E-Commerce and Digital Transformation – Manish Malhotra, SessionAI

This is Textron tv. Hello, I'm Amanda Ani. I'm excited to be here with Manish Mara.

He is the Chief Strategy Officer and head of data Science for session ai. How are you doing? Doing well.

Thanks for having me, Amanda. Glad you're on the show. So can you share a little bit about your background and experience and session ai?

Sure. Uh, let's start with the company session. AI is, uh, in new, uh, type of marketing platform.

We call it in session marketing, uh, which is basically aimed towards helping brands solve the problem, the challenge of anonymous site visitors. How do you convert them into customers when you know very little about them? Um, I am a co-founder at the company, uh, and, um, I, um, had the business strategy as well as data science operations.

Wonderful. So when it comes to e-commerce, how important is that customer experience for businesses? Well, I think, um, the last sort of decade has, uh, certainly, you know, told us that it's paramount.

Um, the fact that, um, you know, um, you know, during covid, which was a big upheaval, uh, e-commerce grew, you know, is sort of a testament to the need, the inherent demand, if you will, uh, that customers seek for convenience and, and, and really the ability of the brands, uh, to essentially meet that convenience, you know, both in terms of the shopping experience as well as in terms of the gratification of getting the products on time when the customer needs them. Well, you know, really, um, uh, I was talking to, um, you know, one of the leaders at Abercrombie and Fitch very not very long ago. And, uh, you know, brands like Abercrombie have seen tremendous movement, you know, to e-commerce.

So these are brick and mortar retailers, uh, who've been around for significant amount of certain amount of time that have essentially seen, um, their e-commerce business grow up. It's now upwards of 50% of the overall business is coming from e-commerce. Uh, so really, you know, if you take a sort of a, a, a, a step back and say, wow, that is pretty significant.

We've grown up from 15% to 20% to now 55%, you know, of the business. So clearly digital experiences, uh, you know, are, are, are crucial, you know, in terms of sort of really where we stand and here how we go forward. I've even seen that is significant and I've even seen some companies that have literally closed their brick and mortar locations and are completely online.

Absolutely. And it's interesting, you know, I I come from India and um, you know, there's sort of a movement there, um, on, uh, to mobile to mobile banking, mobile apps. Uh, there were brands in retail that essentially shut down their stores about eight years ago, and they were perhaps a little too early.

But I think we are reaching that tipping point here in the US as well, as well as globally. I think, uh, the movement towards frictionless commerce, you know, a better experience is undeniable. I think customers wanted, you know, brands, our ability to sort of like as brands, as vendors, you know, is sort of also continuously improving.

That's, uh, you know, advances in technology. I know, even me as a customer, I am expecting to be able to get what I want on the fly very efficiently and easily from my phone or online. So, um, it's very important, I think, for that being said, for businesses to understand how they can better harness the data from customers online.

How can they use the data to understand the customers and improve their revenue? Yeah, absolutely. So, you know, I wanna perhaps reflect on that question, you know, sort of really, uh, thinking about forces of change.

I think the first one you already sort of mentioned in terms of on the demand side, what are customer expectations? Uh, so we've talked about that. I think the second part is more regulatory.

Uh, we've seen over the last, you know, perhaps five years, 2018 is when, you know, safari first implemented controls on, uh, on, on, on, you know, user data. And, uh, apple sort of has been, uh, you know, progressively, uh, you know, getting better with their privacy controls and has done multiple releases since then, not just on, uh, safari browsers, but also on iOS and the operating systems in general. Uh, and, and sure enough, you know, uh, some of that is in response to, uh, the regulatory changes, the environment and privacy overall.

Um, third party tracking of user data is going away. Uh, it's not gone away today, um, but it's on its way. Um, obviously, you know, the, uh, implementations from Google are not there as yet, and that's essentially, uh, what sort of needs to happen.

But third party tracking is going away. First party data tracking, uh, which is pretty cosher. I mean, this is, you know, customers saying you can service me and essentially look at what I'm doing on the site in response to which you can service me better.

First party data tracking is getting harder. Uh, we don't talk about this as much, but, uh, essentially repeat visitors to the same site, um, beyond a week, you know, essentially don't really have, uh, browsers today don't really allow you to track that repetitive visit and really connected to a previous visit. So really from a privacy perspective, and consumers are okay with this, consumers are essentially, you know, sort of leaning into privacy as even as they want better experiences.

So, uh, to, to your question, I think it's, it's, it's, it's, it's a critical gap that seems to be sort of really growing. And, and, and, and, you know, as soon as some of the implementations in, uh, Chrome and Android, you know, come to market, this is gonna get harder and harder. Brands are having difficulties, um, you know, um, collecting data and attributing data to the same visitor.

