Marketing 911
Marketing 911, is the podcast where we tackle the toughest marketing challenges at the executive level.
Whether you're navigating complex strategies, trying to reach your target audience, or facing shifting market dynamics, we're here to provide you with actionable solutions.
From digital transformation to customer retention, if it's a marketing crisis, we're here to help you solve it—before it turns into a full-blown emergency.
Marketing 911
Personalization Power: The AI Revolution in Marketing - Live from Marketverse
Ever wondered how banks and financial institutions manage to send millions of personalized messages daily? The answer lies at the fascinating intersection of AI, data analysis, and privacy management.
Sanjay Ramchandani, Executive Vice President specializing in MarTech, pulls back the curtain on how major financial institutions build comprehensive customer profiles—transitioning from unknown prospects to fully known customers through ethical data collection and analysis. He explains how these profiles feed into AI algorithms that determine the perfect products and offers for each individual, ensuring marketing is relevant rather than generic.
The technological evolution is staggering. Where companies once employed 40 people to manually update rates and personalize financial service emails daily, Gen AI has streamlined these operations to require just 10 staff members for oversight. This efficiency enables sending 4 million customized emails daily, each with dynamic rates and tailored content. Even more fascinating is the depth of personalization possible—some companies analyze 70 different signals from your phone, including battery patterns and pre-sleep browsing habits, to determine the optimal moment to reach you.
But with great personalization comes great responsibility. Sanjay shares a cautionary tale of what happens when permission systems fail to sync properly, resulting in 500,000 emails sent contrary to privacy preferences—a potentially costly violation requiring intervention from marketing, PR, legal, compliance, and data teams to rectify.
Ready to understand the delicate balance between personalization power and privacy protection? Listen now to discover how modern MarTech is reshaping customer experiences while navigating increasingly complex ethical considerations.
Welcome to Marketing 911. I'm your co-host, richard Bliss. We are in the unique aspect here that Brian Baxter and I my co-host Brian, thanks for being here. Oh, pleasure, we are actually here. We are in person with each other in Austin at an event called Marketverse, having a great time having a chance to meet our guests in person Unbelievable, and we're thrilled to have them all.
Speaker 1:And today we have Sanjay Ramchandani, who is a specialist and executive vice president in MarTech. I like that term. Martech that's the idea of the new technology layers. Sanjay, thanks for joining us, thank you. In today's world, we're seeing technology. We've had a panel on AI. We've been talking about going to that. Where are you seeing the technology evolution happening inside your organization? You, you work for a big company. Uh, you have a lot of involvement there, from personalization to technology to flat deployment. Where's that happening? Where do you see that happening?
Speaker 3:so we're seeing technology evolution in two or three places, right. So first thing is, we want to know more about the customer. The buyer doesn't want generic emails or generic messages and it doesn't interest him. I don't even look at them, and sometimes I don't even look at emails which says Dear Sanjay and then do not. Something right, as in the quotes and messages it's a coding thing, right? So we want to know more about you personally, whether you're a customer or a prospect or in the process of becoming a customer.
Speaker 3:So the buyer profile, building a buyer profile with the right ethics, privacy and governance is very important for us. So we get data from Axiom, we get data from Nexus, nexus and all that Build your profile, and it's a journey for us to know you from an unknown to a semi-known to a fully known customer. And, based on how much we know about you, we feed this into AI algorithms, predictive algorithms that are underneath our platform, like Pega, things like that, which gives you the next best offer. So when you are on our website, we basically give you the right products and the offerings that we want you to see for your persona and for your profile, with your knowledge, with the information that we have.
Speaker 1:That's one. Well, no, and I'm going to stop you right there because I like that, because what you're saying is is that you are focused on I'm going to call it selling me something I need. Yes, Not something that you need to sell, but something that I need to be better as a customer. So I like that, that personalization there.
