No More Ad Gaps Introduction

Imagine you’re a video asset, brought into the world by your content creator to be widely distributed and monetized in OTT viewing via inserted interstitial advertising. To support the insertion of ads, your content creator has conveniently placed various locations within you containing empty video slates to be overwritten with advertising at time of delivery. 

In a perfect world, you’ve done your job as a video asset by attracting viewers and filling every ad location, allowing your content creator to sip champagne and make plans for an expensive vacation. But here’s the truth: reality is not perfect. Video assets are often surrounded by ad gaps. And no one likes ad gaps. These contemptible video slates remain unfilled by ads, make zero money and cause consternation all around.

OK, you can stop imagining. 

The No More Ad Gaps blog series is dedicated to chronicling assorted trials and tribulations suffered by those of us burdened with occasional bouts of bad render-rates in ad-supported OTT video. Now, you may be asking, ‘What in the world is render-rate?’ Well, render-rate, and its close relatives use-rate and fill-rate, can be thought of as technical ways to figure out who to yell at when things are less than ideal. (See illustrations below.) But, in simplest terms, just think of low render-rate as corresponding to low monetization. Not good. 

And while we’re at it, here’s some lingo any video asset worth his or her salt should know.

THE VIDEO ASSET VOCABULARY GUIDE   Ad Break – a location in the video where (potentially multiple) avails exist.  Avail  – a location for an ad.  Demand  – the source of ads and advertising revenue.  DSP – Demand side platform; an aggregator of demand.  Inventory – the sum of ad opportunities where ads can be inserted.  SSP – Supply side platform; an aggregator of supply.  VAST – a Digital Video Ad Serving Template defined by the Interactive Advertising Bureau (IAB); it’s XML formatted data carrying ad data.  Wrappers – new VAST requests to other demand sources. With wrappers, demand sources can daisy chain and link to each other.

As a quick refresher, here’s a short summary of how the ad insertion process works. 

• Ad locations in content are marked-up with meta-data indicating where ad locations exist. 
• When a viewer watches a video, a server-side ‘manifest manipulator’ sends video to the player and looks for that ad-location mark-up. 
• When encountering that mark-up, the manifest manipulator makes a VAST call to an Ad Server, which responds with ad references. 
• The manifest manipulator fetches those ad references, processes them for insertion, and inserts them into the stream. 
• Rinse and repeat.

Although we’ll be focused on server-side ad insertion (SSAI) of video, most of these problems are not dependent on that architecture and apply to client-side insertion as well. Now buckle up, it’s about to get technical.

Orientation

To understand the terms, see the figure and explanation below. 

• Fill Rate: The fill-rate measures how efficient the ad decision is at filling the available ad opportunities. It is proportional to the number of ads returned by the ad decision server for the ad breaks where ad placement requests are made. 
• Use-Rate: The use-rate measures how much of the fill-rate was used. It is a measurement of the efficiency of the ad insertion process. It is the ratio of ads inserted into stream(s) compared to the ads returned by the VAST call. Note that if the VAST call is over-filled – that is, if it returns more ads than can be placed in the break –  the use-rate can be artificially low, even if all ads are inserted!
• Render-Rate: The render-rate measures how well the whole ecosystem is monetized. It is the ratio of ads inserted into the stream(s) to the total available ad slots (or time) in the breaks. This measures the real efficiency of the ad insertion process end-to-end. 

The best way to measure these ratios is to take the sum of the times for the returned or rendered ads divided by the sum of the ad break durations – that is, a real ratio of the total break time that was filled with ads. But creative durations are often ignored, and so these rates are often measured by assuming the creatives are 30 seconds long and that a break is divided into 30 second slots. This realization is the foreshadowing that not all is necessarily well in the land of ad breaks. A one-minute break that is filled with four 15 second ads would have a 100% render-rate when measured using time, but a 200% render-rate when measured using 30 second slots and inserted-ad count. 

Problems. Problems. Problems. 

So what can cause low render-rates? It’s tempting to identify the guilty companies: name names, point fingers, gossip! But we won’t. We’re working in advertising, and we need all the friends we can get. 

One reason render-rates may be low is that figuring out end-to-end issues can be complicated. Low render-rates are an area ripe for optimization, and one that many publishers don’t seem to be aware of. Arriving at the OTT party with ad opportunities, it is sometimes a rude awakening to find that there are so many reasons for that inventory to go unfilled. Providing solutions to these complicated issues is one of the things we do at Wurl. 

In the next blog post, we’ll start with our first render-rate issue: bad VAST. Stay tuned!