Customer Strategy: It’s Personal

Hello, I'm Georgie Frost. Before we hear from our guest on this week's episode of The So What, I want to take a moment to highlight a fantastic new video series from BCG that I think you'll really enjoy. It's called CEO Moments of Truth, and it's hosted by BCG's own CEO, Christoph Schweitzer. Join Christoph as he talks to top CEOs about the most difficult decisions they faced in their careers so far, how they navigated them, and how those moments shaped them into the leaders they are today. They share their fears, thought processes, reflections and lessons from the most challenging of situations, something Christoph did too when he joined me on this podcast recently, revealing his own moment of truth. You can watch CEO Moments of Truth on BCG's YouTube channel or on our website at BCG.com/MOT. Consumers don't just want more personalized experiences. They expect them now. For businesses, personalisation is a competitive differentiator, building trusted relationships with customers and giving leaders a serious advantage, but only around 15% of companies are getting it right, leading to wasted money, squandered effort and missed expectations. So what is the key to success? Whatever industry you're in. I'm Georgie Frost, this is the so what from BCG? In the age of AI, brands are increasingly competing on digital customer experiences, not on manufacturing scale, but the scale of insights they have about us, and the speed with which they are able to act on that with AI and technology. Today, I'm talking to Mark Abraham, global leader of BCG's personalisation, business and co-author of personalised customer strategy in the age of AI. In totality, we estimate that personalisation leaders will capture a $2 trillion prize in incremental growth over the next five years. On the one hand, there is at least one pioneer in every single sector in every geography, so people are starting to crack the code on how to scale personalisation. But at the same time, as you said, only 15% of companies are in our estimation actual personalisation leaders. So there is still a lot to do. So what does personalisation leader mean? What does it look like? What does good personalisation look like? And what are the rest, the 85 majority? What are they getting wrong? So take four seasons, for example, if you go back to them, and you've been there before, you get the handwritten note that acknowledges you as a customer and what you might be interested in on this trip. Those personal touches are just as important as the digital experiences that Netflix or Uber are delivering as well. And then what's exciting is that those two worlds are merging. There is no such thing as just digital or physical experience. People are navigating across channels and companies must follow them. You can't be doing personalisation just to sell more stuff. You need to really understand what customer need, what customer problem am I trying to solve? So let's take a favourite example of mine, Spotify. As a music fan, I love the personalisation that they're doing. They know they want to empower me with listening to the music I want, whether it's working out at 5 a.m. or entertaining my friends on a Friday evening. They know me, they don't just look at my data logging into the app, but they have tagged every piece of music by genre, by category, and they're using all of that metadata as well as how long I listen to things, which things I skip, to create a profile of me. Next, they reach me in the app, they'll surface playlists, but also they'll nudge me if there's a new release, or my favourite band is coming to play at a concert in town. So they have this vast library of music content, and that's why they've been able to do this. They have actually more things in their library than people could possibly listen to, and they're curating that exact next best song for each and every user. I think that's what companies need to be moving to, especially in the Jennyi world. Can I create an order of magnitude more content that I can then curate and show exactly the right onto my customers? And finally, Spotify delights me. It keeps getting better. They run hundreds of experiments for users in their app to make little improvements so that every time my experience gets better. A concrete example of this is DJ Xavier, a Jennyi powered DJ. They've beta launched now and are scaling. And that DJ actually adapts to your likes. It explains what song is coming up. It mines what you loved listening to last year, but you haven't listened to in a while. So it's a next evolution using Jennyi of their personalisation technology. What about legacy companies? In traditional industries or just small businesses? One of my favourite examples is from wine, the wine industry. And actually a couple of retailers in the US and in Australia are doing this really well. So total wine or more and endeavor. They're wine retailers in those countries. And they use the same kind of algorithms that Spotify or Netflix might to figure out what next wine variety you might be interested in. Whether it be just for a casual Friday evening or a dinner party you're hosting. And then they contact you in the right channel. They might surface it to you in an email. They might talk to you about it when you're in the store. They might have brief educational videos tailored to you about it. So they're using all of these channels and this content to surface and become basically a Spotify of wine for you. So that's just one example of how a category can really take this and implement the same kind of principles. You can't get much older than the wine industry. What about bricks and mortar businesses though? Can they do anything here? Absolutely. Take restaurants for example. Starbucks famously was built all about the human connection between the barista and the customer. And I know things are getting really cramped in the store with all those folks coming in for their TikTok beverages. So personalizing the app and their personalized offers has been a really important way for them to help customers navigate their menu. Suggest new things you might try out and also give you a great deal for coming into the store and finding what you