Why Programmatic Guaranteed Needs to Add the 'Audience' Layer

Programmatic guaranteed is on the rise, but what will really put it into orbit is the ability to deliver programmatic guaranteed audiences, in which buy-side data is added to supply-side data. In this industry byliner, Timmy Bankole (pictured below), RadiumOne’s head of supply operations, highlights why both sides need to start prepping hard for it.

Although it’s been a challenging time for publishers to grow ad revenue, a couple of positives have recently come to light. Following the Telegraph’s successful trial of header bidding (the method allowing publishers to bid on inventory at the same time, rather than in succession) we’re seeing another advanced trading method coming to the fore to increase revenues – programmatic guaranteed audiences (PGA).

In simple terms, it’s when a publisher, or Ad exchange, only provides an audience that exactly matches the advertiser’s target criteria. In other words, ads will only be served to people the advertiser explicitly wants to target.

For instance, assume a sports brand wants to target Chelsea fans aged 30-50 who play golf. As it stands today, the brand’s media buyer will agree with an Ad exchange/Supply platform beforehand to purchase a certain volume of impressions at a certain price – say 10 million impressions, at a CPM of £5, within a certain time period. The Ad exchange/Supply platform will then try to find that audience, but the reality is they can’t be sure they can meet all the targeting criteria simultaneously, in the time frame allocated.

In our example, it may transpire that only half the impressions served were delivered to 30-50-year-old golf-playing Chelsea fans, while the other half were delivered to an audience who only matched closely, with one or two of the original criteria (play golf or follow Chelsea or are 30-50). Consequently, the sports brand has decreased the efficiency of half its budget (£25,000) on people it didn’t intend to target.

However, with PGA, the exchange agrees to only send audiences that match all targeting criteria. It knows it can provide five million impressions based on the number of users in the given potential segment and time period, so that’s the volume agreed and the sports brand only ends up paying £25,000 and no dilution.

Timmy Bankole, Head of Supply Operations, RadiumOne

The eagle-eyed will spot a downside of this example for publishers in that their revenue is halved (advertisers are no longer buying the useless impressions from them). However, this ignores two vital benefits: guaranteed audiences can obviously command an incrementally higher price, and advertisers’ campaigns will have a higher ROI due to less wastage – ultimately, they will commit more spend in the long run.

This leads nicely on to why both sides need to start prepping for it – I see five key benefits.

It creates buying efficiency for most, possibly all, buyers. At RadiumOne, we currently spend a lot of our own money processing bid requests to determine if the data passed in the requests (e.g. a cookie) matches what the advertiser needs. In some cases, they do and we buy those impressions and in other cases when they don’t, we decide not to bid on those impressions, but we have listened and analysed those requests before making decisions. With PGA, due to the nature of how this works in that audiences have been matched beforehand, you are only sent requests you will buy, and publishers can sell non-defined impressions to other partners.

It also generates close to 99% targeting accuracy, as nothing is ever 100%. Once segments have been matched accurately by an SSP/publisher, buyers can overlay agreed matched data segments on campaigns knowing this will have no effect on delivery. With programmatic guaranteed (without the audience bit), buyers overlaying data leads to issues around wastage as audiences are defined by the publisher.

PGA is also a case of audience quality not quantity. As first-party data is the most valuable data out there, we are safe to conclude that quality will not be an issue but scale. What this means is publishers and brands should be willing to agree a rate that works for both parties. For publishers, they have the ability to monetise audiences they may not have deemed premium, while buyers get the ability to ensure they can hit the defined audiences anytime as agreed – making planning a lot easier.

I also believe this trading method leads to less fraud as the audience segments used are derived from brands’ first-party data. Once data is accurately generated, such as from a user-centric approach, it increases confidence in brands as they know their customers are being targeted and not BOTS.

PGA also leads to an incremental revenue stream. As I mentioned before, publishers can monetise audiences they didn’t deem as premium, but can also forecast better as to how many impressions can be sold on directly or programmatically.

In short, it’s a win-win for both parties – more efficiency and greater ROI for advertisers for which publishers can justifiably charge a higher premium. As a partner of ours at PubMatic, Emma Newman, neatly summarised to me: “PGA has the potential to solve some of the remaining friction between programmatic buyers and sellers by providing a more effective and efficient marketplace. Advertisers have increased confidence in the reach and performance they're achieving and publishers can maintain control over their inventory and, ultimately, their business.”