Cofounder, Codec
How can AI influence content marketing and creative decisions? Martin Adams
Martin Adams, CEO and co-founder of Codec speaks on how AI can influence content marketing and creative decisions at ADV Future of Media conference in Moscow.
Мартин Адамс, основатель Codec в рамках конференции АДВ Future of Media в Москве рассказывает от том как искусственный интеллект повлияет на контент-маркетинг и креатив.
About Martin Adams
In November 2017, Martin Adams, co-founder and CEO of Codec, spoke at the ADV Future of Media conference in Moscow about the role of artificial intelligence in content marketing and creative decisions. He described Codec as a subscription platform that provides intelligence to brands and agencies, telling them "exactly what content marketing their audiences want them to make before they have to guess, before they have to make it." Adams argued that traditional advertising is no longer sufficient and that brands must "think and act like a publisher." He stated that AI helps "put audiences first" by understanding what they want, allowing branded content to be "pulled in by what the audience wants rather than traditional advertising." Adams cited a case study involving Eve mattresses, where Codec's analysis of audience interactions with third-party content led to a creative rebrief. He said the company "doubled their revenue in 30 days and tripled brand awareness from 13 to 41." He emphasized that AI is not about replacing human creativity, stating, "This is not about taking oxygen out of the air for creative people — this is not 'content by numbers.'" He concluded that while search data is useful for e-commerce, understanding audiences through their content interactions is essential for brands aiming to be "part of the cultural lives of their audiences."
Transcript (26 segments)
M
Martin Adams0:10
[Applause]
[Music]
Hello, hello, hello. Thank you for your patience; I appreciate it's late in the day. My name is Martin Adams. I'm co-founder and CEO of Codec. So I want to talk about using artificial intelligence around creative decisions. That is exactly what we do. We have technology that is built for that purpose. We are a subscription platform. We deliver intelligence for brands and agencies to tell them exactly what content marketing their audiences want them to make before they have to guess, before they have to make it.
So we work with customers like this often, direct to brand, but also with agencies, also people that aren't on here like that.
So we've moved away from a world where you can just broadcast, and we've moved away from a world where just advertising is good enough. Audiences have basically forced our hands as brands to do better, to think and act like a publisher.
So what do I mean by that? On one level, I mean that they have to actually make stuff. They have to actually make long-form content, medium-form content, video, things like that, not just display ads. But on another, on a bigger level, it means that they need to stop thinking that the brand or the product is the most important thing. What is actually important is for the brand to ask: what is valuable to the target audiences that I have? They may be existing audiences or they may be target audiences.
Might have not very much; that is the honest answer. Ad tech doesn't really help, so they have to guess. When you guess, when you scatter-gun content out there across all the different platforms and all the different creative formats, one of two things happens. Number one: it just gets lost. It just gets lost among the two and a half billion other bits of new content that come out that day. It could be your friend's wedding photos, christening photos, Game of Thrones content, whatever it is. The brand is up against that and it's tough. That's if you're lucky. Actually, what is more likely to happen is this: we invest millions and billions into forcing audiences to watch the content we want to push. But then they build an ad blocker and they hate you for it.
I'm not going to dwell; we all know many, many examples. Dare I say, we might have been involved in many, many examples. This is Cortana from Microsoft. This is just a couple of pages of the 7,000 negative comments on Twitter after they did a piece of content marketing in 2017. It is not enough, I submit, to just reach your audience. You need to do more. You need to understand what they want. You need to understand what resonates with them before you guess, because your most valuable thing as a brand or as an agency, your most important task, is to safeguard the brand equity. That is the most important thing that you can build out.
So what does AI do and how does it fit into this? The audience wants more generally rather than traditional advertising.
So our technology plays into this in the following way. We have built out a range of machine learning and artificial intelligence algorithms that can understand third-party content. It was mentioned right at the outset of today that we need to be able to understand data and turn things into data if we're going to make them from data. So that is exactly what we do. We profile third-party image, text, and video content. We understand to an explicit and granular detail exactly what is in there: what are the color tones, what is the cropping, all of the things that you might want to know if you were going to make a creative decision.
