Klevu has been positioned in the inaugural 2024 Gartner® Magic Quadrant™ for Search and Product Discovery

Klevu has been positioned in the inaugural 2024 Gartner® Magic Quadrant™ for Search and Product Discovery

Home Discovered Podcasts Episode 3: Unlocking the Potential of Ecommerce Personalization – Insights and Predictions from Digital Industry Leaders


Episode 3: Unlocking the Potential of Ecommerce Personalization – Insights and Predictions from Digital Industry Leaders

Learn about creating a good personalization strategy, and personalization at scale, and discover how to maintain the emotional aspect of a brand while using personalization to drive business KPIs.

Episode Description:

In this podcast episode, the hosts discuss the dynamic world of personalization and its importance in today’s digital landscape. The episode features experts from various fields who share their views on personalization, including Vitaly Friedman, David Mannheim, Rasmus Houlind, Gianfranco Cuzziol, and Rob Boland.

The panel discusses the definition of personalization and how marketers can collaborate effectively by agreeing on a broad definition. They also discuss the importance of understanding the customer journey, personalization strategies at scale, and the emotional aspect of personalization. Listeners can expect to gain valuable insights and thought-provoking discussions from this panel of digital industry leaders.

Guest BIO:

Vitaly Friedman, Founder, Smashing Magazine

Vitaly loves beautiful content and does not give up easily. Co-founder of Smashing Magazine, a leading online magazine for designers and developers, Vitaly is the author, co-author and editor of all Smashing books and front-end/UX consultant. 

Rasmus Houlind, Author, ‘Make it all about me’

With his book ‘Make it all about me’ and his upcoming book ‘Hello $FirstName’ Rasmus helps marketers globally strike the balance of making money from personalisation, and delighting customers long-term without over-investing.

Gianfranco Cuzziol, Group Personalisation Head, Avon

Gianfranco leads global personalisation and CRM for Avon, and previously with brands like BMW, adidas, Disney and easyjet. With a first degree in Astrophysics, he enjoys explaining that engaging with customers in a relevant way need not be rocket science!

 David Mannheim, Author, The Person in Personalisation

David is the author of ‘The Person in Personalisation’ and has conducted interviews with hundreds of retailers. David runs Made with Intent, and previously founded User Conversion, which was acquired by Brainlabs. David is passionate about personalising, Disney, and Manchester United (sorry). 

Podcast Transcript:

[00:00:00] Vitaly Friedman: Look at us. It’s so wonderful to see you all here. Well, welcome to the panel about personalization. I know that many of you are working in the personalization space, trying to maybe do the best out of it and I’m very much looking forward to get into the deep insights about what we should, what we should not do around personalization. I’m very, very happy to have a wonderful panel over here. 

We have a couple of really, really nice people here. We have David who helps businesses mature their personalization strategies, also working on a book on the personalization paradox. That’s a paradoxical word for me, for some reason. Coming up this year, later this year, and also an expert in personalization.

We also have with us Gianfranco, literally a rocket scientist, but maybe we’ll talk about this in a moment. It is not a joke, right? But we’ll maybe talk about it later. He’s working with Natura & Co with four iconic brands, Aesop, Avon, Natura, and the Body Shop to set the personalization vision.

And then we also have with us Rasmus Houlind, Chief Experience Officer at Agillic, an omnichannel marketing automation platform from the wonderful city of Copenhagen, Denmark, also working on a book, I heard, and Gianfranco, by the way as well. I think Rob maybe, too?

[00:01:07] Rasmus Houlind: It’s a club. 

[00:01:09] Vitaly Friedman: We’ll see, we’ll see.

And we also have Rob Boland, senior Solution Architect at Klaviyo, marketing automation platform for email and SMS from London, which is exciting as well, I would say, right?

[00:01:19] Rob Boland: I’m actually from Ireland.

[00:01:21] Vitaly Friedman: Oh, alright. I see. Here we go. All the things we don’t know about you. Excellent, excellent. 

[00:01:26] David Mannheim: Are you working on a book, Rob?

[00:01:27] Rob Boland: Not yet. 

[00:01:28] Vitaly Friedman: We’ll talk about that by the end of the session it might change. We’ll see how it goes. 

So personalization, right? I think that before we even start diving into some of the things that we all care about here, I’m very kind of happy to see hopefully controversial opinions. Rachel promised a lot of controversial opinions. Whew. Let’s see how we can do that, right?

And maybe let’s start with one of the controversial questions. What is personalization in the first place? I think when we think about personalization, very often there is a lot of confusion around segmentation, customization, personalization, A/B testing, how all of these feed one strategy.

So I think that might be a good question for David. David, you like those kind of questions, don’t you? 

[00:02:09] David Mannheim: Sure. Give me the hard one. Yes. 

[00:02:11] Vitaly Friedman: So what is personalization? Why does it even matter? 

