What is programmatic advertising?

“Programmatic” is a broad term used to refer to the automation of advertising. In an increasingly high-volume world of digital communications, the extent of advertising programmes is such that it makes sense to hand over manual tasks to artificial intelligence.

Programmatic certainly doesn’t replace digital marketing experts and optimisation teams, but actually frees up their time to focus on the mechanics of fine-tuning their advertising accounts.

Why is everyone talking about programmatic?

Programmatic is on the tongue of almost every brand manager, with marketing agencies responding accordingly. It’s at this point that specifics need to be defined: the digital world’s interest has been piqued by the real-time bidding element of programmatic, otherwise referred to as RTB.

In a nutshell, RTB is redefining display advertising, which sees brands promote their products and services across the Internet via rich media ads, such as banners, videos and widgets.

Traditionally, display advertising was managed manually across display networks, such as the Google Display Network (GDN), of which YouTube is a large partner. This is still a viable option for advertisers with small budgets and a manageable account, which can be administrated and optimised with a few checks per week.

However, when advertising scope is large, it makes sense to look into RTB, which evaluates each impression on each placement website as it becomes live. As such, there is no pre-bidding and the best cost per impression is guaranteed. To confuse matters slightly, display advertising is almost always evaluated on a cost per thousand impression (CPM) basis, owing to the minute price of each individual impression.

To make RTB work more effectively, advertisers can define their target placement sites (e.g. by topic) and their target audience (e.g. by demographic profiling). This ensures a degree of control to the automated exercise.

Demand-side platforms

Programmatic then re-enters the equation with software that is used to look at each specific placement if required, whilst analysing the RTB impression inventory to ensure that money is spent for on-brand placements with engaged audiences. This is called a demand side platform (DSP), which interrogates the exchange side platform (ESP).

DSPs work on an algorithmic basis, to predict impression share for each website based on historical data, and the day that the inventory is released, along with the time of day.


Crucially, DSPs have developed a form of tracking, which allows advertisers to see where views of an ad have contributed to a sale further down the line. This moves attribution modelling on significantly from archaic, last-click models, which previously made the error of deeming display advertising ROI-negative. The tracking typically employed is called a tracking pixel solution.


A final point to make is that investment in RTB via a DSP is usually pricey. Most DSPs won’t operate without a monthly budget of four figures. That said, if the exercise promotes revenue-driving brand awareness, the investment is incremental.


In light of the investment and the semi-automated nature of this new and increasingly important channel, your activity should always be backed up by an impression forecast as a minimum. Our recommendation would be to bolster this with a revenue forecast based on website conversion rate and basket value data.

Want more information? Contact us here

About the Author