Programmatic advertising is complicated. There’s no doubt about that.
This complexity explains why there is quite a lot of terminology involved, but it can seem quite opaque to the newbie.
Luckily, Econsultancy has a wonderful and thorough discussion and explanation of programmatic - Programmatic Marketing: Beyond RTB.
As a taster, I thought I’d throw some important terms into a glossary. It’s just the basics, but I hope it helps.
N.B. The list isn’t alphabetically ordered, but it should make more sense to the newbie when read top to bottom.
For a bit of fun, you can try Chango’s Chrome app, which highlights and defines programmatic jargon on a webpage, or quizzes you on the words involved.
The number of advertisements, or amount of ad space, a publisher has available to sell to an advertiser.
Simply put, data-driven media buying through a web-based interface. This is the automation of processes involved in buying and selling media, which used to be quite an admin heavy process of order forms and RFPs (requests for proposals).
Raju Malhotra, SVP, Products, Centro:
A media buy can be researched, negotiated, purchased, trafficked, optimised and billed all in one cloud-based interface that has a direct connection to the publisher’s ad server.
Programmatic RTB (real-time bidding)
Buying digital media through an ad exchange (platform) where inventory is secured on a bid for impression basis. Think of this as not dissimilar to Google AdWords and its auction model.
Traditionally, buyers compete for users in an open auction, and the highest bidder wins.
The ads can be targeted, often using behavioural data, giving rise to the idea of buying an audience rather than an ad impression.
Programmatic RTB has given advertisers increasingly easy access to large amounts of inventory (often at a low cost). It has also allowed publishers to sell long tail inventory (traditionally less-valued but now auctioned and reaching a known audience).
Larger publishers may have seen consequent decline in price of premium inventory decline as advertisers look to RTB.
First and third party data (usually behavioural)
Advertisers can apply first- or third-party data segments to find audiences across inventory. They then bid for impressions from this audience specifically (known through cookies).
Behavioural data is often used for targeting. This generally means retargeting existing customers or visitors and ‘lookalike audiences’.
DMP (data management platform)
A centralized system for managing the large amounts of data involved in programmatic buying. Used to manage first-party data, integrating it with third-party data and applying this to an advertising strategy.
SSP (supply-side platform)
Used by publishers to manage their yield, and adjust floor pricing for inventory based on dynamic data.
Publishers and networks add tags to their sites, which enable third-party networks and exchanges to value and sell their inventory to their buyers.
SSPs allow publishers to effectively monetise mid-premium and long-tail inventory for RTB.
DSP (demand side platform)
A technology platform allowing buyers (advertisers or agencies) to plan, target, execute, optimize, and analyze digital media plans.
Check out the Econsultancy DSP Buyer’s Guide.
A third-party company that licenses and supports DSP technology to act as a trading desk for buyers.
Agencies often have a trading desk and use DSP platforms to buy low and sell high.
Trading desks are building a programmatic technology stack that enables the agency to acquire all kinds of inventory, from online video to addressable television.
Agency trading desks add a layer of cost for the marketer but offer data and technology to meet needs at scale.
For Centro’s Raju Malhotra, the missing ingredient is “transparency and value”, which clients will increasingly demand. Some marketers are therefore beginning to buy programmatically themselves through RTB platforms.
Agency trading desks address a business need for marketers. As marketers get smarter and more comfortable with programmatic direct and ad exchange buying, the self-serve model becomes easier to use for broader adoption.