Uh, a recent report that we did at session ai, which was based on a survey of a large number of retailers, we call it the, uh, the, it's an annual in session marketing, state of marketing, uh, you know, found that 90% of, uh, site traffic is anonymous. Um, and this is, uh, you know, grown. Um, this poses a challenge, uh, for brands to, uh, do personalization when 90% of your visitors are unknown.

Uh, that's a challenge. It also poses, of course, a challenge, uh, to their conversion strategies. How do you engage, how do you convert, uh, this large proportion of traffic on your site, uh, that you've essentially, um, you know, sort of really paid to acquire and you continue to pay repeatedly because acquisition on the other side is sort of really also getting, you know, inefficient, uh, as well as, uh, expensive.

So, so those are some thoughts really from the, you know, I think the brands that we speak to, the retailers that we sort of are engaged with, we hear that all the time. So how can we, um, you know, there's a lot of new technology and it's rapidly expanding. How can businesses harness the new technology?

What kind of technology should they be looking at? I know there's AI and machine learning. How can we utilize those in these digital transformation initiatives to improve and uplevel?

Yeah, yeah. Uh, again, uh, it's a wonderful question I wanna reflect on. Uh, so technology needs to service a need.

And so what is this need here? First, let's sort of really, you know, uh, crystallize that, if you will. So we've established that, you know, it's getting harder to acquire traffic, it's getting more expensive to acquire traffic.

Why? Because of, you know, the, the, the tracking sort of of third party cookies and users through that data transfer essentially getting harder. Um, it's also getting sort of tougher to re-target, you know, people once they've come to your site and essentially after you.

So, so you, you have a challenge on acquisition, but you also have a challenge on retargeting or reacquisition, if you will. So that essentially says, you know, that we have to look, uh, take a fresh look at what's happening between the time somebody comes to the site or the app and before they go away. That window of opportunity is what we really think about as in session.

And so that's the motivation for in session marketing. Uh, now technology of course, plays a big role here because one would say that the, with that sort of really, you know, window of opportunity, uh, the data points perhaps may not be that many. You are looking at, uh, understanding the behavior of, uh, visitors within the first few clicks to be able to sort of really service their needs.

Now, seemingly, you know, that's not enough, but if you, again, take a look at where technology is, the state of machine learning and, you know, artificial intelligence, the advances there over the last several years have actually shown, and we've been doing this for about five years ourselves at session ai, um, that, um, within the first five clicks, it's possible to predict with very strong fidelity each visitors' intent to purchase within that visit. And if you can do that, well, then there's sort of a host, a plethora of, you know, actions and opportunities to intervene and influence and service that particular visitor. That opens up.

Again, it's important to remember that we are really talking about using information that is essentially based on the activity of these visitors, not who they are, and certainly not based on, you know, their demographics, social graphics, or third party data that you might be sort of acquiring or thinking about acquiring. You don't need those data points. We are just talking about what we call as in session data.

So that brings up a good point, um, you utilizing that data to know what the majority of customers are wanting to purchase, what were they considering, that's certainly helpful. But then with the customers whom we don't have any of that information and we don't have, um, their contact information, what is the solution there? How can we, um, use tools and technology to perhaps persuade customers that are online to share their information before leaving the site?

Um, just because I know that can be really critical when they have data. I know for me, I will go and I'll put some things in the shopping cart, for example, um, let's say Amazon, and I'll put some things in the shopping cart and then I will leave and I'll get busy and I'm, I may, may or may not remember or might change my mind, but then I will very quickly get an email, uh, reminding me, Hey, we saw that you have these things in the shopping cart. Would you like to go back online?

Here's some other suggestions for you. And often that really does work. I will get back online and I will make the purchase.

So, um, what are your solutions in regard to this? Yeah, um, uh, it's, it's really important question, uh, because again, technology, uh, to surface insights is only part of the solution. We have to be able to take, um, what we call as real time in session actions.

And, um, you know, the way I think about this, Amanda, is the purpose of in-session marketing is to improve conversions and reduce drop-offs before they happen. So to your point, that email that is sent out is after it happens, and that will continue to play a role. Uh, however, as you said, uh, that email can only go out for known visitors on whom you have an email, but remember, 90% of the traffic is anonymous and probably not coming back.

Or if they're coming back, it's gonna be harder and more expensive to get them back there on the site. So what do you do for them? Well, in session marketing, you know, applies real time tactics to address all active users on the site using the intelligence, these propensities that you've determined.

Uh, what are some of those actions? I think those actions can be certainly promotions, uh, which are very targeted. So think about, you know, understanding your highly likely buyers, your unlikely buyers, and people who are in the middle, we call them as on the fence, you know, visitors.

