– Helping marketers think about how to leverage AI and martech for one or more aspects of their campaign creation and execution process
– Demonstrate how Braze can support an AI-supported campaign process
– Explain how high-performing marketing teams are leveraging AI today
Transcript:
A couple of key objectives that we want to make sure come across today. Really what we’ll be talking about are three main elements with a focus on digital marketing campaigns. I want to make sure that we’re rooted in that understanding and foundation to start. But what we’re really trying to focus on here is encouraging you all to think about AI and leverage AI in some aspect of your MarTech and campaign creation process. So that’s number one.
Number two is to demonstrate how Braze can support an AI-supported campaign process for generally available features that Braze customers and folks that we’re working with are using today.
And then number three is to explain how some other marketing teams are really benefiting from a combination of AI and automation to help make their campaign process more efficient and more effective. So we’ll go through those three main things. As you all have questions, feel free to throw them into the chat. I’ll try to answer them as they come in as I see them. Otherwise we’ll have some time at the end where we can go through some of those questions as well.
To start, my name is Bobby Tichy. I’m a solution lead here at Stitch. I lead our solutions team. I’m responsible for thinking through a lot of what marketers are doing today with technology — how could they be doing it better, and how could they be doing it more efficiently. I’ve spent my entire career in marketing technology across Salesforce, Marketo, Adobe, Braze, and Segment. And I’ve worked with a lot of really fun, cool brands over the years — some smaller mid-size, as well as enterprise like Disney and AT&T. I live in Fernandina Beach, Florida with my wife Joanie and our three dachshunds. So if any of you have dachshunds out there, throw a picture in the chat — it’d be great to see them. And I’ll throw a picture of our three up here at the end as well.
So I wanted to start with a poll. Really the thought here is that AI is a big hot topic for everybody right now. So in the Zoom poll, I want to think through who has implemented AI at some level within their marketing team or tech stack. There are a couple of key things to think through here. One, are you using AI? Two, are you thinking about using AI? Three, you don’t use it but you’re starting to put it onto your roadmap. And then lastly, you’re definitely afraid of it and you haven’t made a whole lot of plans to use it.
The reason why I wanted to start here is to think through what the process looks like for marketers to adopt new technology or new processes. And I think the main thing that we’re seeing across a number of our different customers — and in my years of experience working with marketers and marketing technologists — is that a lot of times we can be resistant to change, especially with something that we don’t quite understand.
The way I like to think about this is that a lot of us have gone through this marketing technology landscape, and there have been some leaders, there have been some laggers, but for the most part a large portion of us — if you think about a bell curve, that kind of 80% in the middle — are functioning and doing the same things, right? We’re using a marketing automation platform. We may or may not have a CDP, but we’re doing email marketing and some mobile marketing depending on the industry and company that we’re working for.
But one thing that’s really interesting is I’ve found that over time marketers do a little bit of the Milton from Office Space thing — where they’re very protective of the tools that they’ve used or are using, as well as the processes that go along with them. A good way of thinking about that is someone who’s been in Salesforce Marketing Cloud for their entire career is going to be very protective of that platform. Or someone who is developing emails from a Photoshop design into HTML and CSS every time is going to be very protective of that process.
And what I really want to challenge all of us to think through is how can we make these processes better? How can we look outside of what we’re currently doing to try to make things more efficient and more effective?
So as part of that, I’d like to take a little trip down memory lane. For those of you who are Golden Girls fans, you’ll recognize the picture of Sophia Petrillo. Sophia Petrillo was the matriarch of the Golden Girls, and anytime she wanted to tell a story, she started it out with “Picture it” — it would be, “Picture it, Sicily, 1923. A young peasant girl was married to the Prince of England,” or something like that.
So let’s go down that path for just a minute. Picture it, Indianapolis, Indiana, 2011. Let me tell you guys about a young man, an innovative software company, and a dream. The whole thought here is that, like I said earlier, as marketers and sometimes as technologists we’re resistant to change. We think we have a process down or we’ve done something before, and we think that’s the best way of doing it and it couldn’t get any better.
So this young man at this innovative software company was working with First Midwest Bank — a regional bank, headquartered in Illinois. They sponsored the amphitheater at a capacity of 28,000 people and would have concerts, comedians, all types of different shows. And First Midwest Bank was really ahead of its time — they wanted to do an SMS sweepstakes for concert goers. Again, this was 2011, about 12 years ago.
