AI and Emotional Advertising with Sarah Levinger

Introduction: Understanding Competitor Insights

0:00how many of you know what type of ads your competitors are running I was going to say most most of you probably do how many of you know your competitor's pricing strategies some of you that's awesome how many of you know which emotions they're targeting on their pdps everyone should know which emotions your competitors are targeting on the PDP I think if you could take kind of the guesswork out of analyzing all of this you could compete on a level that no one's going to be able to catch up with if you knew your competitors inside and out and knew exactly what they were doing and which way they were going to go next you could absolutely dominate the entire Market first before we get started on like the

The Power of Emotional Connection: A Personal Story

0:34technical side of what I do and how I'm using AI to decode emotion I want to share with you guys a little story mostly because it really accurately portrays I think how important it is to emotionally resonate with your customers on a much much deeper level so this lovely lady is my husband's grandmother her name is Hana lore we call her Lori for short she is 95 years old and she has 26 grandchildren she speaks three languages she was in fashion for a very long time as a fashion designer one of the most iconic women I think I've ever met in my entire life so fantastic lady the interesting part about her story is she was actually born in the free city of DC in 1929 so this technically free city of DC is now modern day Poland for her and her family this was a really interesting time period obviously cuz in 1929 they had to free flee the free city of D basically because of the the start of

1:30World War II so early in that year she and her family her mom and her sister Ruth actually fled from their Hometown just kind of barely missing kind of the German German invasion of Poland that would happen later that year so when you came into the US from anywhere that you were trying to like immigrate into the US the only thing you could carry was things that were like on your back so they would carry all of their luggage along with like all of their fur coats and everything and it was just kind of a trek to make it into the United States at this time came into Ellis Island kind of right at the peak of summer so at the time the only way to actually get into the United States was to have a sponsor so for Lor's family they were actually sponsored by her uncle so Uncle linger actually brought them in from Ellis Island and took them straight into New York City while they were in New York City

2:18obviously it's very hot they're very tired they're very emotionally spent from their Journey so he took them to a soda shop where Lori would kind of experience a very interesting I guess connection with a product and a brand that I think is just important to tell now this became a very big part of her story as a refugee and I want to share it with you mostly because this shows just how incredibly integral it is and impactful it is to have an emotional connection with your customers because this is where Lori had her very first coat right after she came off the boat on Ellis Island right after traveling for many many weeks just to get to the US really this kind of small experience made kind of an incredible difference for a 12-year-old who was going through this experience of having to immigrate to an entirely new country right so it was a very simple moment but it was more than just a drink it was kind of her first emotional connection to her new

3:07home so very very impactful for her so

Using AI to Decode Emotions in Marketing

3:10today now that we have my correct slide deck I'm going to show you guys how we're using AI to create these kind of impactful connections with consumers and with our customers not just for business but just to bridge that gap between the digital age and the emotional side of the humans that we sell to so at the end of the day obviously like we are humans that sell to humans so we have to be able to correctly identify which emotions people are feeling because understanding that Human Experience is how we're going to change the world in today's e-commerce landscape I feel like AI is is becoming this kind of big Hot Topic and it's a lot about algorithms and data points and all kinds of other things but it's more than just that right it's becoming kind of our emotional window and our compass into the lives of the people that actually purchase from us in general I I want to start by inviting you to think about like imagine if Coca-Cola had known exactly what my

4:00husband's grandmother was feeling in that particular moment how much more prepared they could have been to actually provide this kind of experience to other people who were also experiencing these kind of like terrifying exciting nervous feelings that she was feeling and they really could have helped people kind of find peace in a moment where it was really intense for them I think it's important to also know like this AI is really empathy kind of on steroids because of how much information and how much data it can actually process much much faster than I think human I don't know time can do obviously our brains are are incredibly powerful computers but this is this is just where it comes down to we don't have time as marketers and brand builders to process the amount of information emotionally that's coming out this is kind of bridging that gap between humans I also want to point out that this this is not like about predicting human behavior about what they're going to purchase this is more about understanding why

4:55they feel the way they feel so we can connect better to it and actually serve them better on a deeper level there's a

Case Studies: Brands Leveraging AI for Emotional Resonance

5:01lot of brands that are doing this right now Este Lauder is actually doing a ton of AI basically to recommend product based upon not just skin problems for their particular customers but also how people feel about their skin problem which I think is massively beneficial for this company because they're actually going after emotional resonance not just like problem solution kind of a thing Amazon obviously very very big on using AI lots of data going through obviously their giant ecosystem they're able to actually increase based upon the fact that they're using recommendations based upon behavior that's coming through the site and they're actually able to decrease customer response Times by 40% just by recommending the correct product based upon the behaviors that these customers are doing and the AI is the one that's analyzing that stit fix actually another one that I really really love these guys are able to actually use AI to I guess analyze kind of like style preferences of their customers they're doing a lot of like body type analysis

