eCommerce has never been more competitive. Customer expectations are at an all-time high, while the space has never been more crowded with competition.
You need an edge to get ahead. And according to Preethy Ann Kochummen, Senior Director of Marketing & PR at Argoid, that edge could come from AI.
Preethy joined our sister CMO Alliance recently on an episode of CMO Convo to discuss how AI could take your eCommerce strategies to the next level. You can still listen to that episode, but if you're in the mood for a good read, scroll down to check out everything we discussed.
Preethy’s role at Argoid
Hi, Preethy. Welcome to CMO Convo. How are you doing today?
Hi. I'm doing well, thank you.
Thank you for joining us today. We’re going to be talking about a very interesting subject. We've talked about eCommerce before, but talking about it with AI applications is a new one, and I think it's gonna be quite new to quite a lot of our audience as well – exciting times!
Before we get into that, Preethy, maybe you can tell us a bit about yourself and your role at Argoid.
Sure. Thank you for inviting me. I currently head marketing for Argoid. Argoid is a search and personalization recommendation engine, and we offer personalization to eCommerce and OTT. All of that personalization is AI-driven. That's why I'm here, talking about AI in eCommerce.
eCommerce in a post-pandemic world
Let's talk a bit about what eCommerce looks like right now. It went through massive growth during the pandemic. Is it still like that now? Are we still living in glorious times for eCommerce, or is it a bit different now that we're moving into a post-pandemic world?
I think that growth was expected to sustain for a long time, and that's where certain estimates and expectations have had to change. A lot of people prefer to go out and shop in brick-and-mortar stores, so eCommerce, instead of continuing on that spike, dipped in a lot of regions around the world.
Recently there was a big layoff at Shopify, which is one of the biggest eCommerce platforms; that tells us that there were gaps in estimations. Expectations have changed now, in the post-pandemic world, so it is a little bit more challenging.
Despite the drop in numbers and growth predictions, customer expectations for eCommerce are still high. We expect very speedy delivery, and we expect our shopping baskets to be shared across multiple platforms. Those high expectations are going to continue even in a post-pandemic world, surely?
Yes. I think eCommerce currently has a really tricky job of keeping all its customers happy all of the time. Wherever you are, whatever channel you're on, you expect that the brand or the retailer gets you and gives you great service with an omnichannel experience. So whether you're shopping online or you're in a store, all the expectations are still there.
So yes, I think the customer still wants what they want, and I think that's where eCommerce has to maintain that balance between what works and the customers’ demands, and find the ideal way to bring these experiences together.
There are certain things that online retail has huge advantages in, then there's the whole feel that people want from a brick-and-mortar store. Every retailer and brand is trying to straddle these two worlds and bring a balance. I think that's where the challenge lies.
Leveraging AI for personalization in eCommerce
Personalization has got to be one of the biggest challenges. We've moved beyond the world of people being impressed by Amazon recommending products based on your purchase history, and we expect more from eCommerce now. It's similar to what you were saying people want – that feel from brick and mortar. They want to have that human touch, even in eCommerce.
With that in mind, how does AI offer a solution to these challenges? Is it just making us do things better and quicker, or are there certain things that AI can do that the human touch or even just standard automation can't do?
Yes, absolutely. I think AI has now permeated every component of eCommerce. Maybe it's not as advanced as it will be a few years from now, but there is an AI solution to almost every aspect of managing a store online, whether it be helping to increase conversions, store management itself, inventory, or merchandising. All of those have AI aspects to them.
Personalization is something that AI can do way better than any other solution. I know this from my experience at Amazon, where I used to do something called site merchandising. At that time, all of the site merchandising that happened was simply the site merchandisers’ ideas and thoughts that found themselves on the page.
It wasn't difficult to do from an execution point of view. The more challenging part was to think about what your audience might potentially be interested in while considering the entire audience as one homogeneous mass. You might be able to segment them, but then even then, they are like a mass of people.
AI has been able to make shopping online feel like you’re visiting a store built for your needs. Whatever you see in your store is just meant for you. Even if there's someone else sitting next to you browsing the same site, they see something totally different.
You can't achieve that level of personalization in brick and mortar unless you're very wealthy and have access to personal shoppers. Having a store that's pretty much designed specifically for you – that’s an advantage of eCommerce over brick and mortar.
Exactly. Either you're very rich, or you have a little neighborhood corner store where everybody knows everybody else, then you get that level of personalization. Maybe. If they like you.
AI can deliver that personalization at scale for thousands of customers who are making thousands of visits, placing thousands of orders, and buying thousands of products. Store owners have a lot of products. They’re getting new products every day, and the number of SKUs on their websites is increasing. How do they make sure that people are discovering those products?
If there isn't enough data about a new product, and there isn't enough data about a new user, how do you match them? That's where AI plays a role in product discovery. There are a whole lot of ways in which AI can work to even solve what is called “the cold start problem” and match individuals to products they want and are interested in.