So what they wanted to do was, as you were coming into the venue at the amphitheater, they wanted you to text in — a very natural way of doing some kind of subscriber opt-in or customer acquisition. That way you went into a sweepstakes for a VIP experience. You would text in as you’re coming into the concert, and you would have the opportunity to be one of 10 people that could sit in the first three rows, meet the band afterwards, and have free food and drinks for the entire concert. So it’s a pretty sweet setup.
And so they wanted to create an automated process to do this. They were leveraging Salesforce Marketing Cloud, or ExactTarget at the time. As part of that process, we built this workflow where you text in, it sits in a table, and the process of selecting 10 random people is done and automated.
So as part of this whole process, this young man is building this out, and as part of the testing and QA process for this particular campaign and sweepstakes, the engineering team came in and said, “Hey, we’ve got this whole new process for testing and QA audiences. So that way, whether it’s an SMS or an email or whatever type of messaging it might be, we can ensure that it’s not going to more people than we expect and it’s actually going to the people that we expect it to go to.”
Well, this young man was like, “I’ve got this figured out. I’ll do my own manual testing and QA — I’ve got a process for it, my Excel document, so I’ll just go down that route.”
Well, long story short, it went horribly wrong. That young man was me. The innovative software company I already mentioned was ExactTarget, and I cost them over $50,000. The leadership team had to write a check for $50,000. Thankfully I didn’t get fired, but I learned a really valuable lesson.
And you’re probably wondering why I’m telling this story. It’s because as marketers — and sometimes as technologists — whether it’s testing or automation or innovation or AI, we don’t always do a great job of adopting new technologies or processes.
Not only is that story proof of that, but I want to give you some more. Going back to that time in 2011, we have what the batch campaign process looks like. Again, for the purposes of this webinar and for the tactics and processes, we’re going to talk through thinking through digital marketing campaigns.
So I initiate the campaign — I’m not going to go through each one of these steps. Many of you already know what this process looks like, right? We come up with a strategy, we decide who the audience should be, what the campaign assets should be. We have an agency or some in-house creative folks build up the copy and the content, all those different things. Then we have an agency or some in-house folks develop that from a PSD design into HTML and CSS. We QA it, maybe there are multiple versions that we need to create, and then we deploy it, and then we review it and optimize it. So pretty standard process, right? This is what that looked like in 2011.
So now on the next slide, I’d like to challenge you to identify what’s different about that campaign process today.
And if you can’t spot any differences, that’s because there aren’t any. This is obviously not a rule of thumb, but it’s the majority based on what we’re seeing day to day working with companies that have revenue of $100 million, $500 million, or $5 billion. It’s consistent across many of these different industries and companies.
And it’s really interesting because as we start to look through the future, and as we start to think about how AI and automation can help with these things, you really have to start thinking about the campaign process completely differently. That doesn’t mean that the four key steps here of initiating a campaign, creating assets, QA, and deploying are going to change — because for the most part, those will likely stay the same in some form as we continue to move forward.
But as we think about today, where marketers are spending most of their time — identifying strategies, identifying content, coming up with an audience, a send time, a send channel — we’re really focused more on the tactical elements, much more so than the strategic ones. And that’s where we really think AI and automation can help. Because if AI and automation can help take out a lot of the repetitive work, we can focus more on strategic work.
And that’s where you really see the “tomorrow” piece of it. And you can see there that both of these renderings were done by DALL-E, and our director of marketing does not approve — she graded them a C minus.
But as we think about marketers today, right — the challenges, just like the campaign process, are not all that different from 12 years ago. Technology might be a little bit better, things might be a little bit more automated, but really today we’re faced with the challenges of a lack of time and resources, relying on one-off or batch-and-blast campaigns that have to be created from scratch each and every time. And we’re spending a lot of money on things like creative, development, and deployment where we could be finding areas of efficiency.
So how do we get to the future of marketing, which is less manual work? We’re not just throwing things at the wall to see if they stick, and there’s more of a focus on brand strategy and experiential marketing. We’ll talk more about what that means in not just broad terms.
And that’s one thing I also want to call out here — I hear a lot of folks talking about AI and automation and how that’s going to change jobs and what that might look like. I feel very strongly about the fact that as marketers, we’re not going to be replaced by AI. We’ll be replaced by marketers who use AI to find automation, efficiencies, and repetitive tasks. And that’s where, to my earlier point at the beginning, no matter where you are on the maturity curve — whether you are a $100 million company or a billion dollar company, whether you have three people in your marketing department or 30 people in your marketing department — you can start using AI today.
So what does that look like? How do we get there?