Speaker 3:And then number two so, as an example before I go to number two, if you have a car and then you're looking elsewhere for a deposit or a lending loan or something like that, we can combine that and give you added information and incentives to pull the trigger with us right. Because, we also collect information from various social feeds, like if you visit internet or Facebook or Instagram, whatever right, so we also collect that. It's a 360 view of you that we want and we are collecting that in our platforms at the most.
Speaker 1:I was going to say I'm seeing that because I fly Delta a lot and when I put up my profile on Delta they asked for my Twitter handle. Right, so they're following me on twitter. But I also know that if I get really upset, I want their attention. They know who I am. When I make a comment on tag them, they can go back into my profile and see who I am and then respond. I mean it's kind of. I mean some would say it's creepy because it's tracking me, but I gave them that information because I want them to give them the permission.
Speaker 2:Let's say it was four or five three years ago, prior to AI. How difficult was it for you to gather the individual, unique characteristics of the folks and be real personal? Has AI just dramatically increased your?
Speaker 3:real-time. There are two parts to AI the predictive AI, the collection of all of these states, was there even before three years, right? So we have connectors to all of these social feeds and with Google AdWords and all of these, you're basically able to collect that data and apply predictive algorithms the nearest neighbor or whatnot, kkn and stuff like that. What Gen AI has done is to take the content and give the right content, place the right content for your profile and use that to scale up your campaign offers. So, for example, disclosure management. So, for example, disclosure management. So in our company, the financial services sector itself, everybody's doing this. Rate of your offer could be between 19.99 and 31.99 for your buying purchases or for your check purchases, right, and that keeps changing every day. It's a library rate. It's a rate that.
Speaker 3:So what we used to have people like 40 people doing emails and checking if the rates are correct every day, because the rates change now, and editing the email and personalizing the email and all of that now is done by here, and that's the goal, right, you don't have those 40 people. We are repurposing the 40 people from 40 to 10. And that's where the human is. We still have checks, checks and balances, but you know the 40 people are being reduced to 10. So it's a whole lot.
Speaker 3:So assume a bank has 78 million customers, which is true for big banks like Wells and JPMC and CNB and PNC. We want to send you like 4 million emails a day, right, and four million emails a day need to have offers, prime rates, the conditions of all of these offers and what is the max you can be charged, or when do you get charged? 30 days, and all all these are dynamic, they're not static, and that gen ai is enabled and really kind of made it automatic, right, so you're, you're going to repurpose that, you're going to extract that efficiency and this is where humans and agents and all of them work together. Right, because you can fine-tune the agents to differentiate based on the profile. What content you send Will the right content for you? Suppose you're a Wisconsin Green Bay Packers fan. You send Aaron Rodgers. I mean, he's no longer now.
Speaker 1:He's not with the Jets either. No, he's not.
Speaker 2:Nowhere at all, that's true.
Speaker 3:So incredibly customized, incredibly customized to the right. For example, you're very, very active at 245 at Austin. You would be getting an email at that point in time. That's how they predict. And when do you open? There are some companies which we're talking with which look at 70 signals from your phone Like when do you open it? When is your battery level high? How many times do you open it? When is your battery level high, how many times do you open it? All of these signals also go into the repository of your profile and get aggregated with 360. So how long do you do see before you sleep? When do you get up? When is your charge the highest? When do you look at it for long?
Speaker 2:It's based on your permissions. Can you tell what applications you look at it for law planning? Can you tell it's based on my permissions? Yeah, can you tell what applications you look?
Speaker 1:yes, now again, let's be clear, because I just said that this is not uh by default. It's not right. It is not by default. You have to give it just like that. I went into delta and they said you want to share your uh twitter account? Yeah, here's my twitter handle because, as song gives you value, it gives me value for them to be able to communicate with me, and it's worked. And it's worked. You know, nothing happened. I was just in london, uh, delta's flight got disrupted by the london fire in uh heathrow and suddenly delta's taking care of me because they know who I am, that type of thing. So it's not. Let's make sure, because our audience is going to freak out. Some of them are going to freak out that privacy is being violated here, yeah, but it's not.