[00:02:14] David Mannheim: One of the first questions I got asked when writing the book was from this guy called Shiva Manjunath, and he’s the Experimentation Manager at Solo Brands, this big, big company in the US and he said: David, the one thing that is preventing personalization from going forward is that no one knows what… Can I swear? Well, what the hell it is? 

[00:02:32] Vitaly Friedman: No. That’s okay. 

[00:02:34] David Mannheim: What the hell it is? Someone somewhere needs to define it and needs to put it in a box. So that was my quest. So I went away and looked at all these different definitions and what I learnt was: I don’t think we do need to define it. I think when we define it, it becomes something binary. Are we personalizing or aren’t we personalizing? Really, it’s more of a scale. It’s how much we are personalizing, not if we’re doing it or not. I mean, really it’s this scale of optimization more than anything. You know, this one-to-one, one-to-many, one-to-few, one-to-one. Kind of ethereal movement.

So I don’t think we need to define it. I actually think if we define it, you basically get this dictionary definition of a bunch of static words clumped together in a sentence that creates a generic meaning. So for me, it’s actually more important not to say what is it, but why are we doing it? That’s my take.

[00:03:24] Vitaly Friedman: That would be a nice follow-up question actually. 

[00:03:26] Gianfranco Cuzziol: I think most of us here are marketeers or technologists, and I think that ultimately all we’re trying to do is to meet the needs and wants of customers, right? So what we’re trying to do is start trying, our customers wants to get from A to B, and that’s what personalization, segmentation, optimization, A/B testing, that’s all we’re trying to deliver as marketers and technologies. We’re trying to help the customer do what they need to do, and sometimes we want to do it so that it’s fun, sometimes we want to do it so that we are educating and we want to do it, as the guys from Seasalt this morning talked about, we wanna do it in a way that is right [00:04:00] for our brand.

So, again, I know we’re meant to be controversial, but I kind of agree with you at this point. I think it’s just semantics. We’re here to do what’s right for the customer. 

[00:04:09] Vitaly Friedman: Right. But you also… yeah. Yeah. 

[00:04:11] Rasmus Houlind: I actually disagree with David and I think… 

[00:04:13] Vitaly Friedman: Keep going. Keep going.

[00:04:15] Rasmus Houlind: I’m gonna kick your arse! No, I’m just kidding. 

No, I think we do need a definition of personalization because what I find is that there are different sort of tribes of marketers that are claiming and indeed they are, not only claiming, but they are working with personalization, but they’re working with it in very different angles.

So you’d have the ones who are working with advertising, might be doing personalized advertising. You’ll have the ones working with owned media campaigns. You’ll have the ones working with marketing automation. You’d have the ones working with product recommendations or article recommendations and such. And they’re all working with personalization. 

So, I do agree that the definition will be rather broad, but I think there’s a point in having it broad because everyone who gets a paycheck from working with personalization, you can’t really go and tell them that they got it wrong or that they’re not working with personalization or whatever.

So I think if we can agree on a very broad definition of personalization, that will include all the people who are working with personalization, so we won’t alienate anyone. We can finally start figuring out: ”Okay, so there are different types of personalization”, and I agree there are different sort of degrees to which you can personalize, but I still think we need the definition or else we’ll sort of never get along. We’ll just be sitting in each of our corners saying that we are working with the right form of personalization. 

[00:05:30] Vitaly Friedman: But I think it’s also interesting that I want to kind of build up on top of Gianfranco was saying, because when you think about what personalization ultimately tries to achieve is to help people get their work done, basically. That’s it, right? And, of course, find the right – okay, I’m trying to tell them to find it. That’s probably a dangerous path here giving that you are personalization experts, but basically giving people some sort of way to get where they want to get. And this is of course something that’s very, very typical.

But one thing that Gianfranco was saying is give them the kind of a way to do it well. So how do we then do personalization well? If we maybe not necessarily have different views about how to define it, what makes a good personalization strategy? So that’s actually a question I wanted to bring to Rasmus first.

[00:06:11] Rasmus Houlind: Right? Thanks. Well, so a good personalization strategy. I think you need to look at the customer journey first. So imagine the customer journey be like one large ribbon. I should have brought a ribbon. I promise I’ll do it next time. But imagine your customer journey is one big ribbon, like from the moment you meet a customer until they don’t want to talk to you again for whatever reason.

And throughout many places in this ribbon, it will be some of these places will be like really important moments of truth in the customer journey. Places where you want to win their attention. You desperately want to get them from one end of the ribbon to the other, and basically what you want to dare to do there is make sure, first of all, that the ribbon doesn’t break, and make sure that if it breaks you not only stitch it together, but you tie a beautiful bow. A beautiful bow tie. 

So it’s the main model of my book ”Hello: $FirstName” is actually the bow tie of personalization. So you have the insights on one end, you have the content in the other, and you put it together in a way that it’s doesn’t only sort of fix and remove friction from the customer journey, but you may get pleasurable and memorable and obviously personalized for the end customer and if you’re able to tie those bows on the most critical parts of your customer journey, then I think you’re onto something good. 