If we go back to this batch campaign process today — the slide we just walked through — this is the current state, or what campaigns look like today for most companies. Well, if we take a look at just what we can do today with Braze AI, there are four key components where AI can start helping us specifically within the Braze platform.
Number one is content copywriting with the AI copywriting assistant. Whether it’s a subject line or the actual copy within a message, we can start to do that today with the AI copywriting assistant within Braze.
The second piece is the DALL-E Braze image generator that’s integrated into Braze. That’s another area where we can start leveraging that today and start testing with it. I think some people think that all of a sudden all images are going to be generated by AI and all copy is going to be generated by AI. That’s certainly not the case. But what we’re seeing be really effective for folks is doing the first pass at that — especially for content or copywriting. That way I’m not trying to come up with brand new copy or content for every type of message I’m writing. I have a starting point, I have a draft that I can work from.
For the Braze image generator, what we’ve really seen to be successful is people testing this or doing multivariate testing with images created from their content or design teams versus images generated from AI but then approved by their content teams. The brand is still extremely important, and we’re not recommending that as a brand you start generating all content with AI. But for different types of messaging, or pieces of a message that are maybe brand-agnostic, that’s where we think there’s some value and we can see some benefits and some efficiencies.
The other elements at the bottom are all powered by Braze’s Intelligence Suite — Intelligent Selection, Intelligent Channel, and Intelligent Timing.
Intelligent Selection is really neat because within a journey you have multiple swim lanes or multiple branches, and it is automatically selecting which swim lane or which branch someone should go down based on the performance of that particular variant. So for example, if I’ve got three variants with a 10% conversion rate, but 8% of my conversions are coming from branch one, Braze’s Intelligent Selection is going to automatically route people into branch one. The really cool thing is that this consistently updates over time — the high-performing branch of today may not be the high-performing branch of tomorrow. So it’s consistently seeing what has the highest conversion or engagement and then updating based on that.
The second is Intelligent Channel — this is exactly what it sounds like. If I’m opted in to receive push notifications, SMS, email, and web push, where should you reach me?
And then the last piece is Intelligent Timing, where we’re not only selecting the channel but exactly when the message should be deployed to me for the highest possibility of conversion or engagement.
So this is what we can accomplish today with Braze AI.
As we think through what really future-forward thinking marketing teams are doing, it’s this next-generation campaign process where many of these processes and steps are automated. On the next slide, we’ll walk through a sample diagram of what that looks like and where the differences between the marketer, AI, and MarTech lie.
But for this slide, what we’re doing is the front part — the strategy — is not changing. There’s still a marketer, there’s still a human element to it, because there are so many things that as a brand I’ve got to identify: what am I trying to accomplish, how do I want that to come across, and what are the KPIs that are part of that? It’s not always a conversion, right? It could be engagement, it could be a particular action that we want people to take — not just a purchase.
But the next big branch of the campaign process is creating campaign assets. And what we’re seeing here is folks leveraging some kind of automation or content management to have a treasure trove or library of campaign assets, and then automatically selecting what pieces of content are being used for that particular campaign.
Here’s an example. Let’s say I’m Papa John’s and I have a new pizza coming out — an Ultimate Pepperoni Pizza. I want to drive awareness for this new pizza. As part of that initial campaign, I’m going to set what that strategy is — it’s awareness. We want people to buy this new pizza. For the website, mobile messaging, email, whatever it might be, the content team or the agency is going to have to generate this content. They’re going to have to create imagery for the Ultimate Pepperoni — maybe there’s a wide shot of the whole pizza, maybe there’s a zoomed-in shot of how many pepperonis are on it, all these different things. You can equate that to whatever business or industry you might be in.
So all of that content lives somewhere right now — it should live in a content management system or something similar. That way, as we’re building out campaigns, we have a place where we can go pull that content.
So the first email or mobile message or web push might have the wide shot of the pizza. But what we’re seeing a lot of companies do — and what we’re helping companies implement — is based on how that particular campaign or message performs, automatically optimizing it rather than taking weeks to identify what’s statistically significant and then making those updates manually.
Here’s an example. I launch my Ultimate Pepperoni awareness campaign. Then two weeks later, we’re automatically seeing how it performed based on some kind of baseline that we established, and we’re automatically making adjustments based on that KPI. So maybe our goal was an 8% conversion — meaning someone actually buying the Ultimate Pepperoni Pizza. Well, if we’re at 6%, we could automatically reach into the CMS and say, let’s try some new imagery. Or let’s build out multiple branches and bring in multiple different images, copy, or content and see what performs at that next stage.