Using AI for an omnichannel shopping experience
eCommerce isn't just about people sitting at their laptops and placing orders now. It's a multi-platform, multi-device experience. People might start browsing on TikTok or Instagram, and then they might switch to their computers to make the final order.
Being able to keep track of all those orders and having that shopping basket moving across all these different platforms must be a lot easier to manage using AI than with humans.
Absolutely. You mentioned this off-site thing – when a user is away from the website and you bring them back to that retailer or brand. We have been hearing a lot about AI being used in email for that.
When someone's abandoned their cart or has made a purchase, there’s an opportunity to bring them back using personalized email, which is against something that AI can do. You couldn't imagine doing that manually – curating products for every individual that bought or has an abandoned cart.
AI can do that easily. It just has to look at the products that the customer might have been interested in and what they have abandoned, then send them an email with personalized products or offers to bring them back. This means AI is able to play a role in customer retention as well.
There's a lot of data showing you earn more from retaining your existing customers than trying to acquire new customers. If you can make sure that they stay, or they come back, and they're loyal to you because you understand them and can offer a whole bunch of personalization, that changes the game.
It's not just about calling customers by their first name in the email or showing them something that they recently bought. It needs to go beyond that. You could do it manually, but AI can take that up several notches.
Using AI across the eCommerce customer journey
Customer service has got to be another big application for AI. You see a lot of AI-driven chatbots and things like that these days, especially on Amazon.
There must be AI in the back end as well. You mentioned inventory earlier – it must be possible to create predictive models to work out the levels of inventory you'll need, and do loads of other things with AI in eCommerce operations. Could you tell us more about that?
Yes, absolutely. That's something that we look at in Argoid – mainly the conversion aspect of it and increasing the basket size and average order value. But AI can impact various aspects of the whole eCommerce customer journey, from customer service through chatbots, to a seamless store design, and how they encourage you to purchase and not just leave halfway.
Also, when it comes to product descriptions, there are a ton of uses for AI. Product catalogs are growing in giga- and terabytes every day, so, how do you ensure that you have that volume of quality product descriptions that make sense? AI again.
Search has gone beyond just keywords on your eCommerce website. The words you use to describe your product, may not necessarily match what the customer types into the search box. That needs an AI push.
There are various other things like virtual shopping and automated or shoppable ads. All of that is AI-driven. The entire journey, end to end, is becoming very AI-based. Not everything is very advanced yet, but I think it's great that we're getting there, and we’re not expending human effort on things that AI can do better. That lets us focus our attention on things that we as humans do better, like getting to know people.
Exactly. If you've got AI functions taking care of the day-to-day stuff, it allows the human element to devote their resources and creativity in a better way.
Customer service is a great example of that. If chatbots take care of the FAQs, it allows your customer service personnel to focus their efforts on major problems and give the best customer service they can.
Introducing AI solutions to your eCommerce marketing function
So Preethy, we've talked a lot about the benefits of AI. Let's talk about some of the practicalities. What kind of infrastructure do CMOs need in place to be able to start using AI in their eCommerce? Is it a matter of having a certain level of customers required? Is it about having enough data in place? Is it about having the right personnel to manage the applications?
At the very basic level, the person who is planning to bring in AI has to have enough of an understanding of AI and its applications; otherwise, everybody’s shooting in the dark. While it may not be possible for everyone to have an expert level of understanding, they can learn about the tools that are relevant for their function, and there are a whole lot of marketing tools available now.
When it comes to content marketing, there’s AI that can write articles for you, but obviously, you need human oversight there. AI can even create graphics and images nowadays, and pretty well, actually.
All of these tools are available, and I think they’re going to be everywhere very soon. We don’t have the luxury of saying yes or no to AI. AI is here. It's about how mature you are with it.
Coming to the understanding part of it, I did a couple of courses in data science to understand what AI is like, and what applications there are. I got an understanding of machine learning and algorithms, how they work, and why they’re not always 100% accurate. There is a certain amount of accuracy, and there are different models, so you can learn how AI arrives at certain conclusions.
Once that understanding is set, we go about looking for the tools that will help with what we need to do. It’s useful to think about what it is that we spend the greatest amount of time on.
For example, if I or my team was spending ages on writing and research, I’d look for a tool that could browse the written material on a particular topic for me and let me look at it all at once. When I started thinking like that, I thought there must be a tool to automate this whole process, and I found a whole bunch that did just that.
We work a lot with Shopify merchants, and there are so many applications for AI that sit very well with store management. So whether it be doing product descriptions at scale or cropping and changing images and having a lot of variants, we can do that well via AI.
Similarly, we’ve realized that we can do so much more personalization simply by letting AI do its job, whether that is cross-selling, upselling, or making sure that the offers are tailored to the interests of that shopper.
There’s also ensuring that they don't abandon their cart, that they're given a compelling message while they're checking out, and sending them an email when they've left the store – all of those are things that we realized we can do when we have an understanding of AI.