like. But what's interesting is in restaurants, newer brands that are also built around cafes and actual stores are already building this from the ground up with personalization in mind. One great example of this is sweet green, the salad chain. You know, they only started within the last 20 years, but they built with the digital app from the start. And so they're bringing again personalization and challenges to try out new parts of the menu and get offers as part of the experience from the beginning. What role is Gen.A.I. and currently playing what role will it play? Do you foresee? There's a lot of confusion about AI right now. Personalization is the best way to use AI to drive growth. And I think there's been too much discussion about using AI to cut costs, drive efficiency. And there's actually two fundamental types of AI, predictive AI, which has been around for a long time. It's for example the AI that Netflix uses to score you on all the kinds of shows that you might like. That's been around. It's really important and you have to keep building data and running experiments to hold those models. And then there's Gen.A.I. which is very much putting this all in the hands of customers and asking them to tell the AI what they want and then surfacing solutions, which initially has been chat, but you're going to see much more voice video, all sorts of modalities in that. And the exciting thing about Gen.A.I. is it can combine different systems and write code to execute commands. So what was previously many clicks to do for a customer can become quite simple. And both solutions, both types of AI have to play together, but it's not just going to be a takeover of the machines. We also have to pair that with the human touch. And what I mean by that is guardrails and rules. So brands need to establish brand guardrails within what parameters do I want the predictive AI to operate and optimize. Within what brand guardrails do I want the Gen.A.I. creating my content. They also have to think about interesting rules like anti-repetition or anti-conflict, which things do I not want to show you multiple times or together. Spotify is a great example again, like if I keep feeding you the same music, does it become an echo chamber even worse if I'm a news organization. And how do I want to set guardrails around that to not create these echo chambers. And lastly, responsible AI. So how do we think about the guardrails to make sure that we don't introduce bias into the models thinking through step-by-step from the board down to the working team level. Take an example of personalized offers. It's been around for a long time. It's a big use case and personalization. But one of the ways we have to test and pressure test for bias is are we giving the less rich offers to our lower income consumers because of some quirk in the AI. We need to be looking at quality checks like that and putting in place rules to catch those kinds of things instead of just trusting a black box. Where do you see the role of regulation coming in? You spoke about responsible AI. What role will that play in business strategy, I suppose. At the end of the day, this has to be built around trust. Consumers will absolutely share their data if you can deliver a more personalized experience. So the rate goes up from something like 30% to 90% of consumers will give you their email or other similar personal information if they're actually getting those kinds of experiences. But then you also have to have transparency. So what is all the data you have about me and why are those ads for your products following me all over the internet? So I think you're going to see brands themselves publishing and having places where consumers can go to see all the data the brand has about them. But I think regulation is playing catch up. And so brands have to think ahead and stay ahead in the game of trust. Are there any industries where personalization isn't necessary or you can get away without it? You don't necessarily have to have a tremendously personalized experience to have great customer satisfaction. Take, for example, low-cost airlines that always deliver on time and don't lose your bags. They are not personalizing, but they still have great customer satisfaction scores. But on the other hand, the brands that had the highest customer satisfaction scores overall oftentimes were using personalization and many of the ones I've cited ready are at the top of that list. So it's an enhancer and it's becoming a requirement for corporate strategy. The argument we make in the book is in the age of AI brands are increasingly competing on digital customer experiences, not on manufacturing scale. But the scale of insights they have about us and the speed with which they are able to act on that with AI and technology. I ask that question because I imagine leaders having to make decision about where to put resources and money are wondering, you know, shouldn't I just be focusing on getting our great product, having a great customer service and just doing business well? So the foundation is critical. If you have a crappy brand and customer experience, personalization is not going to make it better by itself. But what we find is the personalization leaders grow 10 points faster than the laggards consistently across industries and even in places like health care or financial services or B2B that are more regulated have been more cautious about this type of personalization because of data and regulations. We're seeing that growth. Increasingly, you're going to see this at the top of the CEO agenda. Mark, do a thought experiment with me. Let's transport into the future 15, 20 years. What will my experience shopping and otherwise be like? Where is this all headed? You know, last year already we were all starting to use chat GPT at least I was to come up with gift ideas. You know, that was an interesting start. I think this is going to go much much further. You're going to see ecosystem solutions emerging. So already I can put in an example of a travel experience I want to give to someone and get ideas for it. Well, in very soon and certainly some startups are already doing this, you can book everything from the flight to the hotel to