How AI-powered Advertising and Marketing is Transforming Campaigns? Insights from Jason Widup, SVP of Marketing, Pixis | Podcast Ep. 22

And so the previous version of our brand before I got here was very focused on the technology. We’re a company led by engineers and so it kind of cascaded down from there. It was too technical though. It really alienated the marketers. It went over their head. We talked about codeless AI infrastructure and no marketer knows what that means.And so our new brand that we just launched about two months ago really focuses on our ICP and their challenges and then leverages things like social proof and value props to get them leaning into our product. And so yeah, I’d say that’s really how we keep it simple. We don’t really get into the technology too much, but more about the problems we solve, how we solve them, but from a pretty like 10,000 foot level.

And then we really rely on social proof like what we’ve done for other big name customers that are like them to help pull it across the line.

Host: That’s really nice, keeping AI practical and impactful. So, speaking of impact, Pixis has over 200 + AI models solving marketing challenges. Can you share a real-world example of how your creative AI or targeting AI transformed a campaign’s performance?

Jason: Yeah, sure. I mean, there’s lots of examples, but I think one really sticks out to me. I’ve talked to a lot of customers and it’s really around our targeting AI that stands out.

And our targeting AI works by seeing all the campaigns that you’ve run. And we actually get audience data from Meta and Google that a normal advertiser wouldn’t receive. So we actually have a little bit richer data about audiences that are engaging with ads.

And so as companies are running all these ads across all these different campaigns, there’s some audience segments that trickle in that maybe weren’t originally targeted, but we see them and we see their affinities to other things like what things do they like? What do they not like? What do they do?

What do they not do? And our targeting can uncover some of these hidden, we call them like hidden audiences. And so one example was one of our customers, they sell healthy ramen noodles direct to consumers.

So you buy them online and they ship them directly to you. And of course you would think I’m going to target foodies and people that like food and people that talk about food, all this food stuff, maybe some families. But what you wouldn’t necessarily target are things like people who enjoy travel or people who are an avid hiker.

Well, these segments are targeting AI where it was able to uncover and say, we’re seeing a strong affinity to these groups, even though it’s not the ones you’ve originally targeted. And so what that does is it often will unlock and open up new efficiencies because these are audiences that just haven’t been targeted before, but they convert at a higher than average rate. And so that’s probably one that’s just pretty ubiquitous across a lot of our clients.

That’s some of the value that the targeting AI piece does. It just uncovers these new audience segments that as a normal, I guess, marketer, you wouldn’t necessarily think would have an affinity to your product. And I think that’s really an unlock for any brand because it’s, you know, your best ICP, but there’s a bunch of other people that might not exactly fit that, that could really have a strong affinity to your product.

And you really want to uncover those. And oftentimes it doesn’t make sense when you look at it, but the data, you know, the data doesn’t lie.

Host: Yep. Absolutely amazing. It’s great to know how AI can turn real data into real results. So, you have grown companies from 1 million to 16 million ARR and led teams at giants like Microsoft and Tableau. What is your playbook for scaling marketing in such diverse environments?

Jason: Yes. And I do have a playbook that I’ve refined over these many years. I’ve been in B2B marketing, I think 23 years, something like that.

I look at it in kind of three, I call it a three-legged stool, maybe. The first is being the cool kids. And what I mean by that is like, it’s a really strong focus on brand because it’s human nature to want to gravitate towards people that are doing well, right? Human nature. It’s we can’t oppose that. It’s just a normal part of being a human.

It’s like, I want to be with the herd, you know, the herd that’s doing well. So we have this herd mentality. And it’s the same in business. So, people will gravitate towards the companies that even just look like they’re doing well or look like somebody they want to aspire to be. So brand is the first leg. And one of the ways that I do that is, especially when you’re a marketing technology company selling to marketers, your marketing has to be incredibly good because there’s a high bar, right?

You’re selling to a marketer who’s doing the same thing that you are. And if they can’t look up to you in some way and even look up to your marketing, I can’t imagine selling a really solid ad tech tool like we have and our marketing being really horrible. Because people will, despite even the tool could be really, really, really good.

But if you don’t do a good job at creating that heat and that buzz and drawing people in, they’ll just ignore us. And so that’s the first one is we’ve got to develop a really strong brand. The second one is a laser focus on differentiation and proof.