[00:07:24] Vitaly Friedman: Okay. Also, maybe Rob building on top of that, I know that you spent quite a bit of time working or thinking around this idea of personalizing at scale and also kind of defining that strategy at scale, which can be quite challenging. Can you share some insights on that? 

[00:07:37] Rob Boland: Yeah. One of the little tricks I’ve found recently is stop trying to personalize for everyone because you can’t do it for everyone, but turn it the other way around and let the customer personalize it themselves. 

So a little example that I have for this recently is: there’s a lovely, a lovely jumper company we work with. They’re called Sheep Inc., and they basically make jumpers out of wool from sheep. But every product that you get, they make jumpers, hats, t-shirts, that have a sheep tag on them. So anyone who’s ever been on a farm, they’ve seen probably what a sheep tag looks like, but it’s got a QR code on it. 

So when you buy your product, you get it in the box and you scan the QR code and that brings you to their site, where there you start to personalize everything about your relationship with that brand yourself. They’ve turned into a bit of a game. They go pick your sheep in the field and they have like a real-time GPS trackers on sheep in fields that they have in New Zealand. I’m not making it up. Go buy their stuff.

You pick your sheep, they show you this picture of a sheep. You give the sheep a name. So I call my Molly. And then they send you updates about Molly, the sheep. So you feel so connected and so attached to that brand, they haven’t had to do anything. I’ve actually given them all of my personal story that I want, and they’ve even turned it into it like, I don’t if anyone here remembers Tamagotchi, those little Tamagotchi games? 

They made a Tamagotchi game where you feed your sheep and you can give them like grass, and mushrooms, and carrots. And I’ve so engaged with the brand and it’s so personal to me. It’s my sheep. They’re my clothes that are coming from this brand. It is really clever where they haven’t had to try and figure out what you want. You’re actually telling the brand what you want to get directly. So I think it’s, it’s a really cool way of looking at it. It’s something different. 

[00:09:35] Gianfranco Cuzziol: Right. And I think that’s really nice because it kind of takes that customization, but the brand is learning from the way that you are customizing yourself, personalizing the experience. Then, as long as the brand remembers that and learns from that. 

So, I stayed in the hotel last night and when I checked in, I was a bit annoyed because they hadn’t remembered that I’d been there two years earlier, but that’s just high expectations these days. But when I checked in, they said, well, what kind of room do you want? Do you want a lower room? Do you want a high room? Do you want a high room with a city view or a courtyard view? And I picked a particular room. I wanted a city view. I didn’t get a city view, but there you go. That’s, that was one failure. But the key thing now in terms of personalization is that – will they remember that the next time I check into that hotel, that’s the type of room that I want. So that should be to the fore when they give me the choice at a check-in. So I think that that the idea of learning from customers and letting them choose, I think is really nice, a really important aspect of personalization. 

[00:10:30] Vitaly Friedman: Well, I guess we’ll find out when you stay here next time. 

[00:10:32] Gianfranco Cuzziol: Absolutely. 

[00:10:33] Vitaly Friedman: Yes. So that’s, we know where are you going to stay next time for sure?

[00:10:36] David Mannheim: It’s in the Travelodge. Was it? We’re staying in the Travelodge. 

[00:10:38] Gianfranco Cuzziol: No, I went up market. Premier Inn. 

[00:10:41] Vitaly Friedman: Also maybe building up on top of Rob was saying, I think it’s interesting that there are this emotional connections which are absolutely incredible. Like sometimes you might remember, or think about the brand without even remembering necessarily what it looks like or what you bought from them, like years later because of those stories. 

But now comes the part when it’s really becoming interesting because at one point we want to use personalization to drive something, and usually it’s going to be business KPIs: it’s probably going to be profit, it’s probably going to be conversion, it’s probably going to be the sale, right? All A/B tests, whatever you want to kind of to see there. 

So how do you then play this long game, without playing the short-term game? Because very often, in my experience, you get to this point where you run an experiment, an A/B test, and then, oh we could do that, but it doesn’t bring the revenue – so, sorry, let’s dismiss that and move on, right? So very much it’s still driven by this idea: does it work? Did it work in this like two weeks of testing or one month of testing or so how do you bring this emotional component to it? David, I know you have very strong opinions on that. 

[00:11:42] David Mannheim: Yeah. Well, so I wore this t-shirt specifically for that question today.

So it’s Scrooge McDuck. I’m a big Disney fan and it’s Scrooge McDuck basically looking at his money. And you think the Disney brand does this guest experience that’s magical and beautiful and you get to meet Mickey and it’s awesome and you know, it’s one of my favorite things, but really they’re nickel and diming the customer, every time you go, short term as mentality to try and draw, everything’s focused on margin. 

