While we’re seeing a fair amount of innovation and inventiveness with the advertising experience on connected TV services, media companies and ad tech businesses spend a lot of time trying to recreate aspects of the ad experience which audiences are used to from traditional linear TV. One big part of this has been recreating the ad break, where a series of ads are shown back-to-back during breaks in programming.
But some things which are quite straightforward on traditional linear TV become much more complicated in CTV environments, and constructing ad breaks is one of them. On traditional TV, each channel has a finite number of ad breaks each day, which are broadcast to all users identically (though that’s starting to change with ad replacement via addressable TV). On CTV services, each ad break is constructed individually for each user whenever they watch content, with ads typically targeted to that specific user. This means manual curation of ad breaks is impossible, everything has to be done automatically.
Hence the creation of ‘ad pod bidding’, a technique which helps solve for some of the complications and specificities of TV ad breaks.
The Basics
Ad breaks – known as ad pods in CTV services – open up a lot of opportunities, but also challenges, relating to the fact that multiple ads are played one after another. Brands will usually have specific strategies and requirements which are unique to ad pods.
The most obvious – and one of the more commonly cited issues with the CTV ad experience – is multiple ads for the same brand appearing in the same ad break (the record seen by this writer is the same toothbrush ad played five times back-to-back). Brands might want to show multiple ads during the same programme in order to reinforce their messaging, but they typically don’t want those ads to appear in the same ad break – and especially not right next to each other.
In some cases however, a brand might actively want to bid on two ad slots in the same ad break, in order to run sequential messaging.
Competitive separation is another challenge, carried over from traditional TV. Brands often want to know that their ad won’t appear in the same ad break as one of their competitors’ ads, to avoid clashing messages.
And the positioning of each ad within the ad break, as well as the wider context of that ad break, also informs brands’ strategies. Brands are generally willing to pay more for the first or last ad played in each ad pod, as these are most likely to be seen by the viewer. And they may want to know how long the ad pod is overall – being the middle ad in a 90 second, three ad long ad pod is very different to being the middle slot in a five minute ad break.
Ad pod bidding essentially allows media owners to send all the relevant information over to buyers, in order to inform their strategies and avoid problems created by ad breaks. It does this by showing the entire ad pod to the buyer, rather than just an individual ad slot within the ad pod.
The Technical Details
Various sell-side platforms began offering pod-bidding to their publisher partners a few years back, so it’s known by a few different names (ad pod bidding, pod bidding, per-pod bidding). However guidelines for ad pod bidding were introduced into OpenRTB, the IAB Tech Lab’s specification for programmatic advertising, in version 2.6 which released a year ago, bringing an element of standardisation.
Prior to the release of version 2.6, OpenRTB already had features which allowed multiple ad requests to be sent within a single bid request. The ad pod bidding specification builds on this by introducing new objects relating specifically to the ad pod – allowing information about each individual impression to be sent through to the buyer. These are:
maxseq: This specifies the maximum number of ads which can be served into a “dynamic” video ad pod (see below)
poddur: The duration (in seconds) of the dynamic portion of an ad pod
podid: A unique identifier for an ad pod. This ID allows the buyer to understand that they’re being offered multiple slots within one ad pod (since the podid will be the same on each of those impressions).
podseq: The position of the whole ad pod within the content stream. The value for this will either be ‘-1’ (for the last ad pod in the content), ‘1’ (for the first ad pod), or ‘0’ (for any other ad pod).
slotinpod: The position of the individual ad slot within the ad pod. The value will either be ‘-1’ (the last ad in the pod), ‘1’ (the first ad in the pod), ‘2’ (the first or last ad in the pod), or ‘0’ (any other ad slot.
As hinted at above, the OpenRTB spec also defines three different types of ad pod – structured, dynamic, and hybrid.
Structured ad pods have a set number of slots, each with a set time (e.g one 15 second ad, followed by two thirty second ads, followed by one 15 second ad).
A dynamic ad pod sets a length for the whole ad pod, but doesn’t specify an exact number of ads, or ad lengths. So a dynamic ad pod with a total length of 120 seconds could be filled with four 30 second ads, three 30 second ads and two 15 second ads, two thirty second ads and four 15 second ads, etc.
A hybrid ad pod is a combination of the two, where part of the ad pod is defined by the seller, and the rest is left open. So for example, a seller might require that a 120 second ad pod starts with a 30 second ad, and then leave the rest to be filled dynamically.
So bringing all of the above together, when a seller runs an ad break during a piece of content, they send a bid request offering up impressions across the whole ad pod, packaged up within one request. Bidders which receive this request then use this data to inform their response, and bid for whichever slots they’re interested in accordingly. The seller can then check for competitive clashes across the ad pod in the bids they receive back, providing competitive separation.
The Pros and Cons
Ad pod bidding brings a lot of basic functionality advertisers are used to from regular TV advertising into the CTV world. This gives buyers more control and transparency into what exactly they’re buying, and allows sellers to charge premiums for their more valuable inventory.
Dynamic ad pods also give sellers more flexibility in their monetisation, seeing as they can take bids from a wider ranger of buyers, rather than having to fill ad slots of a set length and number. This provides higher yield and fill rates.
The difficulty for ad pod bidding is that not all buyers are set up for it. On a basic level, DSPs need to be set up to understand the ad pod information being passed through from the sell-side, and to know how to adjust bids accordingly. And they in some cases need to pass through extra information about the creative to the sell-side, in order for the seller to be able to tell what ads are actually being proposed for each slot. Without this, competitive separation, and even just preventing the same ad from appearing multiple times during one ad break, can be difficult (though some sellers have their own tools to try to solve these problems).
