Client-Side Ad Insertion (CSAI) vs. Server-Side Ad Insertion (SSAI)

Title Image: CSAI vs. SSAI
 

While client-side ad insertion (CSAI) has long dominated streaming, server-side ad insertion (SSAI) is a burgeoning solution to the ad-blocking challenges and delivery complexities associated with video-based monetization.

 

Even in this time of economic uncertainty, advertising helps drive a significant portion of online video revenues. Yet how those ads are delivered — or not delivered, in the case of some older technologies that can be effectively blocked — is critical to helping content owners and advertisers cover acquisition and delivery costs, while also getting key advertising messages out to consumers.

I spent the early Spring analyzing responses from a survey for Streaming Media magazine on video-based revenues and monetization. Now that the survey report has published, it’s apparent that ad-supported streaming media delivery continues to be a major monetization strategy.

In fact, well over half of survey respondents that monetize their content indicated that they use some form of advertising to either increase revenue or lower end-user consumer fees for viewing the ad-supported content. As such, it’s critical for these content owners, rights holders, and media distribution partners that any and all video advertisements not only make it to the consumer’s device — be it a laptop, desktop, set-top box, or mobile device — but also play properly.

Respondents to the survey were also asked about various approaches to delivering video-based ads to end users. The most popular approach was server-side ad insertion (SSAI) at 45% of all responses, followed by interstitials (31%) and client-side ad insertion (CSAI) at 24%. To put it another way, more than half of organizations generating revenue from advertising have yet to make the move to SSAI.   

While CSAI has long dominated streaming and over-the-top (OTT) ad delivery, it has some challenges that SSAI addresses, so it’s not surprising to see SSAI gaining traction in the overall video-based ad monetization landscape. But there are still a number of places that CSAI provides benefits.

Before we delve into the challenges and opportunities of each, let’s look at the consumer landscape for video-based advertising.

 

What Do Consumers Want?

While content owners and advertisers would love to pile on as many ads as possible during short-form and even user-generated content, consumers have consistently said they’d rather watch content without advertisements. In fact, many pay for the no-advertising privilege by subscribing to OTT services that trade the subscription fee for a ad-free viewing experience.

When consumers do watch content supported by video advertisements, there’s often an annoyance factor that drives them to look for ways to view content without viewing ads. And in the cat-and-mouse game of ad serving and ad consumption, a number of consumers deploy ad blockers to suppress the playback of ads while they’re viewing media content.

These ad blockers are an outgrowth of earlier web-browser ad blockers, which effectively kept banner and pop-up ads from displaying. In the online video space, though, they’re as much focused on convenience — shortening the overall length of time it takes to consume a premium content television show episode, for instance — as they are on eliminating the need to interrupt a program’s natural flow with an ad for some product or service that has little to do with the content.

A large majority of ad blockers function on the premise that the video advertising is pre-loaded to the consumer’s video player on whichever device they’re viewing the content. This is known as client-side ad insertion, or CSAI.

 

CSAI and Ad Blockers

In a pre-roll CSAI setting, one or several of these pre-loaded ads will play on the consumer’s device prior to the first portion of the primary content. In terms of classic linear television viewing, think of pre-roll ads as commercials at the top of the hour, between the end of one TV program and the beginning of the next program.

Because these pre-roll video ads play before the first portion of the primary content is played, some early ad blockers would successfully block the pre-roll ads but inadvertently keep the playlist (the file sent to a consumer’s device that lists the order in which advertising segments and primary content segments would play) from properly executing.

The end result would be a total failure of any content in the playlist. In fact, using sophisticated business rules, coupled with ad servers and streaming servers, an online content distribution platform could alert the end viewer that their device was using an ad blocker, and request that it be turned off in order to properly view content.

Improvements in ad blockers, based on hundreds of thousands of users, resulted in more sophistication. Some were smart enough to allow these pre-roll ads while blocking subsequent ads that occur between segments of the primary content. The consumer’s video player, upon not finding a mid-stream video advertisement (known as an interstitial) to play from the playlist, would assume there was an issue with the ad server, skip those interstitials, and continue on with the next segment of primary content.

 

CSAI vs. SSAI

How successful were these CSAI ad blockers? In a test bed that I designed for a major media company a few years ago, we were able to demonstrate an average of 95% ad-blocker success rate on Windows-based browsers and 74% on Mac-based browsers.

With ads successfully being blocked at these percentage levels, why has CSAI persisted for so long?

First, it is much easier to create an ad-playback event using broadcast-standard triggers. Wowza Streaming Engine, for instance, can be tuned to monitor for SCTE35 broadcast signaling, to let a client-side player launch an advertisement. When serving up Apple HTTP Live Streaming (HLS) manifests to deliver content, a CSAI approach can leverage HLS manifest manipulation for direct ingestion of a video-based advertisement at a playback client.