Uh, the opportunity now within session marketing is to do targeted promotions that can influence the on the fence audience while that audience is still on the site before they go away, you know, to essentially get them over that fence, over that sort of, you know, hesitation. How do you do that? Well, of course, in incentives work, but there are other ways to sort of intervene as well, you know, establishing, uh, perhaps the uniqueness and the differentiation of the brands you carry.

Uh, perhaps the service. Going back to that, you know, opening question about, you know, what is it that, you know, let's say if you're a Zappos, well, you know, returns are easy, returns are possible anytime returns are without questions, but there's, you know, really a lot of value propositions that, you know, every brand has that it's time to surface that for first time buyers or for repeat visitors who haven't bought in some time. Uh, those are the opportunities, you know, to sort of really take real time in session actions.

Can we talk about some of the newer trends in e-commerce and maybe not new, but they are more improved. I know, um, one thing I see more than ever is in real time chatbots, they seem to be much more advanced these days. Uh, and there may be some that I'm not thinking about.

So can you share with that? Yeah, uh, I think just wonderful progress, you know, overall, um, you know, sort of really in the chat space every, everything's chat, G p t and, and for a good reason, I've been playing around with that. And I think it's wonderful in terms of the, the advancement of, uh, generative AI and, and really the maturity to which it has come.

Um, you know, the underlying technology here is something that's about four or five years old. It's based on something called transformers, but large language models have this uncanny ability to have a conversation and to sort of really, um, you know, hold that conversation, uh, based on the context of what is being, uh, you know, sort of like, uh, what is evolving, you know, really topically within that conversation. And when you think about sort of that relevant contextual data, uh, that is the, the basis, if you will, of these learning algorithms that are able to sort of really look at contextually, um, you know, what is it that the current question or prompt is, and what is it that it's sort of really based off of one, two, maybe even three, uh, prompts prior.

You don't even have to specify what you're asking about, you know, sort of like in in full detail. It's very different from search engines, if you will, where every interaction is sort of a new context. Um, uh, when it comes to, you know, the kind of work that we are doing at session AI in terms of using in session intelligence, which is not necessarily generative, it's based on sequence analysis, but we think about it in very similar terms, the context of micro behaviors that are being expressed on every single page view on every single click is additive and cumulative.

And, and, and I think that's sort of the sense of, you know, doing predictions at five clicks, cuz it's not just the fifth click, it is every click matters and therefore the session matters. So I hope that gives you an idea of really some of these very visible advancements, you know, based on chat G P T, that essentially are leveraging context, uh, through some very, you know, key underlying technology advancements, you know, these transformers, very similar things that we are doing in terms of these learning models based on sequence analysis. Wonderful.

So last question, looking ahead the next five years, what's the next big thing that you think needs to be focused on when it comes to e-commerce? Gosh, um, that's a great question and I wish I knew the answer. Oh gosh.

Um, you know, let me just again, go back and think about sort of that forces of change because, you know, as digital leaders, I think we really have to sort of really take care of the present, uh, worry a little bit about the future and certainly plan for it, you know, in ways that I think we can put our arms around. You know, budgets are always gonna be tough. Uh, budgets are always gonna be sort of really, you know, cont content.

There are multiple sort of, uh, projects, initiatives that are contending, you know, for limited budgets and limited resources. Uh, you know, obviously, uh, from our standpoint, we believe that the session is a, a a a new focus, you know, for, uh, marketing and for e-commerce. We've done this in the past in, in some ways, but I think those have been, um, really just the stepping stones to where we are today.

I think the next few years we are gonna see acceleration. We are gonna see this convergence between what we think about as customer intelligence or visitor intelligence and then content intelligence, generative sort of content that's, you know, really happening on the other side. And there's gonna be this convergence where content is based on content, but it's also based on the activity of the user.

And so really if you think about it in those terms, there is this significant convergence of technology and advances in technology that I think are really going to go back and give us brand new use cases. So what do we do today? Why I think we have to sort of really play in both codes.

I think we have to sort of really think about in session, um, you know, customer intelligence, customer AI as one sort of pillar of this, um, you know, uh, investments that we look forward into. And then the other part is the content and what, you know, sort of content are these visitors that these customers interacting with, and therefore how sort of that can dynamically be updated, changed, and so on. And that is all generative ai, which is, you know, sort of in the realm of chat c pt.

So I would say that look, as leaders, digital leaders, we need to sort of really think about the customer, the visitor, the content. And I think really make, uh, sort of attempts to move forward on both fronts. I think that convergence is what's ahead of us Looking at that whole picture.

Thank you. Thank you so much for sharing your insights with us today. I know I've learned a lot.

I'm sure our audience has too. And look forward to speaking with you again soon. Thank you, Amanda.

It was a pleasure talking to you. Thank you.