The Kite Factory’s digital team recently had a debate about what area of digital advertising will see the greatest impact from AI technology. Debate winners Iona, Maria, Patrick, Emily and Richard join MP for Digital, Ben Foster, to share why they believe their choice of personalisation will feel the greatest impact from AI in the digital advertising world.
Catch up on the episode here.
Hello everyone and welcome to the Unmuddled podcast. Today’s topic we’re going to be talking about all things AI in the realm of personalization. Now. We actually had a debate amongst the digital team couple of weeks ago. On this topic and this. Was the winning area we thought would feel the greatest impact, more so than other parts of digital. So the team that won that debate are here today to talk us through their argument why it will be the greatest impacted area. So I will hand over now to Iona who will start. Talking about the topic. Cool. So I guess as a team, we went down the personalisation route because of its like vast application across all things to do with. Marketing and we’ll go into a bit more detail of that shortly. But just looking at AI personalisation as a whole, you can use different things across like machine learning, deep learning and also language processing to reshape how your brand interacts with customers, which I think will be really important going forward across all aspects of market. Thing, even the early research shows that people are way, way more likely. I think up to 91% in some cases to purchase from a brand that actually knows them. So using different types of AI learning can actually really help, like understand your customer from a purchase intent, a donation intent perspective. As well. But yeah, I think the rest of the team will go into slightly more detail of the different applications and a bit more research. My question to you on that would. Be obviously people do respond to more relevant messaging that that that is a given, but where? Where do you think the line should be drawn between something that resonates with an audience and something that people find a bit too? A bit too stalkerish, quite frankly, it’s got a bit too much information, it’s too personalised. It actually freaks them out and has the adverse effect. Absolutely. I think there’s definitely a balance that needs to be striked. I don’t think you wanna go down the route of calling people out by name. I think also, just naturally there’ll always be some crossover of someone else on someone’s phone or something like that, but just using contextual areas or different messaging of interest based. Things that people are really interested in can actually help convert them or make them more interested. In your brand. Fit. I believe, Rich, you’re up next. Yeah. So as everyone has said, personalization is all-encompassing now. It’s no, it’s no longer a nice to have and three very strong reasons from three very reliable companies. One is that consumers expect bespoke messaging. According to smart HP, 72% of consumers say they only and only engage with personalised messaging. Secondly, consumers want tailored promo. According to sales force, 57% of consumers are willing to share personal data in exchange for personalised offers and discounts. And lastly, in terms of customer retentions customers. Stay loyal with. You when? When you personalise your marketing and it’s shown when 44% of consumers. Say that they’ll become. They’ll likely become repeat bias. If they did experience a more personalised shopping experience. Great. So in terms of in terms of all of that, is there a specific area where you think that this will really come into its own, you know, see you could be talking about the marketing that had been pushed to people and the marketing that is more the poor marketing, IE the CRM type, you know their actual website? Or do you think it’s actually more in the initial promotion stage? Where do you think the biggest opportunity lies? I think there are definitely strong arguments to be made on whether you’ll have more strong impact on the website side or on the promotion side, but in terms of the more again it’s. Personalization is all-encompassing, so I think when it comes to the biggest impact, it’s mostly on promotions. For example, on websites, there are a lot of at the moment promotions that don’t necessarily attract people. It’s not enticing to everyone, so just for you to harness the data that you have as a company and tailor that per. Audience segment is is strong enough itself to to be leveraged. Perfect. So who was? Who was up next, we believe. Was you, Emily, actually? Ohh no, it was me. Yeah. So I guess Ben and Rich, as you guys just touched on the personalization element can be applied to all areas of digital marketing. So we’re not just talking about the ad that someone sees, but also the experience on the website and the experience of the of the brand as a whole. So through the likes of machine learning and deep learning, as I owner touched on before. And modelling and analysing your CRM data can lead to like personalised and tailored recommendations based on product history, even stuff such as life stage location. In terms of different products that you suggest to people, different experiences people have on the website, improvements that can be made across the likes of conversion rate optimization, beyond that in terms of the sort of ad that gets put out across the Internet, the likes of optimization of. Add copy and creative and can be used and can utilise AI really effectively. So machine learning from different ad copy variants and which ones work best for different audiences and different people. But beyond that, the likes of content creation and creative looking at some AI models that can even create images specifically for brands. So if there’s a lack of imagery or a lack of stock imagery available. AI can solve that problem for you. And then further to. For the likes of natural language models and language learning and things like assistive search can become more and more intelligent and the likes of chat bots on websites can also become a lot more useful, a lot more personalised different customers. So yeah, it’s an all compassing field really. If uh, obviously it is so broad. So