Second, it’s a bit easier to deliver video-based ads separately to an end-user’s device ahead of time with CSAI than it is to integrate ads in between the primary content segments (a process sometimes referred to as stitching ad and primary content together into a single ad).

This means that CSAI also fits well for live events, such as breaking news and sports, where there may not be predefined “commercial breaks” and where a broadcast trigger can instruct the client’s player to deliver ads that are already pre-populated on consumer’s devices.

Third, because ads could be delivered on a per-viewer basis to the client’s device, ad playback could be more finely targeted to a viewer’s demographic and geographic details, allowing a more diverse set of ads to be delivered in a given zip code or viewer demography. SSAI, at least in the early days, suffered from the technical deficiencies of stitching, with the end result often being that the same ad being played across wide demographic and geographic swathes of viewers.

As a bit of a corollary to this point, it’s important to remember that SSAI is much more resource-intensive — oftentimes requiring transcoding as well as manifest manipulation — which could be a game-limiting consideration when attempting to deploy an SSAI solution.

Fourth, ads that were pre-loaded to the consumer’s device would sometimes be counted as played, even if an ad blocker was successful. This meant that, while CSAI was ineffective — meaning there was no way to guarantee that the ad had actually played — it was mature enough that ad-monitoring services would allow the content distribution system to charge the advertiser as if the ad had run.

Finally, the real-time coordination between ad servers and streaming media servers were much more technically challenging than the less sophisticated CSAI approaches. In fact, once media players on most consumer devices were equipped with dual players, an ad served by CSAI could be seamlessly switched — even if the ad used a different data rate or, in some instances, a different codec, from the primary content — without a noticeable glitch between ads and primary content.

 

What SSAI Promises to Deliver

For all the technical challenges of SSAI, however, it represents the best approach for advertisers and content distribution systems. Here’s why.

First, SSAI approaches mean that video advertisements and primary content are delivered to the consumer’s device as a single stream. No content needs to be pre-cached to a consumer device, so ads aren’t out of date (stale) or — especially if the consumer is traveling — geared towards a different geographic region.

In addition, this also means that the content can be delivered to web browsers, which often have only one player engine, widening the audience for content delivery.

Third, measurement of ad delivery can both be confirmed and verified, based on how much of that single stream is consumed on the viewer’s device. There’s no longer a need to guess as to whether the ad served is actually played, which means that ad measurements are much more accurate between the advertiser and the content distribution platform.

But challenges still remain for SSAI delivery, especially when it comes to live content delivery.

 

SSAI Challenges to Overcome

SSAI offers challenges and advantages when it comes to finely targeting ads to demographies and geographies. Ironically, the move towards HTTP-based video streaming, with its small fragments of a primary MP4 file (fMP4) actually adds enough overall latency to make the act of stitching in the transition between a piece of primary content and an advertisement easier to accomplish.

This means that SSAI is well suited for use in Apple HTTP Live Streaming (HLS) or MPEG Dynamic Adaptive Streaming via HTTP (MPEG-DASH), both of which use playlist or manifest files to let the consumer’s device know which rendition (bitrate-codec combination) to play at any given time.

Second, there’s the question of whether video advertising is encoded the same way as primary content. Now that most premium content is centered on a single codec (H.264 or AVC) it’s easier for streaming servers to stitch in ads alongside primary content that, in the past, might have needed to be transcoded to fit together.

Third, media servers can easily transrate content to a lower data rate, thanks to the same adaptive bitrate approaches that HTTP-based delivery uses for their playlist or manifest files.

Fourth, due to the nature of stitching multiple streams together in a just-in-time manner, the number of manifest files created is directly proportional to the number of end viewers. While CSAI allowed for caching of personalized manifests or playlists, SSAI doesn’t have the same capability. At least not yet.

Finally, ignoring for a moment whether the ad is delivered via CSAI or SSAI, there are a number of competing ad-serving technologies — with acronyms like MAP, RTB, VAST — that industry groups like the Internet Advertising Bureau (IAB) are seeking to rationalize into common header information. Until then, the act of inserting video-based advertising in a form that can be properly counted and audited adds an additional level of complexity.

 

Conclusion

In this blog post, we’ve only scratched the surface of video-based advertising as it relates to online video. While CSAI was the most popular approach in the past, SSAI is making progress on the technical side.

Regardless of what you’re looking to deliver, Wowza’s Live Broadcast and OTT Streaming Solutions open up new monetization opportunities by ensuring the best viewing experience possible on mobile devices, set-top boxes, and everything in between.

 

Search Wowza Resources

Categories

Subscribe

Follow Us

Categories

About Tim Siglin

Tim Siglin, who has over two decades of streaming media design and consulting experience, and an additional 10 years in video conferencing and media production, has written for Streaming Media magazine and other publications for 23 years. He has an MBA in International Entrepreneurship and currently serves as the founding executive director of Help Me Stream Research Foundation, a 501(c)3 dedicated to assisting NGOs in emerging markets with the technologies needed to deliver critical educational messages to under-served populations.