And I really like your example before Rob, and I’ve heard very few examples like that because I would like to ask like, what is the attribution from that thing? How does someone get that level of freedom to go away and do something that’s all about the guest experience and in theory, not about the immediate short term – its revenue.

So you look at examples like Netflix. Netflix, you know, they were cited as saying personalization for them was a 10-year leap of faith. Because their main attribution was one of retention, not one of acquisition. You know, we live in a world where more is more, where acquisition is sexier than retention, where VCs are looking over our backs to say, we need more growth on last year and last year and last year.

To answer your question, in my mind, when you’re trying to combat against that immediate, short-term, it’s overly commercial viewpoint, it all comes down to one thing, and that is purpose. Why are you doing what you’re doing? The reason why Netflix personalizes is different than the reason why the New York Times does, when we look at commercial entities, you know, retailers, they’re in the market to make a quick book.

So is it really a question of adding value to customers or exploiting value from customers? 

[00:13:27] Rasmus Houlind: I have a comment to that. So I believe that it’s always a matter of value creation from the company’s point of view, or you wouldn’t be doing personalization to begin with, but then you can argue whether we are thinking short-term or long-term.

 And I find these sort of the sheep tactics, as we all relate to it as the sheep tactic now, that would definitely create some like emotional feelings towards the brand. We can also have people like give up data that we can then later use to personalize and we can send them updates. We have an excuse for sort of being in contact.

I think that probably won’t lead to short-term conversions, but it definitely will. I think. That would be my idea that it would lead to long-term customer lifetime value. Because it’s more like of a branding exercise really. You could also discuss whether it’s strictly speaking personalization, but we promise not to do that, I suppose.

[00:14:12] Vitaly Friedman: But how do you actually measure it? Like, I mean, you need to be in a long gaming, like you have to be interested. I mean, a good strategy in general is both short-term improvements and also long-term vision of what we want to be. How do you really measure this emotional impact of personalization? 

[00:14:29] Gianfranco Cuzziol: Well, personalization is just one part of the overall marketing strategies that you’ll have as an organization. And I think that, ultimately, it’s about showing that there are other metrics you can put into play. You can use to convince different parts of the business that personalization, customer experience etc, drive those metrics. And those metrics might be NPS, they might be word of mouth, they might be customer retention numbers, or customer churn figures.

But ultimately, it’s about playing those numbers in the right way so that you can, you know, practically, you need to convince the CFO that these numbers shift the dial in terms of revenue because ultimately, you’re not gonna get the budget that you need. And purpose is right. It’s about playing the long-term game and saying that actually this is about what we’re gonna be doing in 18 months, three years, five years’ time. But also showing the small, the short-term victories to make sure that you can prove that even at small scale, you are delivering value to the business. 

[00:15:26] David Mannheim: Yeah. Thing is about confidence. So a question, is it more important to know where the direction that you’re rowing in or how fast you’re rowing the boat?

So it comes down to attribution. I don’t think there’s one single metric to define that. I think we’re talking about relationships at the end of the day. How do you attribute your relationship? You know, it’s about familiarity, acknowledgement, friendliness, trustworthiness. 

One of the ways to Seasalt’s point before, one of the ways that you can give confidence to the CFO is through experimentation. So how do you do that? Well, if you do, if you experiment on an individual by individual tactical basis, then you’re most likely looking at very tactical, short-term metrics like, like conversion rate or revenue. So you can collate all those things and you can do what’s known as a backtest.

Victoria’s Secret did this about a year and a half ago, and they collated all their personalization experience. Set 5% away for control of no personalization experience. And hey-ho, they spent 12 months naval gazing just to find out that they made an improvement, but it gave them something to put on the line, on the spreadsheet as a line item to say it gave us an improvement of X, and I understand why we do it. We now can allocate resource to that and understand the return and investment and yada, yada, yada. But there’s a bit of a rant to basically say it’s a combination of metrics. 

[00:16:39] Vitaly Friedman: Right. Rob?

[00:16:41] Rob Boland: I think sometimes people look at the wrong ones. Very often it’s, okay, did sales increase? Yes or no? And that’s kind of, people get laser-focused on sales, but sometimes there’s, there’s other channels that you should be looking at, like a very, atypical one that everyone forgets to look at when they’re looking at [00:17:00] sales they go, ”Oh, sales didn’t go up”. 

Did you check if your return rate went down? Are customers actually keeping your products for longer now? Are they enjoying them? And it’s very easy to forget to check the other stuff that you’re meant to. Like every test has a reaction somewhere else. 

So while you may have constant sales, but you launched this beautiful post-purchase flow or journey, and suddenly everyone’s keeping the product and they don’t return anymore. Or has your referral program suddenly gone through the roof? So don’t forget to look in the other places instead of just focusing on sales all the time. 

