AI is a buzzword that we’ve heard about a lot and rightly so. It has infiltrated all facets of modern business and has the power to be transformational. According to Joanna O’Connell however, buzzwords without real understanding can be problematic, if not downright dangerous.

This article originally appeared in The Madtech Brief, Marketing‘s second print issue for 2019.

Evolving consumer behaviours – multi-device usage, decreased brand loyalty – mean that brands and their marketing teams need to be more agile to succeed. Artificial intelligence (AI) and machine learning are powerful tools that can make everything, from ad creative to media buying and planning, more effective.

Today, AI has infiltrated:

• Media planning

The media planning process has traditionally been highly manual. Marketers can now use AI in the form of lookalike modelling to determine who and where their most valuable prospects are and make budget allocation decisions accordingly. 

• Media buying and optimisation

Programmatic platforms have been running decisioning logic for years to power automated digital media buying. Now machine learning algorithms, the most common application of AI in advertising, learn to predict the probability of media buy outcomes based on past user behaviour, market dynamics and marketer historical data, along with target KPIs to make real-time buying and pricing decisions across billions of available ad impressions.
This is something a human simply could not do.

• Media performance measurement

The speed of media buying and optimisation that’s become necessary to keep up with modern consumer demands means marketers must move faster from buying to measurable results to thoughtful action. Forms of AI are increasingly penetrating advertising measurement, including deeper pattern recognition to untangle complex customer behaviour and faster insights that are increasingly capable of pushing recommendations directly into media buying platforms.

• Creative development, optimisation and insights

Ad creative is demonstrably critical in driving consumer behaviour. But the traditional creative process, which is highly manual, unidirectional and lengthy, doesn’t match the speed or nuance needed in today’s real-time, audience-driven advertising world. A new crop of creative advertising technologies has emerged to make the process more streamlined – while also vastly increasing creative output – to better match ad creative to a multiplicity of target audiences and/or media scenarios. The leading platforms in this sector offer a variety of AI-powered features, such as computer vision to analyse creative objects; pattern recognition to understand buyer behaviour, machine learning to optimise products, features or creatives shown; and natural language generation to create textual content.

 

In an era of programmatic advertising, AI brings about obvious benefits, but it can also mean significant risk. While the machines’ capabilities continue to evolve, marketers’ understanding of their role in advertising is not even close to keeping pace. These risks become amplified when AI-fuelled machines run without human understanding and oversight. 

Marketers risk:

• Wasted ad dollars due to fraud

Marketers in the US lost a whopping US$7.4 billion in ad spend to invalid impressions in 2016, and that number will climb to nearly $11 billion by 2021. And concerningly, if not surprisingly, fraudsters are masters at outsmarting industry efforts to thwart them. The automation at the heart of programmatic advertising has opened the door wide to fraudsters looking to dupe ad fraud detection algorithms when human eyes aren’t monitoring the machines’ work.

• Harmed brand reputation due to inappropriate ad adjacency

When human oversight of ad placement gives way entirely to automated ad platform-driven decisioning, the results can harm, rather than help brands’ advertising efforts. The news is rife with examples of brands’ ads running adjacent to inappropriate content, ranging from the embarrassing to the downright scary.

• Marred consumer experience due to unethical or otherwise harmful targeting

Ad targeting is designed to help a marketer find and deliver relevant, useful communications to any intended target consumers. But without thoughtful human oversight, the results can have the opposite effect.

With these benefits and risks in mind, marketers must ask questions that will give them the necessary confidence in their adtech partners’ approach to and applications of AI and develop the right interaction model between human and machine to maximise every dollar spent on advertising. They should also apply the 80/20 rule when preparing for an increasingly AI-driven ad future: build a solid foundation of performant advertising with the bulk of your spend, but make room for ongoing experimentation.

Joanna O’Connell is VP and principal analyst at Forrester

Forrester is a Marketing Content Partner – a leading organisation with which we collaborate to bring exclusive content to readers.

 

* * * * *

To purchase a subscription to Marketing magazine, visit the online shop »

* * * * *

 

Further Reading:

• What will it really take for Facebook to fall? »
• The consumer tech stack – Forrester analysts breakdown the Australian tech drive »

 

 

 

Image credit:Christopher Burns