Machine Learning in Digital Media: Control vs Conversions

By Georgina Wall | Sponsored

Georgina Wall, General Manager of Product and Partnerships at Resolution Digital

Machine learning. The buzz around this buzzword is going nowhere, with more and more platforms investing heavily into automated campaign types. The question is - is the promised increased scale and efficiency from machine learning worth the loss of control and insight to advertisers?

There is no doubt about it, the way we buy digital media has seen a dramatic shift over the past few years. Gone are the days of deeply segmented campaigns and having to tweak campaign settings at the drop of a hat. 

Most platforms now focus on consolidation and the element of a goal-based bidding approach. Set your campaign objective, throw in some creative, and away you go. Sure, this saves time spent over-analysing and labouring over a multitude of targeting settings. However, the control advertisers have over placement, target audience, and even display of their messaging has now massively decreased.

Less levers to pull has its benefits; media buyers now have more time during the day to invest into other projects. Plus, giving more control to machine learning allows for a more strategic lens of campaign performance.

How much scale is too much?

A key element of successful automation is the ability for the machine learning to continue to learn and adapt, enabling quicker and greater decisioning. As a result, many formats tend to achieve this through broadening the mix of inventory included within a campaign. This is nothing new – advertisers have long had the ability to scale campaign reach through partner networks and alike, however with an automated campaign type, this choice is taken away, and in most cases, so is the reporting granularity of where their ads have appeared, subsequently raising the brand safety alarm.

What about the impact to platforms?

Creating a black box of targeting has the propensity to further create divide and fragmentation within the digital landscape. The size of many walled garden platforms has the likelihood to cannibalise smaller players who can’t outperform their heavy investment into machine learning and automation. At Resolution Digital, we are seeing first-hand more traditional campaign types disappear, and the focus pivoting towards placing our trust solely on “trusting the machine”.

The method of targeting users has become far more predictive – taking everything we know about the campaign by way of the goal and the message context, and combining this with what know (or think we know) about the end user.

In addition to targeting, the method of measurement is changing. Use of modelling to report on campaign performance has been another area seeing increased prioritisation for platforms, as privacy changes in the digital ecosystem begin to take effect. This in turn makes it harder to identify users and link their behaviours online to a certain campaign tactic.

What does this mean for personalisation?

Dynamic ad formats lean on this predictive nature of automation to achieve the outcome for the campaign. Whilst each different platform will have varying amounts and levels of detail around their audience make-up, their goal remains consistent; produce the most relevant user experience to drive an outcome.

Personalising one-to-one messaging experiences does not fit within this game plan however, forcing advertisers to look at the way they talk to prospective audiences, and instead align better with the changing tech. For example, using contextually relevant assets that talk to a group of users to maximise conversion viability.

In summary

Humans do not like change – and especially when we are up against a machine! As with all shifts in technology, a period of adaption is required for the masses to be content with change, and this is no different. Like the machine, we are learning and adapting to a new way of working in the digital environment, and it has potential to be a great change.

Reduced control is confronting, however to drive change, we must embrace it. The convenience of consolidation, and the amount of intelligence available from machine learning allows for bigger picture thinking for advertisers, and the potential to scale marketing campaigns to new heights.

Have something to say on this? Share your views in the comments section below. Or if you have a news story or tip-off, drop us a line at

Sign up to the AdNews newsletter, like us on Facebook or follow us on Twitter for breaking stories and campaigns throughout the day.

comments powered by Disqus