So we’re starting to automate, and that’s a key differentiator here — automation versus AI. We’re automating the process of bringing these campaign assets together and optimizing them, in conjunction with AI writing different content or copy, as well as figuring out what are the best days, best times, or best channels to communicate with those folks.
And the best example I think of this is an actual use case for a customer that we work with — what this process looks like end to end. At the very beginning, again that marketer focus — you can see the legend in the bottom right-hand corner. That generation of the campaign, the strategy of that campaign, is not changing. It’s still with the human, it’s still with us, it’s still with the marketer.
But then once I actually create that strategy and build out whether it’s a campaign request form or something similar, from there we’re using a lot of automation and AI to then generate those messages.
In this situation, this particular customer is using Airtable as their workflow management tool. As a marketer, I come up with an idea of what I want to distribute or what my campaign is. Going back to the Papa John’s example — I want to send out an Ultimate Pepperoni campaign. I go into Airtable, I type in all of the different elements, and we can even get advanced with different tones for what I want the subject line to be. We could do multiple versions — five different versions with different tones, some subject lines with emojis, some emails with GIFs, whatever we want. We can set all of that at the Airtable level.
From there, Airtable is communicating with a number of systems to generate the message. It’s communicating with ChatGPT for the actual copy. And then Airtable is communicating with Sitecore as the CMS to pull down the content related to that Ultimate Pepperoni campaign. And then Airtable, Sitecore, and Braze are all talking to each other, so that once we have the copy and content, we’re pushing into Braze to create the actual message. We also have Movable Ink as a key part of this as well, because if we’re pushing out an offer — especially if it’s time-boxed or event-driven — we want to make sure that people see content that’s relevant to them at the particular time they’re opening it.
And then from there, the proofs are automatically being sent to the marketer. So this whole process — from the time I decide what I want my campaign to be, to building out that campaign request form in Airtable, to getting proofs — is a matter of hours, not days or weeks. Then I can review those proofs, and those campaigns are then built through that automation in Braze, whether it’s a journey, a one-off campaign, or whatever it might look like. That campaign is now available to send, but it’s also part of our reservoir.
And what I mean by reservoir is that as we start to build out these campaigns, we have a catalog or library of all these different campaigns that AI can then choose from as we start to optimize. And a big portion of this, especially as we think about content, is building out what we think of as the lifecycle campaign process. And I don’t mean building out a journey for this process — I mean building out the types of content that we want to share.
This is a very simplistic view of what I’m talking about, but hopefully it provides context into how that content should be ideated and then tagged. So for example, we’re in Sitecore using the Papa John’s example, but we’re thinking about it in terms of onboarding. We have these four different lifecycle stages — inactive, lapsing, active, and loyal. I should have all of my content tagged based on these different stages as well as what the actual content is. So for example, a piece of content for the Ultimate Pepperoni might have the context of “Ultimate Pepperoni” but the lifecycle stage of “active.” That way we can quickly bring in these different pieces of content.
And this is really where the future of campaigns and marketing technology is going — it will not be the step-by-step process of building out a campaign. We’ll be really focused on the strategies, and then have a library or catalog of all these different campaigns that we could pull from at any one time to send to a subscriber, a customer, or a user.
So as we start thinking through that in the future, and where humans, humans plus AI, or AI alone will function, this is a key framework we put together to identify what that looks like. You can see across these different ones — red for human, purple for human and AI, and blue for AI alone. AI is not going to completely take over, and I think that’s another message we want to make sure comes across here.
But there are enormous opportunities to use AI to do data analysis, to perform repetitive tasks, to optimize the campaigns that we’re sending, and then leverage automation to bring these things together to make them faster and more efficient — as well as make things not only more efficient and more effective, but also a bit more foolproof. Going back to my story about costing ExactTarget $50,000.
So hopefully this was helpful in understanding where we’re seeing AI and automation being leveraged in the campaign process and also what the future looks like. If you’re not sure where to start and you need some help identifying what that looks like, we would love to help you. Go to this Bitly link or scan the QR code and fill out the form. We’d love to jump on a call with you to understand what you’re trying to accomplish, where you currently are on that digital maturity curve, and where we would recommend you implement some AI elements. It’s not a paid engagement or a paid call or anything like that. We’d like to talk through that with you, help you build out a roadmap for where you could use AI or automation in your MarTech stack or your process. And then if you’re interested in us helping you, we’d love to help. If not, you can use that roadmap internally to start leveraging and implementing some of those different things.
So thank you all for joining. Really appreciate it. If you have any questions, feel free to fill out the form or hit us up on LinkedIn. We really appreciate you joining. Take care, everybody.
