Dylan Temple-Heald, Director – Precision at Publicis Media, deep dives into the basics of programmatic buying.
How would you define programmatic and what role does it play in the current advertising landscape?
In its most basic form, programmatic is the automation of buying digital ad space – combining inventory, data and technology in order to automate the delivery of the right message to the right person at the right time. There are a lot of things that go into making the buying of digital ad campaigns programmatic, but this is the simplest way to describe it.
Programmatic advertising allows for the consolidation of an ever-expanding and complicated landscape of inventory providers, tech solutions, and data vendors currently available in the market. If you were to run all of the individual publisher and tech solutions independently, you would have vast amounts of overlapping audiences; frequency capping would be redundant due to isolated media buys, and wasted media spends would be astronomical. Running your activity programmatically allows advertisers to consolidate most of their digital display and video activity, allowing for effective frequency capping across a majority of the publisher buys leading to greater efficiency, less wasted ad spends, and greater control of your digital media delivery. On top of that, programmatic allows brands to target the right audience and deliver the most relevant message through the utilization of audience data, making it possible for brands to deliver more engaging messaging to their desired audience.
What are the different programmatic models?
Open Exchange or Open Auction is the most common and widely recognized programmatic buying model. This is where the vast majority of available inventory is bought and sold. Traditionally, [it is] the most cost-effective entry point into programmatic advertising.
Private Auction or PMP (private marketplace) is an auction buying model where a deal is set up by the publisher or network with a fixed price that buyers agree to. This is usually open to a large number of advertisers on a non-guaranteed deal.
Unreserved First Look or Preferred Deals is a non-guaranteed buy where a publisher and buyers agree to a fixed price. The difference between Preferred and a Private Auction is the number of advertisers invited to access the deal. A smaller number of advertisers are invited to see the available Preferred inventory before being available in the PMPs or open auction. [It is] normally set with a premium that advertisers and publishers agree on beforehand.
Programmatic Guaranteed is, as the name suggests, a guaranteed buy where the number of impressions is agreed and fixed along with the price. This buy is similar to a direct insertion order (IO) buy; however, it’s transacted programmatically, which allows for greater efficiency and automation of both planning and delivery of your direct IO buys. This is also a particularly practical way of planning and buying in busy periods of the year where competition for inventory in both Open Auction and Private Deals is high and is not guaranteed.
What is the role of data in programmatic?
Data plays a major role in programmatic – without data, we wouldn’t be able to target specific audiences or understand how campaigns are performing.
We’re able to access data from three main sources:
Data that comes from the advertiser through any number of ways, such as behaviors captured from their site visitors, CRM data, app downloads, social media, etc. – any data that is owned by the advertiser.
Data that mainly comes from publishers (it’s their first-party data). We use this in a second-party fashion where we access the data through their inventory deals on behalf of advertisers. This data comes in similar flavors as a client’s first-party data like site behavior, interests based on the content being accessed, purchase behaviors, app installs, and location.
Data collected from various sources and sold through a reseller or aggregator. We access this through either a data management platform (DMP) or directly from resellers through integrations within the demand-side platforms (DSPs). 3rd party data could be anything from standard demographic, interest, location, and purchase intent, just to name a few.
What about AI and how could it drive change?
AI in programmatic is not a new development; machine learning has been around for quite some time. Machine learning allows us to analyze and automatically take actions based on vast amounts of data points that would not be possible by a human in the fraction of a second it takes to deliver an ad in real-time. This is the foundation of programmatic today.
The way I see AI evolving in the advertising space would not only drive change in programmatic, it would drive change in the advertising landscape as a whole. By removing the need to run marketing campaigns in siloed channels, AI would be able to determine the best environment/format to serve an ad across all channels as we move to a greater digital reach in all areas of advertising. In saying that, there will still be a need to have a human touch to ensure the right goals are set and the delivery still follows the brand safety guidelines required by brands.
What are the next steps and possible future, in your opinion?
AI will evolve further in the direction of greater automation around campaign creation and delivery that not only incorporates programmatic, but automation of campaign planning and execution across all marketing channels.
I’m also excited to see how we all move forward as an industry when it comes to third-party cookie deprecation – we need to figure something out before 2022, that much is certain. We’ll see a rise in available formats across programmatic OOH and CTV opportunities, as well as how 5G will enhance IoT and VR experiences and how that can be incorporated into programmatic. The future is exciting!