And it allows us to do two things once you have that technology. So on the first level, we can understand content. We can see all of the clusters of people that have been interacting with content that is about a certain thing, whether it's yoga or ad tech or artificial intelligence. And we have a database of about 1,500 different cultural tribes, as we call them, that can be accessed. And then we also use those algorithms to say, and this is what else they care about, so that as a brand or an agency, you are now able to learn about what to make creatively that is outside of the scope of what you would normally AB test within. If you're a soap brand, you probably talk about soap a lot. The ability to learn about all of the other content that exists on the web—the two and a half billion bits of content—is very much reduced unless you can profile and understand that content to explicit detail.
What color schemes work with that audience? What content, what influences, who they should cast, etc. And then the results are things that we cherish and things that we love and things that we find entertaining. And that's exactly what we need to be doing as marketers now.
So let's bring it down to an example for one of our customers. We worked with a company called Eve Mattresses; they sell direct-to-consumer a mattress. And they came and they wanted to move beyond demographics. They wanted to target an audience defined by its mutual love of green smoothies and juices and things like that. So they plugged them in. And then from their interactions with a broad range of third-party content, we were able to deliver creative intelligence to the brand before they had to commit mistakes and spend money on media.
So we were able to see from clustering the thousands of others that we have in our system. We were able to see the creative formats, tones of voice, things like that. We were also able to see how content changed over time for this audience. So we were able to see that meditation content, content around spiritual awareness, emotional intelligence, was on the up and up month after month after month, because our system could classify content. It's not a social listening approach; it just genuinely classifies it based on what is in that content.
So the result was that Eve didn't commit a big mistake that it was going to commit. It was going to spend money on this type of content, very product-centric, very image-based. Actually, that was not the right answer. So they rebriefed their creative agency and produced content that played into the area within culture that their audiences cared about. The results were huge: they doubled their revenue in 30 days, tripled brand awareness from 13% to 41%. They IPO'd a year later.
So we can do this for lots of different clustering things around things like color tones, leading to huge content engagement uplift. We can do this to show how one persona or one target tribe might manifest itself through content differently in one local market versus another. So this is for Unilever to show them how one persona expresses itself through content and creative very, very differently.
So in that way, you can start to solve some of the kind of local and global dynamics that are very, very common these days.
It means: what content does this audience want? What is going to resonate with them? And it means that we have to be able to do that not just for our target audiences, our existing consumers, but also for any sort of new tribe or prospect or persona that we might identify. It means that we have to build for the right people. So we've seen why we have to make so much content these days.
So let's make intelligence and deliver the data for the people that make content decisions, rather than just for the analysts or the data scientists who then have to translate it over to the planners or the creatives. So we build for creatives, planners, budget holders, CMOs, etc.
You also need to provide it in a way that's useful. If you give that to a creative, they tell you to go away, basically.
And you need to deliver the right forms of intelligence. So we need to understand who these people are as a whole: personality analysis based on the text that they create themselves and that they interact with; visual clustering around types of image that are compelling and resonate with the audience; things like passion points and how they change over time, just like we saw earlier; things like influencers, so that we can distribute it once we know what content to make; and things like the content that is bubbling up with this audience right now.
And then finally, I think maybe just as a kind of macro point, we need to be building, and as an industry we need to be thinking about what type of data we're using. So we make a distinction between empathy data and the logos of our brands or the product shots, etc. Search data is amazing; it's absolutely amazing. It shows a specific intent of a person at a specific point. Amazing if you're trying to drive an e-commerce transaction with display. But what if you are trying to play in this new world? What if you are trying to be relevant? When meaningful brand studies show that 76% of brands are not relevant, what if you want to try and fight against that? Then, if that's probably why you're using content, you're trying to be part of the cultural lives and the everyday lives of your audiences, then you need something else. You need to be able to see the world through their eyes. And the best way to represent that culture better, I believe, than asking in surveys and things like that, is to look at the content they engage with. If you can do that, you will be in a position to make the right content for them.
And yeah, I guess the final thing to say is this is not about taking oxygen out of the air for creative people. This is not content by numbers. No machine can, or hopefully ever will, be able to tell you exactly what to do in every circumstance, and there you go, and your audience loves you, and you're part of culture, and off you go to the pub. No. It is about helping in a very, very, very difficult circumstance. There's never been more platforms. There has never been more creative formats. Each day, the platforms are telling us that we have to adopt the new VR, AR, IR, whatever you know, instant articles, etc. They're trying to push these on us. But when we shift our focus towards content, we can use artificial intelligence to really make the right creative. Thank you.
[Music]
[Applause]
[Music]