"Seamless frontend demands a complex backend":HDFC Bank's Deepak Oram on the reality of Martech stacks

Time of India

06-05-2025

HighlightsHDFC Bank's Deepak Oram decodes the shift from Martech stack fatigue to stack fitness and effectiveness. Why Agentic AI should be seen as a revenue-generation engine, not just a cost-saving tool. The myth of 'simple stacks': Why complex backends power seamless customer experiences. Consent-led personalisation and the role of privacy-first strategies in modern Martech. How brands can future-proof their marketing infrastructure by building Gen AI-native Martech stacks.
In an era where artificial intelligence is redefining the rules of engagement, building a high-performing
Martech stack
isn't about chasing perfection—it's about staying agile, integrated, and customer-obsessed. As
Martech
tools proliferate and tech capabilities deepen, the line between
marketing
, product, and customer experience continues to blur.
In this candid interview with ETBrandEquity,
Deepak Oram
, Sr. VP – Growth Marketing & Martech at
HDFC Bank
, shares his perspective on stack fitness vs. fatigue, the rise of Agentic AI, and why consent-led personalisation and cross-functional ownership are key to Martech's evolution in India's digital-first economy.
Edited Excerpts:
Given the pace at which Martech is evolving, is the concept of a "perfect stack" achievable, or is it more of a moving target?
There is no such thing as a perfect Martech stack. A Martech stack is like a garden—you have to periodically water the plants, prune the excess, and use manure to rejuvenate growth. What I'm trying to say is that it's always evolving because customer behaviour keeps changing, channel preferences shift, and the rules and regulations keep evolving. This is the ecosystem in which Martech operates.
From Stack Fatigue to Stack Fitness—what does an optimised Martech ecosystem really look like today?
Stack fatigue is when people get bored of the old systems and want something new—for better agility, faster turnaround times, less data manipulation, and so on. That said, stack fitness is not just about speed. It's not about how fast or easy things are—that's a common misconception. Stack fitness is about how effectively you can perform a task. Sometimes, with fitter stacks where you don't need to manipulate data, you actually have to work harder. Ultimately, this is about customer experience—and faster is not always better. Stack fitness is all about effectiveness with the customer at the centre, whereas stack fatigue is something every marketer experiences.
Does that essentially mean having fewer tools and more usability?
You can definitely have fewer tools, but we're in an age where people will create their own tools—AI has made coding easy. People will build their own wrappers and hacks. However, the stack won't become simpler; in fact, it will stay complex. I think it's a misnomer that a simpler stack is always a better one. Imagine guests coming over and you cook something simple like Maggi noodles—will it impress them? Whether it's the customer service department or marketing, the backend will always remain complex. It's the complexity behind the scenes that ensures a seamless experience at the front end.
Martech tools promise simplicity but often add complexity. What have been your key lessons in simplifying integrations and ensuring seamless data flow across the stack?
I think many folks take their organisation's needs too seriously and end up customising way too much. Think of any vendor—they often serve 100, 200, or even 300 customers across India and abroad. They're constantly refining what suits marketers best based on tests across hundreds of organisations. These solution providers are very knowledgeable. It's not always the marketer who knows the most—we need to be humble about that. First, leverage what the vendor offers before customising. It's the customisations that become a headache five years later.
Have AI Agents begun to change the very architecture of the Martech stack?
Let's start with value and put the customer at the centre. For any Martech tool, the customer is the marketer. What does the marketer want? Globally and in India, one of the main goals is to reduce cost—cost of acquisition, cost of service, cost of communication, etc. Now look at a marketer's KRAs: have you grown customer revenue, customer satisfaction, the topline? That's the lens through which we must view Agentic AI.
A lot of the chatter today is around using AI for workflows like image generation—but is that solving the biggest organisational problem? We should look at Agentic AI not just as a cost-saving tool but as a revenue-generating one. For instance, humans don't iterate very well—we don't consistently change subject lines or A/B test landing pages. Every CMO sees iteration as a challenge. Imagine an AI agent constantly optimising subject lines or reducing bounce rates on landing pages. That's value creation beyond cost savings.
How do you ensure your Martech stack balances hyper-personalisation demands with the increasing importance of customer privacy?
Customer privacy is 100% paramount—regardless of industry. If you breach customer privacy and they find out, your brand reputation is irreparably damaged. You cannot take shortcuts for short-term revenue or experience goals. And it's not just about regulatory compliance—it's also about common sense. Don't do to others what you wouldn't want done to you when you're in the customer's shoes.
Then, how do you balance the hyper-personalisation goals?
Very simple—take consent. There will always be customers willing to provide access to some data for personalised offerings. Many people wrongly believe that privacy and marketing can't coexist. But look at cookie data—60% to 80% of users give consent. They understand that if they want an Amazon-like experience with relevant product recommendations, they need to share data. They love it when those recommendations are accurate. Customers appreciate receiving a call that directly addresses their query. Once you get consent, what's also critical is how you store, encrypt, and ensure regulatory compliance for that data.
In an age of platform proliferation, who should own Martech orchestration—Marketing, Tech, or Product?
Orchestration is essentially how an organisation defines its customer treatment. Let's set marketing aside and consider a call centre: is it defined by Product, the Call Centre Manager, the Regulator, the CEO, CFO, or CIO? The answer is—it's defined by all of them. The regulator says don't call without consent. The CFO says don't make too many calls—there's a cost. The CIO says use tech like VoIP for efficiency.
So, Martech today is no longer just about marketing. It's a technology that drives customer experience. And customer experience is a cross-functional mandate. Hence, Martech orchestration cannot be owned by a single function. It must be co-owned across the organisation.
If you were to build a Martech stack from scratch today, what's the first capability you'd invest in—and what would you consciously leave out?
Let me give you an example. We've had several tech revolutions, and one of them was mobile. Yet, even 10 years into the mobile boom, many large brands had websites that weren't responsive. That's how difficult it is to change technology.
A similar revolution is now underway with Gen AI. But everyone is asking—how do I mount Gen AI on top of my marketing stack, my call centre, or my app? That's not the ideal approach. Why can't we integrate Gen AI from the bottom—at the API level?
Think about it—what I show on my app already comes through an API. We're entering a space where stacks will be Gen AI-native, just like we once had to become mobile-native. It'll take time because legacy systems don't support this immediately. But if you're coding a new website today—make sure it's Gen AI-native from the start.