Marketing is Still Not AI Ready

Jim Marous (0:00)

Your growth plan has no fuel. We just published this year's State of Financial Marketing Report, and more than half of the institutions we surveyed have generative AI up and running in marketing right now. Yet, not one of them told us their customer data is immediately available and AI ready. Not one. And I want to be clear about where that comes from, because we didn't grade anyone. We handed them a ladder and asked, which rung were you on?

Some said their data is fragmented across systems. Some said it's centralized but stagnant. A few said it's real time. But every single one of them put themselves below the top rung.

These are people like you telling us their data isn't ready and then planning next year as if it is.

Let me start with the part that stung a little. Hyperpersonalization might be the most used phrase in financial marketing. I've used it myself on stage more than once. Roughly one institution in a hundred say they're actually doing it. We ask people what they meant by the word personalization.

Two-thirds of what we call personalization turns out to be choosing which audience receives the message.

That's list selection. When I ran marketing at a bank, that's a form of what I did. The list came from a core tape, sorted by zip code, and balanced here. We picked the segment, we picked the offer, we dropped the mail, and we measured the whole thing by counting the coupons that came back. We call it targeting.

And an enormous amount has generally improved since then. I can reach that customer wherever they are now, not just in one mailbox. I can do in an afternoon what used to take my team months. I can even test things I couldn't have tested at any price back in the day.

None of that is small, and I don't want to wave any of it off. But look at the decision beneath all of it. It's still about which list gets which message. Better tools, far more data, and the same call I was making using zip code and balanced years ago.

Our own research three years ago found that segmentation was the strongest data capability in this industry, and nothing in this year's numbers have moved beyond that. Here's the thing about targeting though, the customer never experiences it. Nobody's ever opened up an email and said, wow, they segmented me really well. What a customer actually feels is the offer, when it shows up, how the onboarding went, and whether anyone noticed just what happened in their life. Those are four dimensions we've invested in the least. We put our money into the one part of personalization that the customer can't perceive. And that's a reporting problem before it's a technology problem. Your organization can honestly report a strong personalization capability, yet still deliver something the person will never call personal.

That's how a gap this size survives every report we sent upstairs.

So why is this gap persisted? Because everything beneath it still runs through the data. People hear fuel and think budget.

Here, fuel is whether your customer data is unified, up to date, and usable by a system the instant it has to act.

In our research, roughly two-thirds of the industry describe data that's either scattered across systems or sitting in a batch update. Almost nobody described it as real-time, and nobody described it as ready for AI to act on. That explains our second finding. Most of us aren't using predictive AI at all. Even fewer have moved to eugenic AI. Generative AI writes more messages, while predictive AI decides who should get them and when. I've talked about this in a few recent Banking Insight videos about AI making us faster without making us better. And this is the whole story with a number on it. The zero didn't surprise me. I'll be honest, I have expected it. What surprised me was the effect in this chart. We asked which tactics are actually working, and AI-driven content and creative came in dead last. Almost nobody in this research considers it effective, and it's taking one of the biggest budget increases in our whole study. I want to be careful that nobody hears this as a reason to wait, because that would be the wrong lesson entirely. You don't need perfect data to run the machine. Good solution providers build strong targeting from imperfect data every day, and most of you should start this quarter. Think of it as grades of fuel. You could run a high-performance engine on regular. It'll start, it'll get to where you're going, and for a lot of what we do, that's honestly okay, but it won't allow you to do what you bought that engine for. Targeting on a lower grade is exactly what I did with that tape from the core, and it worked. What we can't do with it is notice what has just happened to a customer, and respond while it's still happening. That takes the highest grade fuel, and not a single institution in this study said it had reached it. So, the machine we're running is the one I ran decades ago, just a lot faster. There's a chart near the back of our report that I keep coming back to because it's the one thing here that's entirely within our control. We outsource our strengths and protect our weaknesses. The research showed that most of us bring partners in for media planning. About half of us is bringing them in for creative production. Personalization and decisioning are the least outsourced functions in the entire study. The two disciplines we say we don't have the people for are the two we're least willing to let anyone help with. Some of that caution is legitimate. Decisioning touches customer data, governance, and fair lending. But protecting our weakest capability from outside help is also how a temporary gap becomes permanent. That same shape shows up on the growth plan itself. Respondents to our research said deposit growth is the number one priority, and they rank it at the top of almost every scorecard in this study. The two things that actually produce deposits that stay, like being the customer's primary financial institution and growing share of wallet, sit near the bottom of the same list. Acquisition wins in the landslide. Advocacy and organic growth are afterthoughts at banks and credit unions whose entire competitive claim is the strength of their relationships. I want to be careful here though, because the easy read is that this is a big bank story and the rest of us are just behind on budget. That isn't what we found. We ran a separate cut on our research for every organization under $10 billion, and the maturity, the methods, the technology stack and many other dynamics were the same. But so was that zero.