[00:17:37] Vitaly Friedman: Yeah. So I think I agree, having this holistic overview from the business perspective about all the things that we actually care about, right? And then looking at all the different facets of that experience, the customer experience, the financial revenue, the revenue, and things like that.

That’s really, really critical. But we still need to organize it all in some way. So on the one hand, let’s imagine we want to personalize and want to personalize better. Oh, well, like that also requires a team that requires some skills, that requires some setup, that requires a buy-in from the top to experiment, right? To do all the things that we mentioned. So what would make a good setup or what kind of skills would be required to run a good personalization? Rasmus, I think you have some thoughts on that.

[00:18:13] Rasmus Houlind: Yeah, I do. Well, first of all, I think they are sort of three main things that you need to consider. First of all, which people and skills do we have? And you need technology, which is also why we’re sitting here. I mean, you can’t, there’s no marketing without technology anymore. Almost. I mean, obviously like super clever and creative. And third, you would be needing to, to put, those things into the structure and apply the right governance.

And there’s sort of three maturity levels within these three disciplines. In my book, I put into like sides of a pyramid, the backend of the pyramid. What is it that supports the actual personalized communication that is meant to make the big difference? 

And your question was, which kind of skills do you need? Well, it depends, because if we’re talking about personalization, there are many ways of working with personalization. So you can choose to have like an like website-only perspective on personalization. And in that case, you’d be needing software like Klevu, for instance, or you’d be needing other software similar, working on platforms, and you’d be needing people who are specialists in working with UX and CX on the web platforms or in apps and inbound platforms. 

Whereas if you’re working with personalization within marketing automation, you’d be working with email marketing and outbound marketing, push notifications and text messages and such. A whole different skill set, more outbound mindset to work with, or if you’re working with an advertising person… So there’s really no simple way, depends on where you are, which department are you in? You’ll be needing different skills depending on that department.

But once you get to a certain level, the funny thing is on the top of the pyramid things start to merge. So what drives the highest value if you go with the errand that won’t upset the customer. So the call to actions that you’d be suggesting will be equally in favor of the end customer, as it will in the favor of the organization. You’ll be working more towards sort of a next best experience approach, and you’ll be working from an omnichannel perspective. So in the end it merges, but I think personalization starts in the different inherent marketing disciplines. 

[00:20:12] Vitaly Friedman: Okay. Yes? 

[00:20:14] Gianfranco Cuzziol: No, I was gonna give just a practical example of how we’re, we’re doing that right now. And the way that we’re setting it up is, is actually because we’re looking at it from a omnichannel perspective, we’re setting up a essentially a global team which originally I was gonna call the global omnichannel activation team, but then I realized it, that goat, yeah, that, I wasn’t gonna do that. But it’s about, but it’s about having the central team that actually looks at the customer. 

[00:20:38] Rasmus Houlind: They could become the scapegoat. 

[00:20:39] Gianfranco Cuzziol: That’s the problem, yes… They’re not gonna be, they’re not gonna be called goat. 

So it’s about creating that central team, and at the core of that, it’s having someone who own the customer, but also having someone who looks after the technology perspective. Because I think that having that dual focus within that global team enables you then to kind of, to put the customer first and try to [00:21:00] understand then what are the right technologies that will support that customer vision that you’ve got. So, technology comes second. It’s the customer first. So think of it that way. And then you build out the teams sitting beneath that, depending on what channels, you’re using at the time. Yeah. 

[00:21:17] Vitaly Friedman: And also think about customers first. I think I cannot have the session about personalization without thinking about one thing that’s been kind of very prominent over the last couple of years since the notion of privacy.

So by the way, hello everyone. My name is Vitaly. I’m working with European Parliament. It’s not a joke. And privacy is very, very important. And obviously collecting data is a story of its own. 

And very often what you see, and this is kind of one of the stories I found, I think I read in the New York Times, I think recently, where the idea was, well, there was a company that was actually customizing the offboarding experience to make it more difficult for people to leave, to cancel a plan or to leave kind of close an account based on their preferences, so make it more difficult to leave. That sounds to me like an unethical decision in many ways, and a very deceptive practice, but there is no regulation against that

So I’m wondering where do we find that balance. We want data, obviously personalization partly at least depends on having the right amount of data about the customers, but also being respectful about just the right amount of data and not kind of having this kind of deceptive patterns at all.

I know again, David, you have a very strong opinion about things. You have a strong opinion on that, don’t you? 

[00:22:26] David Mannheim: What comes first business or customer? I think that’s just all it really comes down to. There’s a really good example that I like to often cite and it’s, you know, we wouldn’t be in a personalization panel of, we didn’t mention Amazon.

But you have, you know, Jeff Bezos, the pioneer of personalization, arguably, who’s like customer first, customer first! If we have 4.5 million customers, we shot up 4.5 million different stores. He’s also the chap that said, your margin is my opportunity. And you know, considering they’re the pioneers of recommendations as it were, you know, it’s how YouTube even 20 years on still use their recommendations algorithm.

