow this.
> Pre-roll adverts are less constrained to fit in with the established 15, 30 or 60 second slot model, and hence offers advertisers more editorial flexibility
However, pre-roll has some disadvantages:
> Consumer tolerance for pre-rolls is lower, since it represents a delay before the content plays. This restricts the number and length of ad slots that can be sold
> Advertising can only be displayed when connecting to a service, which for the longer-form programming typical of linear multiscreen services means only a couple of ad slots per user per session. In contrast, mid-roll advertising typically offers 15 to 20 minutes of advertising per hour.
Mid-roll involves replacing advertising already present in the linear broadcast. Achieving this requires several key elements to be in place:
> Agreement from the programmer to do this (given it is typically the programmer who sold the original adverts)
> Potentially some form of revenue sharing arrangement with the programmer
> Reporting as to which ad slots were replaced so the programmer knows not to pay the advertiser for the original slot
> Frame-accurate metadata showing where each replaceable advert is in the stream. This requires cooperation from the programmer for them to either insert SCTE35 markers in the stream around replaceable adverts, or to share playout server metadata. In this second model, the playout server metadata is injected into the transcoder via the CableLabs ESAM interface Mid-roll advertising is therefore more complex, but balanced against this are a number of advantages:
> Consumers are not asked to accept additional advertising
> With more advert slots available per user session, mid-roll offers a faster growing revenue stream
> Mid-roll slots are more valuable than an equivalent length pre-roll slot, thus giving a better return to justify the complexity
> Although the need for programmer cooperation may seem a large hurdle, there is a motivation for programmers – targeting. Targeted adverts are more valuable than non-targeted, and so there is a motivation to replace non-targeted mid-roll adverts with targeted. Only operators of OTT networks can easily deploy targeted advertising, forming the basis of a symbiotic relationship In summary, therefore, pre-roll offers a simple foundation on which to launch multiscreen advertising and learn about operating such systems. However, there is a limit as to the revenue that can be generated, creating a motivation to work towards mid-roll advert replacement.
Mid-roll
1.3
in-hOuse vs. OutsOurced Ad sAles
Selling advertising requires an inventory of adverts to sell. This is managed by a campaign manager, which maintains a database of adverts and metadata about the campaigns each advertiser wants to run (for example “show this advert N times per week, with at least 60% of those impressions served during primetime hours to 30-40 year old males”).
The multiscreen workflow interfaces to the campaign manager via one of two industry-standard interfaces – SCTE-130 or VAST 2.0. As a result, it is easy to pick-and-choose campaign managers, or even run multiple in parallel for different applications (for example, pre-roll vs. mid-roll). However, more importantly, these interfaces mean it is as easy to have the campaign manager co-located or external to the head-end. This creates a market for advertising networks to act as
The main advantages to using an external ad network are:
> Allows multiscreen advertising infrastructure to be built up in stages, rather than needing to learn several new components at once
> Much easier to scale down to smaller operations, for example if the current strategy only involves selling pre-rolls on a subset of launched channels
Although different ad networks will serve adverts in different formats, the MVP advertising solution includes file-to-file processing to ensure a segmented version of the asset is seamlessly made available for delivery over a specific OTT deployment.
The two main disadvantages to using an external ad network, however, are:
> The ad network will take a slice of the revenue generated from each slot filled, thus reducing the potential revenue stream
> Ad networks are very country-specific, and not all countries are as well served. In addition, not all ad networks sell video adverts, and instead focus solely on banner ads. Thus in some regions it may not be as easy to find a suitable partner
For these reasons, the most successful strategy would seem to be to launch with an external network and once the operation grows, bring the ad sales in house.
To offer a complete end-to-end proposition, Ericsson has partnered with BlackArrow to be the selected reference partner for in-house campaign management. BlackArrow’s suite is mostly a hosted solution, meaning little IT infrastructure needs to be deployed on location. In addition, BlackArrow’s success in this field means their UI will be familiar to many advertisers.
The ability of the advertising solution to route different advert requests to different campaign managers facilitates transition from outsourced to in-house operations. For example, initially only pre-rolls may be run from the in-house operation, and the lower volume of pre-rolls means it is easier to manage a sufficient inventory of adverts.
external Ad sales
1.4
tArgeting OPtiOns
OTT delivery is inherently a unicast delivery of broadcast content. This greatly simplifies individualising the stream for, in theory, every individual user. As a result, individualised targeting of adverts becomes easily implementable for the first time, and the extra CPMs1 this generates is
the desired goal for many.
