Flavour of the month or a genuine innovation shift, the term artificial intelligence (AI) is wont to divide the opinions of performance marketers. Many see it as a disruptive technology with promising land to be grabbed, others call it a buzzword and a relabelling of an existing toolset.
The debate isn’t set to dampen at PI LIVE next week, with no few sessions levelling their sights on the state of the technology and its potential to process and act on vast amounts of data at the blink of an eye.
Tackling this discussion head-on is Webgains’ CEO, Richard Dennys, in a session entitled ‘Can AI Give Performance Marketing the Revolution it’s Waiting for?’. PerformanceIN cornered Dennys to get an early take on some of our most pressing questions.
First of all, artificial intelligence; how can it really benefit the performance marketing industry, compared to the technology we have had available up until its advent?
Richard Dennys: Performance marketers have long analysed data and online activity to pick out patterns or trends that provide opportunities to engage with customers – AI massively reduces the time needed to identify these trends.
Using AI in performance marketing will help businesses make decisions based on crunching a huge amount of data that would take a human a serious amount of time to get through. This is the real benefit.
For example, at Webgains we deal with 250,000+ publishers and a billion clicks per year. Now you might be Alan Turing incarnate, but you still couldn’t read and analyse all of that contextual data. But give this data stream to an AI platform and it will crunch through it to identify trends and make future predictions of purchasing behaviour based on previous activity or browsing history.
This means that performance marketers can now get in front of customers, not just with the offers and products customers want, but at a time when they are most likely to want them.
Do companies need to fundamentally change in order to facilitate AI technology, or does it simply supplement existing work?
RD: No, but they may need to adapt how they work. Where AI can make a real difference, and offer exciting possibilities for performance marketers, is in delivering ‘actionable insight’ much quicker. i.e. giving performance marketers access to campaign performance in real time.
This allows performance marketers to adapt campaigns quicker and boost results.
Real-time insight is essential for performance marketing, however, businesses will not go from human data analysis to full AI implementation overnight. Current use of AI revolves around improving the data analysis processes that already exist. For the moment, I see AI working with existing platforms and processes, as opposed to working autonomously.
Are there any examples of AI or machine-learning in action now that are just not so widely recognised?
RD: You’d be forgiven to think that businesses are releasing the latest and greatest AI tools every day – especially in the consumer tech sector where the big players like Facebook, Google, Apple, Samsung etc are dedicating their huge resources to AI.
However, it is the smaller, start-up businesses where I think there is real progress. Those businesses that don’t have AI as an added feature, but rather baked into their original design concept. For example, I think WordSmith is an interesting AI platform - last year it created more than 1.5 billion pieces of content. It’s these original concept designs and uses that are going to be really exciting as the technology matures.
What is Webgains doing with AI technology, and is it an area you’re planning to develop further?
RD: Earlier in 2017 we announced our partnership with IBM Watson to create an AI solution for the performance marketing channel. This made Webgains the first in the market to offer an AI solution that moves away from post-user behaviour analysis, and towards intelligent product upselling.
The development of this AI solution will see an increased focus on intelligent ad placement - processing both structured and unstructured data, including popular trends, web queries, user data and social media insight. Our mission is to give brands and retailers new insights about their ad spend – and how they can improve their ROI.
We aim to use AI to solve real-life challenges and uncover actionable insights hidden in billions of data points that are close to impossible to identify by humans at scale. Tapping into unstructured data will provide a better understanding of the why and how decisions are made when consumers make a purchase. This moves performance marketing from using basic verticals of personalisation to a hyper-personalised campaign that considers the weather, time of day, what was in the news, what was happening on social media before presenting you with the correct offer.
For example, the travel industry is seasonal - Spring, Summer, Autumn and Winter - and marketing around these seasons is at a basic level e.g. summer breaks in the school holidays and or short winter breaks in the Autumn half term.
What if, at the moment a click is made on a campaign advert you take the weather data based upon a location associated with that click at the exact time? That gives you another source of data and opens up a universe of questions to build cognitive models around. Are your customers more likely to buy that Maldives holiday on a cloudy day in January, or a rainy weekend at the end of the month in the summer?
If they are more likely to buy that holiday on a rainy day in the summer, wouldn’t advertisers want a predictive tool that would suggest this data so that they can quickly create campaigns that go to the newly found targeted publishers, and leverage on those sites to boost those conversions? Of course, they would.
Our mission is to use AI to make affiliate marketing at Webgains intuitive, predictive and proactive via cognitive technology and strategic services - moving away from the “work to live” or “live to work” paradox, and towards an equal balance of both. The results are already transforming our business from a ‘What now?’ reactive approach to a ‘Now what!?’ proactive approach based on intelligent insight.
How would you respond to concerns that increased automation could result in the loss of certain jobs in performance marketing?
RD: I think there is a fundamental misunderstanding about this and we need to be careful not to give into misconceptions founded in Science Fiction. The introduction of AI into a suite of technology isn't about making jobs disappear - just the opposite in fact.
By reducing laborious admin tasks, AI should enable employees to be more efficient and make more intelligent decisions. While Sci-Fi movies interpretation of AI will have many of us believe that the next Skynet is just around the corner, AI is essentially still a piece of software and like any software needs to be programmed, monitored and optimised by humans on an ongoing basis.
Automation and machine learning is certainly an exciting element of performance marketing, but in isolation, it won’t be effective. To execute the insight provided by AI, human interaction is still needed and businesses will still require human operators to intervene if things go wrong.
Could this technology also ultimately cheapen the work of its operators and bring down profit margins?
RD: AI will remove a lot of the manual work from performance marketing and make the process of collecting and analysing vast amounts of data easier, which will lead to account managers being able to create more personalised campaigns.
As a supplementary tool, AI can improve efficiency and productivity, providing the all-important actionable information sooner, while reducing the amount of time, effort and money wasted.
Used correctly, the end results should increase productivity, employee happiness, business revenue, and produce a reduction in costs. What business would say no to that?