In this edition of Affiliate Spotlight, we speak to Chris Tradgett. Chris is the co-founder of Publisher Discovery, which develops machine learning and artificial intelligence (AI) technologies for affiliates. Within affiliate marketing, Publisher Discovery’s main focus is forex.
At the moment, Chris operates in a Chief Marketing Officer position at the company.
In this interview, Chris speaks about why AI and machine learning are important for affiliates. He also discusses possible barriers related to adopting this technology, plus more.
Affiliate Industry Review: Let’s start with a bit about your background. How did you first get into affiliate marketing and how has the industry changed since?
Chris Tradgett: “I first encountered ‘affiliate’ as one of the team that launched the buy.at affiliate network in 2002. It’s a pretty fast-moving industry and certainly in the early years, we were inventing the way it all worked – alongside other young networks, mainly TD, Awin and DGM – and jockeying for advantage.
“Since then, the industry has matured enormously. It’s far more technically-advanced and accountable today. The biggest changes have been around how advertisers can now interact with publishers – and indeed the nature of affiliate publishers themselves. Many are large corporations, several of which grew from an innovation – and M&As have meant consolidation into larger media groups.
“Advertisers can now interact commercially using multiple models. These include influencer placement, to retargeting as well as good old affiliate links still paying on a CPA.”
AI: Compared to if affiliates were sourced manually, how much higher is the quality of affiliate partners found through AI and machine learning?
CT: “If you’ve searched for affiliates in the usual long route using search, you’ll know how ‘hit & miss’ that can be. There are over 1.7 billion websites in existence. Even for a term like ‘forex comparison’, 12.5 million results appear in Google. With only Google’s rankings to rely on as a guide, dealing with this volume of results is impossible.
“Machine learning allows a lot more parameters to be applied, along with secondary data sources brought in to achieve tighter filtering. It also enables the machine to continue learning about the sites discovered and adding further layers of useful data to the platform.”
AI: As the co-founder of Publisher Discovery, can you tell us how the idea for this business initially rose and developed thereafter?
CT: “We first came up with the idea as a development of the Linkdex SEO platform, to dig into affiliate network tracking and find affiliates linking to networks. The first big test came when Google closed its affiliate network overnight in April 2013. All tracking literally stopped, going to a 404 on Google.com!
“We were able to provide the link data, so that other networks were able to quickly identify the affiliates using GAN. This meant that we could provide them with links to alternative merchants.
AI: In your opinion, what are the biggest barriers to affiliates using AI and machine learning? What can be done to help them?
CT: “For most smaller affiliates, using machine learning may appear daunting.
“There are off-the shelf-tools that affiliates can use to understand their own data sets more effectively, helping them to better-target their activity as a result.
“Any affiliates with programming skills should be able to take advantage of these using popular languages, like C++ and Python. I’ve no doubt that more and simpler, off-the-shelf solutions will be launched for many simple applications.”
AI: Other than what has already been mentioned, what else is unique about Publisher Discovery?
CT: “There are a couple of things. We combine multiple machine learning techniques within a huge data lake.
“The volume of data is also pretty huge. The simple interface gives access to potentially over two million affiliate websites, linking with over 330,000 advertisers. As we grow into new regions and uncover new tracking relationships, that volume is ever-growing.