if you were to get if uh clients would say, look, I’m really interested in this, where should I start first? Where is the quickest, easiest win out of those examples you’ve just given where? Where do you reckon it? I mean, Emily. Will go on to the sort of functionality that we can do immediately with the with the partners that we’ve currently got access to the likes of Google and Meta. However, I think it’s probably most important for client, for clients and businesses to make use of the data that they already have. Start modelling their CRM and and learning about the customers that they’ve got access to already. It can be very low cost or free to model what you’ve already got access to, so it’s a very quick and easy win. And you know that that’s obviously acquired with a sort of attitude of embracing it. Some may be slightly more risk averse and also slightly more cynical. They may say this is. Another sort of technology that’s coming in that that, that we all say that everybody needs, you know, wasn’t that long ago that every client apparently needed a huge DMP to be able to manage their their marketing when actually only a fraction of the market did. So how would you convince them? More reluctant, you know non early adopters to sort of start looking into this space. Essentially, it’s just a conversation. You need to have about being on the front foot. I don’t think this is something that’s gonna go away and you don’t necessarily have to embrace it whole hog and start creating all of your content through AI and all your different messaging to to customers through machine learning. But it’s something you should at least be starting to dip your toes in because this is. I think a way that certainly a light touch stage is a way that digital marketing is going to pivot to. So you’ve got nothing to lose by investigating how you can make it work for you. Yeah, completely agree with that. Hopefully I got it right this time. I think it’s you now, Emily. It is indeed, yes. So obviously, yeah, we’re all paid media specialists so that we’re most interested in how AI and personalization can kind of work for us is in, in kind of different campaign types that we have and. Within kind of the sphere of paid media, so the arrival of automated campaign delivery through and things like forms Max and advantage. Plus, have kind of been quite a big pivot for the. Kind of tech giants in the last few years that kind of allows us to show users personalised ads, and I’m kind of seeing these giants like Meta, Microsoft, Google, they’re leading the way with a lot of their AI and personalization and we expect that a lot of other kind of small channels will follow in the next few years. We’ve seen a movement. From kind of segmentation to personalization specifically, so going for more for that one to one interaction with the customer, which hopefully will enable them to feel a little bit more kind of personally engaged with the ads and bring up our engagement rates and conversion rates. We have kind of seen the results speak for themselves so far. So in testing that we’ve done for instance, we’ve seen kind of more than 800 conversions coming from performance Max, whereas other PPC was kind of driving around about the 250 mark kind of under 200 for social and round about 150 for display. So kind of really skewing towards these more. Automated and personalised approaches. I’m kind of seeing the results from our side. We’ve also seen advantage plus campaigns driving really strong Cpas compared to more kind of manual targeting. And it is kind of generally expected that moving forward, we’re going to see some really big benefits to marketing from this kind of more tailored offering. So kind of higher quality leads, better opportunities, better engagement with potential customers and hopefully just more efficient ad spend for us in general. So that we can make. All of our budgets go even further. That’s that’s great. I mean, obviously, yeah, we’ve seen some strong results in those, those new ways of buying, you know advantage plus and P. It does take away some element of the control and the transparency. Some clients, quite reluctant with that and nervous about losing that control and handing it over to the, you know, those big media owners. How would you? How have you gone about sort of trying to comfort those clients to get them to go ahead and test these new platforms and they turn this this AI world? Yeah, I guess it’s kind of about explaining the fact there is always going to be some sort of trade off with this sort of activity, so. We know that if we are kind of getting more power of tube channels, the likelihood is we will get better CPMS for instance. So we might get kind of some some more effective and more efficient traffic, but that might come with the caveat that we don’t necessarily get the same amount. Of data back. However, I think in kind of quite a changing world, clients are beginning to understand that that’s not necessarily always going to be something we are able. To provide now that we’re kind of in this post iOS change world where we can’t get a lot of the data back that we used to anyway. We’re not kind of losing out as much as we would have done maybe four or five years ago with the amount of insight that we can provide on the people that we are reaching and engaging with a lot. What we’re getting back is kind of model data anyway, so I think there is definitely a trade off there and I understand that more risk averse clients might be hesitant to give quite a lot of power over to kind of algorithms and. Tech giants, but in a lot of sense, we already were doing quite a lot of that, and we’re not necessarily losing out on vast amounts of data like we might once have done a lot of what we were kind of getting back in recent years was kind of more approximations and modelling. Perfect. Great. I definitely agree with all of all of those points. And finally, uh Maria, I think you’ve got a few points that you’d like to make in this topic as well. Yeah, I guess just tying back to your previous point on why should clients care and how they can implement this? I think it’s worth noting that AI implementation in terms of personalization, in marketing, it’s already happening. Thing the big brands have definitely