the car rental to even dining experiences and on site experiences all via a virtual assistant. So this notion of agenda AI not brands pushing personalization to customers, but customers actually pulling the personalization they need is where this is going to go. Think about commerce as well. Search is going to change. You're not going to be clicking through 15 websites for a parallel product. So you're going to be consulting a virtual fashion assistant and getting recommendations across brands of the types of products. That you might gift for the holiday season and maybe you'll buy some for yourself as well. And what's interesting is this completely flips on its head the marketing model for brands. One CMO I was talking to said within three years, about a third of my marketing budget will be about marketing to AI, these virtual assistant agents because it's a great intuitive way to search, discover, get inspired, especially in categories like travel, apparel, gifting, but even food and the like. What an interesting thought. So if I have a virtual assistant making decisions, which would be lovely, by the way, as to what I have to wear for work, for example, oh, I could do with that. Who then is being advertised to? It's not me, is it? It's my AI assistant. So then do I trust my AI assistant to make the best choice of me? How will that work? Chad GPT-5 is coming out. One of the improvements they're making is really looking at your preference history in a much more tailored way and overweighting that in the system. That's one way in which the AI will better take into account your preferences. Of course, there's, again, this huge importance in the guardrails and the rules that we have to think through. Because how much I weight those preferences and show you the same stuff will really shape what you look at and what you consider and then what you buy. So the algorithms have to be tweaked for how much in units, how much different things should I show you? And that's a really interesting part of the behavioral science of this that we have to figure out as marketers, as brands, and frankly as consumers, because I think this is an area for consumer advocacy, perhaps regulation as well. As a journalist, the area that most concerns me, well, there's lots of areas that concern me, but probably the one that sticks out is the echo chamber. We know people are getting a lot of news from one perspective. Let's put it that way, which is not particularly useful. We need to be broad, well-read human beings. We need a lot of different opinions to be put into our feeds and we're not getting that. How do companies avoid that? Because surely it's working on what you like. So you know, the idea of, and often when I'm running along to a dance track, you know, all of a sudden my spot if I will put on some classical thing, yes, I like it, but not in that context. Exactly. So this is a solvable problem with AI, plus what I call, oh, I, organizational intelligence. So really bringing the marketers, the human-centric designers to the table, along with the technologists and the AI and data scientists. For example, we worked on the next best action engine for a large airline. So they had to decide how to talk to you about flight destinations. You could be going to next selling you additional products after you make a booking, but also about things like getting you to sign up for their credit card or their travel insurance or health insurance or other offerings. So there were hundreds of things they could be talking to you about. Well, in their old approach, they would have the product teams behind each of these figure out who are the best customers to talk to about my product. And of course, it was always the highest value customers. So the highest value customers got bombarded with everything 10 emails a week and lots of ads and so on. But there's a better approach. We call it less is more you can build a next best action engine that actually scores these customers on what's best to talk about next. What's the next best conversation I like to call it to and then decide how much newness do you want. So there's actually a factor we built into our we call it Galileo algorithm that is actually a toggle for how much you're going to test new things I've never talked to you about. And it has to do as well with how likely am I to get my recommendation right. So if I just launched a new product, I really don't have any data about you or other customers like you, I'm much more likely to surface that in your recommendations. But if I really know you like this and that's the next best thing to talk to you about, I'm probably going to surface that. So there's there's a lot of techniques that we can design to deal with this. We just have to have the whole team at the table and the right ways of working to get it done. If you're just starting out on your personalization journey, Mark, what question should you be asking what should you be doing? I always say just get started. And the questions to ask are how can I get started in a simple way really boil down the use case to its core that challenges you and the team to get it out the door in a month or two. And then start testing and improving with a view to scaling it. One technique we use a lot with our clients is just taking a set of customers offline. So for example, for a large retailer, we took 5% of their customers offline and we tested a much more multi-channel, optimized, personalized approach and showed that there was a better way to do marketing in this case than a channel by channel and product by product approach. This allowed the teams to reand design the process as well as show the value that could result from this kind of approach. So it's a great way to break the business as usual silos and mindset that companies get stuck in. Mark, thank you so much. And to you for listening, if you want to check out Mark's book, personalized customer strategy in the age of AI, the link is in the show notes of this podcast. We'd love to know your thoughts to get in contact. Leave us a message at the so [email protected] and if you like this podcast, why not hit subscribe and leave a rating wherever you found us. It helps other people find us too.