They just got fined 900 million pounds by the EU commission because of their practices of putting higher margin products in certain recommendations areas rather than what is actually best for the customer. So to go back for the use of data, for me, it always comes back: it’s customer first, not business first.

And I actually think these regulations are a bloody brilliant thing, because I think it forces us to be more creative in our practices to do what’s actually right for the customer. Not what is right for the business, gone to the wild, wild west days of the cookie. So I think it’s great. 

[00:23:34] Gianfranco Cuzziol: I think that when I had the opportunity to work at group level for a year or so, one of the notions that we sold into the various brands was the fact that, think about, think about the fact that data, customer data might at some point sit on your balance sheet, right?

A brand might sit on your balance sheet at the moment, but think about how the quality of the data might sit on your balance sheet in five years time. So actually the [00:24:00] rules, regulations, guidelines that are now in place. If you go above and beyond those, and that was always our intention,, not to set a minimum level in terms of trust, ethics, and privacy.

But imagine in five years’ time, the quality of the data, the quality of the consented data, engaged data that is sitting on your balance sheet is worth US dollars, pounds, euros, whatever it might be. It frames that conversation in a slightly different way so that the business does think about the fact, oh, actually, you know what shit we should be… Sorry, I did swear then, didn’t I? 

[00:24:30] Vitaly Friedman: That’s alright. 

[00:24:31] Gianfranco Cuzziol: You think about it that way and then the business does start to think about the quality of the data and the value and the consent levels that you have in that data. 

[00:24:40] Vitaly Friedman: Right. And of course, kind of related to that, I think we have a lot of conversation also today in the first session about ChatGPT and the future of AI in general kind of be really defining our future in many ways, right? Obviously, we have to bring it into personalization as well, because we have so much data we can do so much with just so much data. The question, however, is… This is something that may be very personal. Sorry for a personal question, but are we going to get in a position where things will look a little bit too strange to customers? Like, this is too personalized for me. They know maybe a little bit too much about me, this uncanny beauty of personalization. Are we getting there already?

Like sometimes you can’t really tell is it a personalized messaging that I’m consuming or is it just a general messaging that I’m consuming? So are we getting to this position already and how do we get out of this? If we are. 

[00:25:28] Gianfranco Cuzziol: Well, I think you’re just trying to find that level of creepiness and you don’t want to cross the line in terms of creep. I remember once getting an email from, this was quite a while ago now, and they don’t do it anymore, but I got an email from, I remember it was John Lewis saying that we noticed that you spent a bit of time on our bedroom department online.

And I kind of went, I have, but just that’s like, just a bit too much. And so it’s just trying to find the right balance on what’s right for customers, but also using personalization in the wrong way. 

So everyone talks about the Spotify Wrapped bit of personalization, where they tell you about the music you’ve been listening to over the last 12 months, and I think Spotify are great in terms of the way that they then give you live music recommendations based on what you are listening to on the app. 

But I then remember that Caffe Nero then started sending me an email telling me about the amount of coffee and what coffees I’d been drinking over the last month or so, and I start thinking shit! Do I really drink that much coffee and it actually stopped me drinking coffee or you know, less coffee? So you kind of need to think about why you’re doing this personalization. Is it just because you’ve seen it somewhere else as you think it’s a really cool idea or is it really adding value to the customer experience.

[00:26:36] Vitaly Friedman: Right. 

[00:26:37] Rasmus Houlind: We were discussing creepiness, and I think there are certain things that you can do to steer clear of creepiness and not end up in the uncanny valley. But first of all, I think it’s important to say, most companies are no way near any kind of uncreepy valley, because they don’t have the data and they just get started and they’re basically just doing “Hello, $FirstName” personalization in emails.

[00:26:56] Vitaly Friedman: Or maybe they have data but they don’t what to do with it.

[00:26:59] Rasmus Houlind: Exactly. So what tends to drive the sensation of creepiness, if you’re using data that the consumer isn’t aware that you have on them. So if they weren’t aware that you were collecting browse behaviour from the website, for instance, say that you were browsing Victoria’s Secret and you were looking at a fairly intimate underwear, and suddenly they started addressing that to you’re in well-meant retargeting emails.

You looked at this super kinky outfit, would you care to see more of those? That may come across as a bit too creepy for some because they weren’t aware that you were collecting this data. 

Or what you could do instead is to, I think, distinguish between explicit and implicit personalization. So maybe not say directly, make it completely evident to the customer that you are using data, make it seem as more of a coincidence that now you are communicating about these fairly kinky outfits, but it’s just a coincidence. So you’re not showing it directly that this is an act of personalization, it’s just random. To some extent. 

