Generative AI is all the rage, from the excitement around the tech’s potential to accelerate almost every industry, to unease over how it might upend nearly every industry — but not in a good way.
On one side generative AI can help a publisher boost SEO rankings by producing SEO-optimized content. On the other, with the technology being baked into search engines, publishers fear it will cut their search-referral traffic.
When it comes to the advertising ecosystem, generative AI is mainly being used on the creative side of the pipeline to generate creative assets at scale and remix and deliver it relative to consumer behavior and interests. For example, using a site visitor’s behavior to rejigger messaging or colors that appear on creative exclusively for them. Both Meta and Google have plans to use the tech in their ads business to pump up ad revenue, and Snap showcased AI-based ads in its My AI chatbot at its NewFronts.
Now let’s take it back to the sell side of the equation for a moment. What if an ad tech company could figure out a way to enhance their offerings with generative AI to provide publishers with incremental revenue? Well, that’s exactly what Dan Rua, CEO, Admiral, the Visitor Relationship Management (VRM) Company, told me his company is offering publishers through their platform.
He and I discussed how Admiral is implementing GPT to use personalized optimization to generate value exchange messages for publishers to serve their visitors and the importance of using intelligent messaging to drive conversions while building trusted relationships with them. We also talked about privacy and ethical concerns when it comes to generative AI.
What is Visitor Relationship Management?
Lynne d Johnson: First, let’s get the basics out of the way. Can you remind people what VRM is again?
Dan Rua: We were publishers before with a company called Groveshark, which was in the music streaming space. Then we started looking at how to get around ad blockers and drive toward user empowerment and privacy. The existing ad tech industry wasn’t built for that. Ad tech is more about tracking people and shooting ads. And this new way means publishers have to build relationships with their visitors. And that required a new kind of platform, which was less of an ad tech platform and more of a martech or marketing automation platform.
At its heart, that’s what Visitor Relationship Management is. It’s like marketing automation for media publishers. And we help publishers manage the entire journey with their visitors, whether it’s talking about privacy consent or turning an ad blocker off or email capture or registration or donations or subscriptions or social follows or mobile app downloads, or first-party data capture.
Every one of those things is part of the journey. But if you try to cobble it together like a Frankenstein with five different vendors, to pull off those relationships, you’ll shoot yourself in the foot. You benefit from an intelligent journey system that can talk to the visitor about the right thing at the right time and talk to them in the right way based on who they are and where they’re coming from.
GPT Meet VRM
LdJ: And now you guys are making VRM even more personalized lives with the integration of OpenAI’s GPT. How will that work for publishers and help them build more trust with their visitors?
DR: When you think about an intelligent journey system, any one of those asks, whether it’s about adblocking or subscribing, is a value exchange at heart. So first, last year, we implemented AI around when publishers have the most value to exchange — because you don’t always have the same amount of value exchange — any given article at a certain point in time may be more value to exchange.
A good example of this is what we call surge conversions, where the journey can pick up on traffic surges for an article, maybe it’s going viral. That’s an opportunity to ask for a value exchange because you have more value to exchange at that time. You want to realize that in real time and take advantage of that moment. Publishers using that AI system, gain 40% more recovery monthly because a single article spiked and they were able to make the right ask at the right time, to either turn blockers off or get them to subscribe.
And then when GPT rolled in we launched the ability to deploy value exchange messages on the fly. First, we created push button messaging that takes the content of your site and then asks: “What’s the offer?” “Is it a subscription?” “Is it email?” “Is it adblock?”
It will take the length you want and the sentiment you want, and then after choosing a couple of choices, push the button, and it generates a message on the fly, which is a really solid message. So, now you’re generating that content in a quick and relevant way.
Then, we wanted to drive revenue for publishers, so we layered on that second piece, that automatically generates A/B testing of messages, running them against one another and then getting rid of the worst performer. Then it auto-generates another message and runs it against the other one, and kicks out the worst performer again, so it’s iterative. It’s A/B testing on the fly, creating variance, and auto-optimizing revenue generation for the publisher. The system auto-optimizes itself.
Journey Intelligence, Your Tool For Combatting Consent Fatigue
LdJ: One of the things I ask everyone in the consent business is what about consent fatigue? It’s a real thing. As people are surfing around the web, they’re seeing popups everywhere. Then, when people have too many choices, they just sort of zone out, which would equate with opting out in this case. What are your thoughts about these challenges?
DR: Even before all of this AI stuff, with journey intelligence, you’re not going to get away from having to talk to visitors and build a relationship with them. This is about user empowerment and giving users control over the environment. You’re not going to be able to hide behind your site and shoot ads at people and it is okay. Publishers who separate themselves from the rest are going to have to talk to visitors. And then, how do you drive the most conversions with the least number of interactions? How can you have an intelligent system to drive the most results with the least necessary conversations?