And so taking a look at our product and really understanding what are the, let’s say we have 20 competitors. What are the true, true differentiators? Why are we even here? You know, it’s like, why are we here? If we’re doing the same thing as everybody else, we shouldn’t even be here. We should just combine forces and create a bigger company and then go to market.

Well, we are our own company. So what makes us vastly different than everybody else? And let’s laser focus on those and blow those out as reasons to believe. And then the other part is kind of in the same vein is proof. So I mentioned that like social proof piece, having existing customers sing your praises, showcasing the performance benefit that they’ve received, show showcasing the amount of time that they’ve saved all of these things and going to market with them with their brand. That kind of goes along with the first one, which being the cool kids.

Because if I’m selling if one of my customers is a very cool brand that you also look up to, that’s going to tell you something like, oh, maybe they are. Maybe they are for me as well. So that’s the second stool is differentiation and proof.

And then the third stool is really around tech operations and execution. So you can have the best ideas ever, but if you can’t execute on those things using a very streamlined and efficient technology stack, it doesn’t matter. And so so I have a maniacal focus on execution and will always out execute other marketing organizations.

And so those three together is kind of my general marketing approach. And then, of course, every place I go, there’s all these uniquenesses. But that’s where I start on those three areas and making sure those are really working well. Yeah.

Host: Yeah. So human centric and tech driven marketing approach is really amazing. So next up, Pixis champions ‘stay human’ as a core value. How do you ensure AI enhances rather than replaces the creativity of your marketing tea

Jason: Yeah, this is a good one. Marketing or just leaders in general are not sure how they should think about AI. And right now, everybody’s using AI. So that’s there’s no doubt everyone’s using it. But it’s in like this level one maturity stage where people are just using an LLM like ChatGPT BT or Claude, and they’re using it to just ask it questions, get some brains, you know, brainstorm with it, have it help me write some content. It’s really like a level one maturity use case that everybody but everybody’s doing it.

And so when there’s that use case, it’s not really about replacing people. It’s not about reducing my headcount. If I’m at level one use of AI, that’s really just going to be I’m augmenting my staff with a little bit of help.

And so a lot of people are worried about what is a level two, three, four, five, let’s say there’s gonna be five levels of AI maturity. What are those levels look and at what point does it start to get close to replacing people? Your question was more about replaces the creativity of marketing teams, which that will never happen.

So the creativity piece will always have to come from humans. Because if you think about AI, AI can really only tap into things that have already been done before. So it’s referencing the internet, you know, and other things and documents, etc, that have all been done before.

AI is not really thinking in creative ways about how to put how to take all these things and create something brand new that’s never been done before, or a creative and AI doesn’t understand human emotion. And most of the creativity that we do in marketing is intended to draw on human emotion. I’m either trying to create, I’m trying to be funny, I’m trying to be informative, I’m trying to anger you sometimes, sometimes I’m trying to make you feel sentimental, you know, I’m trying in marketing to elicit emotions and change your thinking and behavior about things.

And that’s not what AI is good at. So the way we think about it at Pixis is, and in fact, it actually comes through in our logo. So the X in our logo is two semi circles. And what it’s supposed to represent is AI on the bottom, supporting humans, which are at the top. And so that’s how we really think about it is AI from a creative standpoint is really here to augment your thinking and get you there faster, but it won’t replace you. That said, there will be roles where AI will replace things.

But when we think about enhancing creativity, it’s going to be more of a, and I hate this word, but sidekick, you know, it’s going to be like, but also like a, an advertising strategist in your front pocket that you can ask questions of, and that has all this vast knowledge so that you don’t have to have that vast knowledge. Now you can focus on, all right, have AI distill all this for me, go back and forth with it. I’ll chat with AI, get some ideas, but then I’m going to pull all those together in a unique way that only a human brain can really piece together.

And then I’m going to, you know, basically move that forward. That’s kind of like, it’s not that far apart from, and I’m going to get deep on you here, but like in the sixties when LSD and acid was a big like drug that people were taking, well, guess what? That unlocked a lot of creativity in the human brain because the drug created connections that weren’t there before.

That’s the kind of, AI will never do that. AI hallucinates today, which I guess you could say is kind of like taking acid, but it hallucinates in a different, you know, and not a creative way. It hallucinates and just like, I got it completely wrong way.

So yeah, I kind of find it interesting to think about it, but it’s going to take humans to continue to advance creativity and understand taste, you know, and AI will not understand taste anytime soon. But it will be interesting to see how close that it gets, you know, as the years come go by. But, but yeah, that’s how we think about AI supporting humans in creativity and not replacing.