This may be a bit grey hat in this area, but say that you are addressing topics that are touchy like divorce, for instance. So one of the expert panels for the group was a pension company talking about how they could collect website data and see through an algorithm who was most likely going through a divorce and addressing that explicitly super dangerous. Addressing it implicitly or maybe even getting people to supply data like from a zero-party data perspective that ”Yes, I am going through a divorce”, would open up for really interesting conversations between the pension company or the bank or whatever.

So I think the topics, if they’re touchy – be more implicit. If you’re using data like that people’s underwear you’re collecting, be more implicit or don’t do it at all. Because you’ll be too much. 

[00:28:51] Vitaly Friedman: And also maybe, probably just being polite about those things instead of just making assumptions about what people are like, what they’re going to think and all that. Just maybe asking as questions.

[00:28:59] Rasmus Houlind: Assumptions are super dangerous. 

[00:29:00] Vitaly Friedman: Yeah. I think, Rob, you wanted to add something as well? 

[00:29:03] Rob Boland: Yeah, I was just going to say kind of, I think we hit the creepy phase like 5-10 years ago and we grew out of it. As a humanity, it’s a big win for us, I think. 

Everyone’s probably heard or heard some variant of Tesco. Tesco Clubcard where people would scan as you shop. And a father found out his daughter was pregnant before she told him, because Tesco was being a bit too creepy and sending him all these club card offers for all these baby stuff, and he had no idea what was going on. So I think brands learned a lot from that.

And we actually learned there is this level of too personalized, too creepy, and we kind of scaled it back a little bit to be a bit more helpful, as opposed to being creepy, I think. 

[00:29:49] Vitaly Friedman: Right. And Gianfranco, I think you also wanted to add something. 

[00:29:51] Gianfranco Cuzziol: Oh no, just on the back of using data for recommendations and my wife bought me for Christmas, a book on Amazon using my account [00:30:00] called ”This is how your marriage ends.” So I have been getting some weird recommendations on Amazon. 

[00:30:09] Rasmus Houlind: Talk about divorce. 

[00:30:10] Gianfranco Cuzziol: Oh yeah, a little bit. 

[00:30:12] Rasmus Houlind: I think it’s a good point, Robert, and it was a Target example, so if you’re looking for it was Target from the US that had this pregnancy example.

But what they did, the solution to this is not to annoy any more fathers by predicting pregnancy with their teenage daughters. What they did is that they actually toned it down so they would still present these products like there was an algorithm saying that these are the right products, but they would sort of not make them sort of the main story of the message.

So instead they would, like, in the hero position of an email, for instance, they would put bananas or something like pretty less dangerous product, but the pregnancy test would be immediately beneath, but it would appear as if it were a coincidence. So that’s kind of how they solve that. So making it less obvious.

Also, you could exclude products. So for instance, we’re working one of the biggest health and beauty retailers in Denmark. They’re excluding [00:31:00] products that they feel would be offensive. So, for instance, if you have like a teeth, teeth prosthetics, then you need certain glue for that, apparently, and it’s filtering that out. 

They’re also selling vibrators and they would never choose to have those products in an email, for instance. So for outbound communication, they’re totally filtered out. 

[00:31:20] Vitaly Friedman: Okay. Okay. Gianfranco? 

[00:31:22] Gianfranco Cuzziol: I was just saying, the thing is with AI and the way that personalization is going at the moment. We’re becoming really super clever about this, to the extent where we think we know what the customer wants before the customer thinks that they know or they need it. And this is maybe a little bit philosophical, but you know, part of that is kind of going well, are we actually removing free will from customers? Right? 

And are we just using our own devices to move customers down a particular route that we want them to go down? Or, you know, where’s the serendipity in finding new products now? So if we’re being super clever, are we actually doing the customer an an injustice? 

[00:31:59] David Mannheim: I think it’s more than that. So, I completely agree. There’s the ethics, there’s the echo chambers and filter bubbles that we’ve all heard off. Mark Zuckerberg once said that people may be more interested in a squirrel in their back garden than people dying in Africa. There is all this ethical debate around it. Do not get me wrong, but I also think AI in general is a science. It’s a pure science. 

Whereas personalization is all about being more personable. It’s a verb to be more personable and that is a human connection that is all about relationship and connection. When we think about it as a marketing principle, when we think about market as a whole, there’s this combination between the art and the science and slightly more ethereal, like you said. I don’t know whether we are tipping the balance more in favour of science than art.

[00:32:44] Gianfranco Cuzziol: I mean, if you guys have ever been into an Aesop store when you walk into an Aesop store, you get an amazing experience. The consultants there are fantastic. They’ll spend 30 to 40 minutes with you at the sync trying to find the right products for you.

And our challenge at Aesop was not to try and replicate that. Because you can’t replicate that human experience. It was trying to find a consistent approach so that once you left the store, we could talk to you in a way that was very Aesopian without becoming, without becoming Kiehl’s. I love Kiehl’s, but the number of emails I get in my inbox that are offering me discount after discount after discount is just excessive.

