When we think about the words Artificial Intelligence (AI), we’re drawn to the notion of Hollywood blockbusters or, you know, that episode of Black Mirror where Domhnall Gleeson’s character Ash, following his untimely death, is replaced by a synthetic AI version of himself that mimics his very being through the Internet and social media history trail he left during his life.
The reality is AI does not prevail purely in the realm of science fiction. In fact, we already unwittingly use more AI than we realise. A global consumer study by Pega titled “What consumers really think about AI” revealed only 33% of consumers believed they used technology with AI. The truth was that 77% were actually using an AI powered service or device.
Think about your everyday life. Are you using Siri, Alexa, or Google Now? Do you use Facebooks image recognition for tagging friends? Do you shop online through personalised recommendations? Do you use e-mail spam filters? If you’ve answered yes to any of these questions, you are already using some form of machine learning whether it is predictive analysis, facial recognition, voice recognition, or deep learning — all rudiments of the science of AI.
Now, think about how AI is being practically applied within the publishing industry. Behavioural analysis is supporting better decisions, smart content curation drives better audience engagement, propensity modeling allows us to understand how a consumer is likely to behave, ad targeting allows us to serve the right content at the right time to the right person, and dynamic pricing allows us to identify that subset of audiences who need that extra boost to convert to sale.
In terms of audience segmentation and targeting, machine learning has helped us come a long way by enabling us to draw correlations between consumer behaviour and engagement. It facilitates smarter campaign optimisation with improved targeting, in turn maximising return on investment (ROI).
At Independent News & Media (INM) our advertising team of experts uses first-party intent data for direct response, working closely with brands to build their ideal audience and target campaigns at scale across programmatic, native, and display.
First-party intent data is gold dust for marketers: It contains valuable signals that indicate the likelihood of a person to be in the market for specific products or service.
We recently conducted a six-week case study to demonstrate the positive effect machine learning has on campaign performance and in identifying that audience intent. The results were strong.
Working with a travel company as the client, INM leveraged machine learning to find the right campaign audience and increase engagement by identifying readers with traveling intent. This initial audience was monitored by an AI marketing solution to see how potential customers interacted with the advertisements they were being served. Machine learning was used to understand the characteristics of the readers most likely to engage.
We conducted an A/B test where readership was split randomly into two evenly sized groups. Each group was served the same campaign using an INM travel audience segment or the AI marketing engine via our in-house ad server. Over the lifecycle of the campaign, both audience creation approaches were optimised to uncover which achieved the highest engagement.
This A/B test proved the AI marketing engine created audience segments that were consistently:
- More dynamic (33 times more new audience members).
- Larger (five times as many members).
- More relevant (390% improvement in click-through rate).
Through the use of AI, we are now able to harness the power of intent data and identify audiences with real intent, maximising ROI for our clients.
However, it doesn’t end there. This is but one example of utilising AI solutions to identify audience and optimise performance.
We are currently in the midst of trialling IBM Watson with our native team, bringing cognitive computing to the native advertising landscape. IBM Watson is a supercomputer with extreme processing power. Combining AI and refined analytical software, it has the ability to interact with people and to learn and reason, interpreting Big Data at scale.
Our native editor Hugo McCafferty explains: “When a brand comes to us looking for suggestions for branded content titles, we can see what people are saying about that brand at a global level. We can look at their competitors and see what the conversation is around them. It means that even if it’s a one-off piece of branded content, there’s a coherent, researched context to it. It means every article has a content strategy behind it.”
The future is bright for AI development and the potential applications it may have within the publishing industry. AI is already reshaping business strategy across industries. No matter what stage of the adoption curve you’re at, one thing is clear: Artificial Intelligence is no longer something to merely consider in the far off, distant future. The future is now.