That’s how you get this idea of propensity to convert and when do you have the most value to exchange? You need intelligent messaging, because the goal isn’t to hit people over the head with the same message every single time, you want to do it intelligently so that you engage them the least number of times necessary to drive the relationship that you’d like to drive.
Not to Sound Cliché, but Right Person, Right Message, Right Place, Right Time
LdJ: So with consent, you’re saying the same thing we’ve said about ads for years, that hasn’t always worked — right message, at the right time to the right person in the right place?
DR: Dead on, It’s the same concept. How do you make it relevant to someone? How do you get them at the right time? It’s like that surge example I mentioned earlier. You know that one article out of the thousands on your site will achieve conversions for you and you don’t need to hit people over the head, or across all of your articles. It’s a time when the person has a greater appreciation for your content and it becomes the right time to have that conversation.
But we can even talk about going further in the direction of personalized ads with this. So you’re not just A/B testing messaging, you can change your CTA text or change your CTA colors.
And going back to messages, let’s talk about the level of granularity you can get. You have people coming from Facebook versus coming from Google Search, and you can do the same kind of tests for those referral paths. Or VIPS versus non-VIPs, or brand lovers versus casual. When you start to do that, you have granular personalized optimization based on where they are coming from and what content they consumed previously combined with anything else you might know about them.
Changing the First-party Data Capture Game
LdJ: Sounds like a game changer for first-party data as well.
DR: First-party data is another thing that drove us in this direction. When you get into declared first-party data capture, you’re getting into an infinite number of conversions, right? Like Ford vs Chevy? There are so many opportunities to grow the relationship with visitors that you will want and need a really intelligent system to figure out when and how to talk to people about that.
If you want to know a thousand things about someone, you’re going to need a really intelligent system that builds a conversation with something like a frequency cap built-in in a way that you’re not spooking people out. You might want someone to say yes to privacy consent, and you might want them to turn off their ad blocker and you might want them to subscribe, you can’t talk about all of that stuff today. You have to figure out how to pace that over time.
You don’t have to get married on the first date? You go to dinner and then maybe you go off for a weekend, you need systems that can pull that off with visitor relationship management.
AI Regulation Is Inevitable
LdJ: I’m sure you heard about Italy banning chat GPT over privacy concerns. The EU has this Artificial Intelligence Act. The US is talking about regulating AI in general. Does AI need to be shut down or regulated? Do we need standardization around all of this stuff?
DR: I think we’re just in the infancy days of this whole thing. I think regulation is inevitable. I think lawsuits are inevitable. What does the future hold if the percentage of zero-click searches becomes the majority? Looking at the degree that these engines are learning off of publishers’ content, then you’re going to find yourself in licensing conversations.
I tend to see at least two different aspects of what’s going on with GPT and the Italy example that you used. There’s what the AI can deliver to you for the questions that you ask — the personalization that might be delivered. And then separately, there’s the information being provided to the AI. So it becomes, what’s happening with the data that’s being handed to the AI?
There was a case where Samsung was trying to use AI to better inform their business and they were loading confidential documents in the process. So there’s this question of feeding content and data to an AI and the implications. On the flip side, there’s the output of AI, which has a brighter opportunity while the world tries to figure out how we handle what data is going into the AI to fuel it.
How Else Can Generative AI Generate More Revenue For Pubs?
LdJ: What else do you think generative AI could make possible for pubs, especially when it comes to future-proofing their business?
DR: Readers come to publishers for their content or for their knowledge, which are two different things. There are the individual articles that are written, but then there’s the knowledge that is infused in those articles that readers are getting. There will be publishers that are differentiated by the knowledge that they bring to the world and then there will be publishers that are differentiated by the presentation of knowledge that they deliver.
Every publisher should have an AI brain or an AI agent that has leveraged the knowledge of that publisher and that knowledge is some combination of all of their content. Maybe there’s even some unpublished knowledge that could be part of that agent and their readers could absolutely derive great benefit from that. Imagine reading an article and then the reader could ask, “Whatever happened to so-and-so in that article?” And the system could deliver the rest of the story to you. It can be a knowledge resource that visitors are tapping from you as a publisher. That would give the publisher the ability to keep people on the page.
I am really energized by generative AI that generates revenue. And I get a little nervous that so many use cases are on cost savings. And that immediately runs in the direction of jobs at risk. Publishers themselves, and vendors, need to keep working on ways that generative AI generates revenue, and that can fund jobs.