But it’s about, you’re right, it’s about maintaining that human touch as part of all this, and it is finding the right balance between art and science.

[00:33:30] Vitaly Friedman: I think at this point, just before, I’d like to open up the conversation for the wonderful audience, maybe just to wrap it up, I would love to get maybe one final word of wisdom from you if you wanted to say, to send a message to the wonderful people attending here.

What would that be around personalization or anything really, that really matters to you at this point. David? Just a quick one sentence. 

[00:33:50] David Mannheim: Personalization is the verb to be more personable. So be more personable. 

[00:33:54] Gianfranco Cuzziol: In the words of Mike Tyson, everyone has a plan until they get punched in the mouth. So it’s an evolution. You’ll never have the final solution from day one. 

[00:34:04] Rasmus Houlind: I think my message would be watch the space of generative AI and explore how that can help you create more content variance more easily, or even make that as a democratized process for your employees making that happen. I’m so curious to see how that’ll evolve. 

[00:34:19] Rob Boland: I always say this stuff is really hard, so don’t hit yourself. If you’re not doing it well, it’s actually really hard to do it, but don’t be afraid to try and do it. Think of my little sheep. Think of Molly, the sheep in the fields. 

[00:34:31] Vitaly Friedman: We always will. We always will. This is never going to change anytime soon. 

[00:34:35] Rob Boland: And just imagine your brand. Try and do that for your brand, that experience.

[00:34:39] Vitaly Friedman: I think this might become like a hashtag or something. We’ll, we’re just going to start a movement here. I’m sure that sheep personalization. Alright. Thank you so much for the wonderful replies.

But before we wrap up… Sorry, we have running out of time? No, we have time for questions. Right? Do we have any questions in here that you’d like to bring? My wonderful colleagues will be happy to throw you a microphone. So I think we have one question all the way on the top over there. Throwing would be a little bit too difficult I think at this point.

Oh, here we go. Oh, okay, that worked.

[00:35:10] Audience member: Hi. Hello. Hi. Mic check. Thanks for that. That was brilliant. Great panel there. I guess my question is, how we take personalization and also take into account like generative AI and how things in a way of becoming a little bit more monolithic. You type in how to write an email marketing newsletter and you’ll kind of get very similar responses a lot of the time. And it’s how do we make the personalization adapt to the multifaceted nature of like human beings. 

So you talked about the Travelodge, sorry, Premier Inn an example. You wanted a city view, and the next time you go back it’s got, he wants the city view, but you know what? You didn’t really enjoy that city view. You want the Southbank view, you want the car park view this time. I guess it’s like, maybe not, but I guess it’s like how do you take into account, how does personalization take into account that we are constantly changing people as well?

[00:36:03] Gianfranco Cuzziol: So I think that as brands, we have opportunities post that experience to try and understand whether it was the right experience for you. So again, going back to the Aesop example, when you go through that experience in the Aesop store, you might buy three products in there, and one of the first things that you get is actually we don’t try and sell you more products.

We try and get elicit some experience from you, after about 30 days, to understand if actually those three products that the consultants spent half an hour talking to you about and ultimately you buying. Were they the right three products for you, and if they weren’t, come back into us and we’ll readdress that for you.

This is why I think it’s with customers, it’s an iterative process. You are constantly learning, constantly getting information from different touchpoints to understand what is right for that customer. 

[00:36:54] Rasmus Houlind: But I think there’s a good point that you’re making actually, that we will never actually know what’s going in the real customer experience because the customer experience belongs to the customer and whatever they’re doing in their life, whatever else is going on, we can’t really know for sure what their intention is. I mean, we can be less wrong. And basically, I think that’s what personalization is all about.

So we can suggest things that we think are less irrelevant, and sometimes we have better clues, we have better insights, and if we’re dead sure, like for instance, if you have like explicit data from a customer saying this is my preference or whatever, then you can choose to be like really explicit or you bought this very product. This is how other people are making the best or the most of it. 

So we’re working with a retailer in Denmark. They’re selling health and beauty, and the moment that you buy into a new category, for instance, the follow-up that you’ll get. I mean, instantly you are in a like a suppression list for all campaigns, and instead, you’ll be receiving helpful communication about how you apply this new type of makeup or skincare or whatever.

And actually, in terms of long-term and short-term revenue also you’d expect this to sort of benefit the long term and the customer relationship. But guess what? It also impacts the short-term revenue as well, which is a surprise to most people involved with this. Plus they can sell it as retail media to their suppliers. So it’s like a win-win-win kind of situation. 

[00:38:10] Vitaly Friedman: One more question. No, no, unfortunately. Well, I’m sure that while our wonderful speakers are not going anywhere, do you? No, just hanging out here. So if you have any questions, please come. Yeah, have a conversation, chat, coffee, water, anything. They’re very much available to you and thank you so much for the panelists for being here and to you of course, for joining in and being a part of this. So thanks everyone.

More Resources