Speaker 3:It's not. You need full permission and there are how do you call that? There are top-level permissions that you give. There's more fine-tuned and bottom-level permissions that you give. For example, what channel you want to be communicated yeah, I forget the word that is used, but it's very fine-grained permissions, right? So, in terms of how you want to be communicated, when do you want to be communicated, what channel you want to be communicated, and like it's even more Like SMS, even that, like between 4 to 5, I do not want to be communicating Only on a watch, only on a mobile. So all of that can be established with compliance. And we have some tools in the market, like OneTrust, which integrate with marketing tools like SFMC, email providers, sms waterfalls all of these are coming together in an ecosystem play which look at privacy, governance and offer management.
Speaker 2:So one of the things we talk about on Marketing 911 is when you know something goes wrong and in this particular case, with managing this entire platform, what is something that either has or could go wrong that you get concerned.
Speaker 3:There's a lot which can go wrong, right? For example, I know from my past experience what has happened is Salesforce Marketing Cloud embeds the attribution, like when you click an email, the email click data and your data is sent back to the company which is collecting your data, whether it's bank or anybody else, right? So if I have changed the permission settings during that time and the email attribution has not changed, then it's a big deal. So what happened in one case was there was a sync between the permissions granted. It was not synced in real time and we had already sent out about 500,000 emails. So most of it and it's a violation. It's basically you could be fined of it and it's a violation. It's basically you could be fined millions of dollars because it's violation of privacy. So we had to basically understand what was collected, remove them, clean them up huge effort and then stop that bug. The bug was not syncing in real time between the permissions that was granted in one of the software.
Speaker 1:Yeah, I've seen this. We had to redo it Now. That explains why some companies, when I opt out, they'll say okay, it'll take a week before this to take effect and they're giving themselves that buffer to make sure everything gets in sync.
Speaker 3:Yeah, and that was not clear. Even if they gave a week, there was a bug that didn't like in this case. It's not a week for us, it's much sooner one day, but it didn't happen, yeah.
Speaker 1:It was a. It didn't happen. It was a real time integration issue or a sync issue, and then, when that happens, then the marketing team is talking to the PR team, which is talking to the legal team compliance teams, data teams, because we have to identify first which email is it.
Speaker 3:It's a difficult task. We have to identify all that has gone through. We have to identify which parts of the data which is collected is valid and which parts are not. Because they might have just said I don't want you to collect data between one and five and you have to understand that and remove that data from the database and assure them that it's not good. So you have to send back confidence emails to the customers and consumers.
Speaker 1:And the good news is that's not something that AI can easily fix right now.
Speaker 3:It's not something that AI can easily fix right now. It's not Right. It has too many. Yeah, there are too many interactions and too many sources that AI is not able to get. It can help you with the messaging, can help you with the content generation, if you keep bringing in the data, but identifying and solving it's just not yet?
Speaker 1:Not yet. I think that's the answer to every AI comment Not yet, because today, right now, the recording of this podcast is the worst that AI will ever be Right the worst, that's right, that's right, and so it might not do it today, but it's certainly going to be something that we look at.
Speaker 3:The newer models of reasoning are getting way better. I think they can reason and if you give them the right information, they can go and look at certain databases and get it. They'll be much better than two, I'm sure. Yeah.
Speaker 1:Well, thank you, sanjay, for being here with us. It's been a fascinating conversation. We've had a chance to get to know you a little bit here before we started recording and we'll have the chance to interact with you as we continue to be here at Marketverse in Austin. You've been listening to Market 911. My co-host, brian Baxter, and myself. Our guest has been Sanjay Ramchandani and we've been talking about market tech and kind of how marketing fits into all of that. It's been a fascinating conversation. In this case, we hope you found it interesting. Fascinating, I know, brian, and I always do. Definitely. Thanks again, thanks for listening.