embraced it fully and they’re actually building custom platforms powered by AI to provide better customer experiences. So a few example are Nestle who have built to this creative guidelines essentially powered by AI where they plug all their creative into create. Personalised messaging and they have seen much better our ROI efficiency from this API verified ads, if you will, than from standard ads that a creative agency may create, for instance. And then conversely, brands like Petco, who are just the Pets Wellness and. Food, I guess supplier they have a different approach to this. So they’ve created an eye powered customer journey whereby they each each member, each user on their side receives A personalised product offering and a personalised journey and as a result site site metrics are much. More favourable and bounce rates are much lower, so there is definitely a wide span of applications. They’re already taking place. I think it’s worth noting that it’s not just all about ChatGPT and writing copy for search ads, or, you know, in terms of language generative AI, it goes much broader than this, and its applications still don’t have like a limit. We’re still learning new ways that we can apply to our work. Yeah, it’s definitely something that you need to be at the forefront of, because otherwise you just risk being left behind and. Then I guess it to discuss a bit more on how can we at the kite factory UCI for personalization into our work. I think as media leaders, we should be paving the way for our clients. And you know, give a good example into this space. Try to bring innovation into their ways of working. So I think we can definitely make our work a bit more automated and AI powered as well. So I know Ben, you’re also working with the CHI studio currently to install AI into our creative processes. One of my favourite is writing up client emails, so especially in more delicate situations you can just give it bullet points and it will create the perfect e-mail for you and obviously save a lot of time in the process. Well, in terms of a delivery, I’m really touched on this. It’s not really something that we ourselves do because obviously we’re still at the mercy of the app. Platforms that we use. But pushing more ad spend towards that staying performance maps and advantage blood shopping campaigns and trying to follow the best practise that the Giants. I’m trying to tell. Us and then in terms of scaling and productivity as well also, I think it’s worth noting that not all applications of they are are in terms of like ad delivery, there’s a lot of tools these days that help our employee productivity in terms of time management and whatnot. So just by being more efficient. We’re able to add more value as humans to our work and therefore scale our profit as a result. And then finally trying to push more CRO technology onto clients wherever possible to optimise their UX in a more AI driven way. So I know we have a couple of partners that we work with already, but it would be great to. Get them. Get them over the board and more on more campaigns. Do you think we’ve it, you know, obviously initial results have been very strong for these big brands that have been testing AI and it’s because they’re, you know, they’re early adopters and they have got the scale. Is there any danger if everyone started adopting this, that there would be a huge? You know, homogenisation of everything, every all the creative and all this sort of approaches would be the same as people are becoming too reliant on on the AI which is giving similar answers to to the various challenges brands face. Is there a danger of that? Well, I think personalization is inherently unique. So I think what we’re doing now is delivering the same match to everyone, but I think we would actually. The to the exact opposite, if you were to apply AI for personalization, essentially every user would. Get a right. That is an app that is best for them and they don’t. They may. They may or may not know that ad was created by an algorithm, but what they will know is that that brand has thought about them that brands knows them. And we all know that in this day and age, particularly with Gen Z, they’re they’re not just purchasing a product, they’re purchasing the brand and they want the brands to completely get them completely get their values. Them and really feel like they’re a part of a community. So I think AI and just the on the personalization front that will definitely help to build better Connexions with customers. Right. And in terms of, you know, obviously AI’s impact is as everyone has said, is so broad and all-encompassing, how do you see it affecting your what you actually do day-to-day in the next couple of years, it may not have had the biggest impact quite yet, but we all know it’s coming so. How do you see your role sort of evolving over the next couple of years? Yeah, I I’ve. Always been an optimist when it comes to technology, I’ve always embraced it and tried to see the best in it. Obviously there is always the argument or befalls into the wrong hands. It has power to do evil, but I think even open AI have been very strong on the use of ChatGPT and they have been very like. Transparent with how it should be used in terms of how I see moral evolving I. I think it’s inevitable that our jobs will become increasingly more automated. I think we’ll definitely keep a role there as moderators when it comes to the output of the AI campaigns. Obviously it’s learning, so it’s not going to get everything right the first time. So us. We have as humans we have unique abilities that a machine would wouldn’t have. So far, to keep that machine in cheque and just make sure that the output is really the best thing for the client. So I don’t think I think our rules will change for sure. They will not disappear. But I think we will work with the machine side by side to deliver results rather than try to constrain the machine. And you know, try to put our human spark when everything that we do. Well, I think that that’s a brilliant place to finish up landing on a very, very strong point there. So look, please subscribe to the our Unmuddled podcasts, share it with all your friends and look forward to speaking to you, getting a future episode.