Host: I totally agree. Balancing between creativity and tech is important in marketing, at least using a tech, tech tool as  you know, support system is more of a beneficial approach than planning for a replacement. So next up with the data privacy concern rising, how does Pixis balance hyper-personalized ads with user trust, especially for Fortune 2000 clients?

Jason: Yeah. Yeah. And we do have a lot of Fortune 2000 clients. And so we do have to think about this. But on our end, Pixis doesn’t, isn’t enabling hyper-personalized ads. The ad platforms have already done that.

So the ad platforms, Google, Meta, TikTok, they have enabled this hyper-personalized advertising capability. That’s not something Pixis does, but what we do, there’s still things and protections that we need to have in place to make sure we’re protecting our clients’ data. And the way we do that is a couple of different ways.

First, obviously we’re, we have AI models that are being trained. We’ve trained that on advertising data in historical advertising data, like over two and a half billion dollars of ads. Then we’ve trained on close to three, but our current customers, we don’t, we don’t pipe their data back into the platform to continue to learn.

So, or we give them that ability to opt out. So you don’t want your data used in training our general models. No problem. We’ll keep it out of there. That’s one way. Another way is through creative, sorry, brand guidelines.

So our generative AI product that, called AdRoom, that builds ad creative using a prompt interface, it has an entire module around brand protection. And that’s really you uploading all of your brand guidelines, not just the colors and fonts, but even your tone and anything else you have about your brand that you want it to adhere to. And it will not go past those guardrails.

And so that’s where, because a lot of people are worried about, you know, one of the strongest use cases for generative AI is ad creative iteration, right? So like the, your creative director creates a couple of versions and then you put that into AI and it creates thousands of iterations on that. Well, people don’t want to approve every one of these thousands of ads.

So they need the, they need the confidence that it’s going to adhere to the brand guidelines and it’s not going to do something that’s going to embarrass us or get us in trouble or, you know, go against those guidelines. So that’s one of the ways we do it. And then of course, we’ve got the normal like SOC 2 compliance and ISO compliance that we make sure we adhere to.

And then the final thing is we don’t deal in personally identifiable information. So we’re not passing email addresses, names, titles, anything between our system and our customer systems or the ad platforms. We’re completely agnostic of PII.

And so those ways are, those are the ways that we’re keeping like data privacy concerns at the, you know, at the forefront. But again, we don’t have to think necessarily about how personalized these ads are because that’s enabled by the ad platforms. One thing though on that is that’s interesting to think about is in the not too distant future, ad creative will, with the ad platforms will be more about giving it components of the creative, but then every single different, every single person will see a different ad, not just like I have this ad and it’s going to, you know, 500 people.

And that’s fairly, that’s very personalized out of my TAM of maybe like 10,000, that’s not going to be the same. And you won’t even know what creative they saw because it’s going to be put together on the fly for that person based on their behaviors, based on their demographics, you know, all these things that the ad platform will know about you. And so that’ll just be an interesting change that we’ll see that you might not even know what creative they saw and you didn’t even develop it.

It’s just been pieced together by AI on the fly based on who you’re targeting. And so yeah, and that’s not in the too distant future.

Host: It’s great to hear how Pixis is prioritizing user privacy while considering personalization as well. So, you have spoken about generative AI’s potential. What is one underrated AI trend in advertising that will explode in the next two years?

Jason: Yeah, it’s something that we actually just started to develop about six months ago. So there’s a new underlying AI technology called MCP or Model Context Protocol. It’s very complex and heady, but the basic way to understand it is that it’s basically like a smart switchboard operator for AI models.

And so the first thing it does is it figures out what kind of situation you’re in. So it’ll understand like, is this a legal conversation? Is this an advertising conversation? A medical question? Is it a casual chat? Then based on what it’s understanding of the situation you’re in, it’ll choose the best model for that context.

So you might have hundreds or thousands of models and it’ll intrinsically know which of these to choose. You don’t have to tell it ahead of time. It’ll just know which one of those to choose.

Then it’ll take the data and prepare it in the right way to give to that model. So it’s kind of a transformation layer. Some people also call MCP like a very smart API.

So it’s going to transform that data in the format that that model needs, and it’s going to hand it to that model. And then when it gets a return from that model, it’s then going to take that and it’ll return the most relevant, accurate information back to the user adjusted for their situation. And it does all of this automatically in real time.