Like I said before, it's not a yes or no thing anymore – it's just how soon. When are you going to adapt to AI and bring it in? A lot of things – your Facebook ads, your Google ads, and all of that – are very AI-driven already. It's just a matter of time before stores and eCommerce start to use them very consciously, but I think in many ways, we're already there. We're already using a lot of AI solutions.
It's certainly becoming the norm that you're utilizing these tools in some respect. At the same time, even if the marketing lead who’s going to run this project has done all the research and become an expert on AI, they're still gonna have to convince other stakeholders who might not be so familiar with AI.
They might be a bit nervous about it. They might see the cost of the applications as something to be worried about. Is there a way to show the benefits of AI for your business? Can you trial it on certain aspects of the business, or does it have to be AI over everything all the time?
Absolutely. That's the approach I take to any AI tool in my tech stack. I always either A/B test it with another tool, or look at the results I'm seeing from the tools and then decide to scale it up or down as the case may be.
Take Argoid itself – we always offer the ability to A/B test our product against other tools or solutions, and there’s a very transparent dashboard, so there's no way to hide behind anything. One way or the other, it's all reflected in your balance sheet at the end of the day.
So it is definitely possible to start small, see how it works and whether it integrates with your ecosystem, understand the benefits, and then scale it from there. That would be my advice.
Of course, I am at a startup, so things are slightly different here. There aren't a ton of approvals required before you get something done. It's still challenging because you have to prove ROI for whatever it is that you're suggesting, but maybe the risks are fewer. Also, being an AI startup, it’s easier to convince people about AI.
I can imagine that legacy companies and bigger organizations might have trouble bringing AI in, but there are a lot of AI SaaS tools available that make it possible to start small, experiment with a small cohort or a small amount of data, and then if it works, scale it up.
Let's say this test has gone well, and you're looking into rolling out AI in a major way. We've been talking about this mainly from a marketing perspective, but what other departments need to be involved? Who are the other important stakeholders within a business that need to be involved in bringing AI into eCommerce?
I think product is one of the most important functions because a lot of what an eCommerce company or a retailer does depends on the product. And when I say product, I mean the website or app, not products in the store.
The way the customer journey is designed is vital, so anyone who's involved in that customer journey needs to be involved in these decisions. Product and IT or engineering need to be involved because that involves changes in the tech, integrations, and data. So I think product, IT, and marketing are probably the biggest stakeholders here.
And then, of course, the top-down approach is always useful. If you have a CEO or a leader who is encouraging of it, who understands it, and also understands that it's not a magic bullet, that’s a huge advantage.
The future of AI in eCommerce
We're still in the infancy of AI, so let's talk about the future of the technology. There are a lot of cool ideas floating around that could eventually be applied if the technology develops.
For instance, personalization could go even further, with product descriptions tailored directly to a customer – not just the products, but the product descriptions. Netflix already does something along those lines, so that could happen in eCommerce.
What else could we see on the rise? And what are you seeing out there that could be exciting to get involved with at this stage, so you can hit the ground running when it’s rolled out more broadly?
Netflix is a great example. I read that they customize visuals to be more attractive to you and encourage you to watch that movie. They've got multiple variants of each movie poster, and they show you the one that's most likely to get you watching that movie. It seems like a no-brainer to bring in AI to do that for you, but then you start thinking, “Is it so easy to manipulate me?” Seems like it is!
There are a lot of possibilities in eCommerce. Something very interesting to me is the virtual shopping experience, where you get the feel of the offline experience but without the hassle of having to go to the store, fight crowds, and wait in line for the fitting room. The virtual try-on technology is especially interesting. You can have the experience of trying something on and knowing if it suits you without visiting the store.
I definitely see this coming together in offline and online retail. I think it’s similar to how we're struggling with hybrid versus fully remote versus in-office work. No one has the right answer yet. Some people want to go into the office; some people don't want to go into the office; some people want to go in some days; some people don't want to go in at all.
It’s the same with retail. People want to shop online, but they want stores as well. It's difficult to run so many brick-and-mortar stores if people aren't visiting them, but people seem to prefer to visit the store and then order online. As a retailer, what do you do with that?
I think the future is about finding the best way to bring these things together. Retailers need to be able to straddle these two worlds and straddle them well. They have to identify the things that work well online, and then offer the offline experience with the convenience of the online purchase, while maintaining a great customer experience.
I'm not sure if there is one ideal way that it would work, but retailers are already doing things like letting you shop online and then pick up your item in an offline store. There are also all these easy self-scan and checkout and self-checkouts available in offline stores now.
I think those are ways of bringing the best of both worlds together, and that's where I see technology moving. The ones that win are going to be the ones that bring them together well.
It's going to be very interesting to see how things evolve over the future and see what happens with eCommerce in general. As we said earlier, it’s been through some exciting times lately, and there are surely more exciting times ahead.
Thank you very much for joining us today, Preethy. It's been a really interesting conversation. We've talked about eCommerce before, but looking at it from an AI perspective has been absolutely fascinating.
Thank you very much. It's been great.