However, individualised targeting brings three associated problems:
> Integrating the campaign management system with the advertising system and the subscriber database
> Collecting user demographic data to provide a sufficient description of each user to enable correct targeting decisions. For example, the CRM2 system may only collect a user’s name and
address, but not their age, occupation etc
> Having a suitably scaled ad sales operation to provide a sufficient inventory of adverts. In other words, there is little sense investing in targeting infrastructure if only a few advertisers are providing adverts
However, there are other options which avoid needing to launch day one with a fully targeted advertising system. Alternatives include:
> Network-wide: offering adverts for all users using the OTT service. This in itself provides some ready-made demographic targeting, since OTT users tend to be early adopters, and hence are typically younger and more affluent. Thus adverts can be biased towards higher-end brands and products
> Simple: although the operation may not be sufficiently scaled for individualised targeting, it may be possible to group users. One option could be regional-based, by routing users from different parts of the network to different advertising servers
Therefore as with the other two dimensions covered above, it is quite easy to break the targeting problem down into simpler steps.
2. MoNetisiNg LiVe-to-FiLe
As discussed at the start of this document, multiscreen delivery up-ends existing business models. One of the reasons for this is that multiscreen devices mix lean-forward (i.e. interactive content consumption) and lean-back (i.e. passive content consumption) usage modes.
For example, content discovery occurs in a far more interactive usage mode than is possible in standard linear TV, meaning multiscreen services need not be constrained to the grid EPG. Instead it is possible to seamlessly mix content suggestions from different stages in the content life cycle – live (“What’s On Now”), catch-up (“You May Have Missed”) and VOD (“See also these box sets”). The different stages can cross-sell and support each other – for example if the user has recently watched a particular show, older episodes could then be offered through the UI.
The term live-to-file is used here as an umbrella term to cover a number of different use-cases:
> Catch-up TV: offering “backwards EPG” content to allow users to catch up on missed shows
> Start-over TV: offering the ability to jump back to the start of the current show if a user tunes in part way through
> nPVR (as in Network PVR, also known as Cloud PVR): offering users the ability to control which programs they record and save for future viewing
In all cases, three main monetisation models are possible:
> Advertising-based: current live-to-file solutions capture the original adverts embedded in the live to file asset, but multiscreen opens up the possibility to re-sell those slots
> Non-Advertising: sell the service on a pay-per-view basis, or as part of a higher subscription tier
2.1 Advertising-bAsed MOdels
One of the challenges with existing live to file services is that the advertising embedded in the asset during the record stage quickly becomes obsolete during the catch-up window.
For example, most advertising in the days before a major national holiday or festival period will be offers relating to that period. However, such adverts are quickly obsoleted during a 14 or 28 day catch-up window. This presents an opportunity to re-sell those advert slots every time the asset is viewed, in order to allow more relevant advertising to be inserted.
The same is true of start-over TV and nPVR, meaning that the original advertising slots can continue to be resold.
2.2 nOn-Advertising MOdels
Live-to-file content can also be monetised via means other than advertising. The lean-forward nature of multiscreen devices makes it far easier to enable content discovery of both linear and non-linear content. Live-to-file is therefore a service that is valued by consumers, and hence one that they are willing to pay for. Restricting access to live-to-file content only to subscribers of higher tier packages means OTT services create a straightforward means to increase ARPU. In addition, the different live-to-file use-cases offer increasing value, allowing different tiers to offer increasing propositions. For example, the lowest package could offer no live to file services, while the next package could be limited to start-over. Moving up from there, access could be added to catch-up content, in a time-limited window, whilst the highest package could add nPVR with a certain amount of data storage to allow viewing outside the catch-up window.
2.3
hybrid MOdels
The final model covers the mixed mode model. Advertising and non-advertising models can be combined in a variety of ways, owing again to the unicast nature of multiscreen delivery. One of the main models is to steadily reduce the amount of advertising in live-to-file content as users move up pricing tiers. For example, at the lowest-level package, catch-up content retains the full 15 to 20 minutes per hour of advertising. The next tier up could reduce that to, for example, 5 minutes per hour, whilst the top tier removes advertising entirely.
The manifest manipulation element of the advertising solution makes such models very attractive, because with this technology end users are always given a manifest that is the same as a linear manifest. Such manifests only offer devices a limited “moving window” onto even VOD content, meaning the ability for users to skip through adverts can be controlled.
coNcLusioN
In keeping with the concept of Discovery-Driven Planning3, the ideas presented in this document show
how the new and exciting area of multiscreen monetisation can be broken down into smaller steps. This reduces the size of the upfront investment, and hence also project risk. In addition, it allows learning along the way. It also offers different entry points into multiscreen monetisation, with the entry points matched to the capabilities and prior knowledge with which each organisation starts out.
reFereNces
1. “Cost Per Mille” – the rate paid by advertisers per 1000 impressions of an advert 2. Customer Relationship Management system