So I can’t even understand it. My brain can’t get to understanding it because it’s like, well, how does an AI figure out the context of an entirely new data source? And so we’re building a new platform on top of this we call Prism.

And in just weeks, we’ve been able to advance this thing so quickly because it has this intrinsic understanding of these other models. And so for example, when we want to, so one of the biggest challenges in performance marketing is the amount of data that I have access to, to analyze, to understand what’s going on and to then make changes. Most of that data is limited to what’s available in the ad platforms, right?

I’m getting my spend, impressions, clicks, conversions, et cetera. But what it doesn’t have access to is things like external data sources. So one of our customers, Helly Hansen, who’s this outdoor wear brand, they wanted to know, do sales of their rain jackets go up in cities where it’s raining? Well, guess what? Raining that day. You know what I mean?

Not just like it generally rains in Seattle, which is where I’m from, which is common, but no, this day it rained for these hours. Did our sales of rain jackets go up during that time? That’s not something that anybody could answer with an ad platform before.

And to answer it, they’d have to find a good set of weather data for every city, you know what I mean? And it’s just, if you think about it from like a human analyst perspective, you’re just like, that’s going to take so long. And is it worth doing that analysis?

In two days, we were able to connect the data source for Helly Hansen, have them interrogate it, and sure enough, they were able to get those answers. And so it gives us these new ways of thinking about all the data that we can bring in that’s even external to the ad platforms that will give us better recommendations and let us take action or just understand the performance better than we ever have. And at a very fast pace.

So for example, we just integrated the entire Shopify applicant platform into Prism in four days. And so things just are happening so quickly. So MCP, this general, is going to unlock so many things that will allow AI to be better for performance marketers that in the next two years, I don’t even think we can predict today what that’s going to look like in two years from now.

We just know it’s going to be huge. And it’s going to really improve advertising, analysis, planning, recommendations, but even broader than just advertising, all of marketing. And so yeah, it’s going to be an interesting time.

Host: It’s absolutely an impactful trend to watch out. So, you grew your LinkedIn following by 10,000 and own the title of ‘Creator of the Year.’ So what’s your secret to building authentic engagement in a noisy digital world?

Jason: Yeah, my secret is, I would say, confidence. I’ve been doing this for so long. And I’ve done some very interesting things in my career.

And I talk a lot mostly about my career, but also on LinkedIn, I talk about personal things sometimes or like when I was planning to climb Mount Rainier, I had a series on every Friday, I would post about my road to Rainier. And when people meet me in real life that have been following me on LinkedIn, they tell me, you’re exactly how I thought you would be based on your posts on LinkedIn. And that’s music to my ears.

I love it because it is about authenticity. People gravitate to people that are authentic. So what that also means is when I post, it’s not always about sunshine and rainbows. I post about things that don’t work too. So here’s what I failed at this week. Here’s what I messed up on.

I messed up on this. Here’s how I messed up on it. So you don’t make that same mistake. And people grab, if you can be self-deprecating, not take yourself too seriously, but be smart and intelligent, have a perspective and a point of view and be yourself, then people will gravitate towards that. And it just happens over time. It’s taken, like I said, you said, I grew it by about 10,000 over four years.

I really didn’t start to post on LinkedIn until 2020. Well, I guess it was 2020. So I guess we’re coming up on five years now, but it takes that long sometimes.

And it’s a long game, but it pays dividends in so many ways. It gives me so many opportunities. It helped me grow my own fractional consulting business when I was a fractional CMO before I took this job.

And I just think it’s important for us as humans to share as humans, because not 100% of my life is awesome. And so if I present it that way, people will already know like, you’re not being authentic because nobody’s life is 100% perfect. And so posting things about failures and being okay with that and being okay failing, I think is really important.

And yeah, and people just tend to gravitate towards that. So it’s been a fun ride on that too, just kind of building my voice and growing my audience and seeing what benefits that brings.

Host: Yeah, indeed. Being real, confidence and authenticity absolutely win together. Such a great, great lesson. So that’s a wrap. Thank you, Jason. That was an incredibly informative session.

Thank you for sharing your expertise on AI, marketing, and the future of advertising. Your insights are truly valuable for us and our audience.

Jason: Awesome. Thank you.

Host: Thank you, everyone, for joining us today. I’m your host Rittika, signing off. See you in the next episode of ExtraMile by HiTechNectar with our next extraordinary leader on board sharing their thoughts and knowledge.

Stay tuned.