5:55and customer feedback analysis and they're doing this too just so that they can emotionally resonate with people on a deeper level and increase the chance that people are going to stay so loyalty LTV huge huge piece of this puzzle glossier glossier however you say it they're using AI right now to analyze customer sentiment as well they're actually doing this on all kinds of different platforms they're using AI to do social media sentiment customer reviews and customer service like tracking what people say during their customer service experience so these are not just efficiency markers they are really really good moments of actually connectivity right so just like Coke did for my husband's grandmother this is how AI is kind of helping us to make these deeper connections and resonate

Practical Guide: Implementing AI for Emotional Analysis

6:38emotionally some of the most powerful tools that I currently use to do this are large language models that are built to analize customer language in particular the words you say as a human have powerful powerful impact so when we're looking to capture kind of large amounts of data on this we need something that's kind of capable of understanding how humans talk and what it means when they say certain things I'm doing this mostly because I'm trying to figure out what our customers feel a lot of times I see Brands come in with this idea of we need to figure out whether it's positive or negative happy or sad right those are the two that I see the most prioritized but there's emotions that go way way deeper than just happy sad frustrated those type of things and so I want to show you kind of how to pull this emotional data and how I'm using AI to basically generate emotional models so that we can resonate deeper show hands how many of you flew

7:25here on a plane okay and how many of you had a great travel experience okay a few how many of you had a me experience that was me how many of you want to sue the airline after this is done if I was the airline and I wanted to analyze kind of your emotional sentiment around my particular like airport and I wanted to make sure that you came back and flew with me again I would probably collect the data and then I would analyze it for positive or negative sentiment and basically try and make a a guess as to whether or not you were going to fly with me again now this would be really great information to have but the problem is you can't really dig deeper than that right so what if I wanted to understand more than this what if I actually wanted to understand whether my customers were more likely to turn on a Tuesday what if I wanted to actually analyze the connection

8:13between shopping behavior aov and customer identity how would I go about doing that is there a world where I can actually take that data then and explore the idea of advertising specific emotions to a specific customer at the exact time that they're most likely to purchase based on their experience as a human how would I go about doing that this is what AI is currently able to do which is kind of terrifying but also so so fascinating it's actually analyzing the emotional Behavior life of customers and we can figure out pretty quickly what they're going to do where they're going to go so I'm going to real quick kind of show you how I'm actually using this to decode emotion using AI first one here I'm a huge huge huge fan of NLP it's natural language processing takes down down a crap ton of data from the brands that I consult for categorizes them into nine different emotional categories so that I can tell from the core of it what people are actually feeling during the purchasing process

9:12not just after it so this is based on frequency numbers I can tell kind of just from what we pull out of the actual data not just which emotions they're feeling but also how confident the AI is that these emotions are going to work s Sarah where's this export from usually where did you kind of go like where am I doing this export from this one yeah yeah these actually come out of the llm that I'm that I built to to analyze these numbers so you can actually do this in chat if you would like which I have a tutorial that I'm going to show you how to do but yes so these scores are just based on the llm that I built mostly because I wanted the system to tell me how confident was that it was going to work and then also how motivating these emotional messages were going to be great case steedy shout out to Nate we actually did this for original grain it's a good looking watch

9:58though I know oh they have such nice watches go get a watch from dat cuz it's amazing and when we did this work I actually generated ads for them based upon the emotional Foundation of what we found so this was an original message that I created for them Nate made fun of me because he said it was girly but this particular message ran for 6 weeks during Father's Day and aov increased by 35% LTV increased by 33% and revenue increased by 25% during that particular time that it ran so NLP based emotional ID that's one way that I'm using AI can't say enough for that it's just so much fun so take a picture if you would like this is how you actually do it through chat if you wanted to kind of see if this would work for your brand first gather that review data very important organize it by character count I always do this first up because organizing by character count gives you context llms need context they need a

10:52lot of characters a lot of words to be able to analyze it so pull anything that just says like this was a great product that's not usable to us big long pieces of text is what we need what's the range you like what's like 20 words 30 words I like to say anything over 300 characters so I think a tweet tweet I need a tweet or longer yeah yeah is what I need so drop that into chat categorize it into positive negative emotions and then ask chat to give you the top three core emotions that are coming out of that data now you can do this for anything it doesn't have to be just reviews I've taken full customer like interviews transcribe them and then put them into chat and does the exact same thing pretty powerful once you start to pull the emotions out of it so now let's

Advanced Techniques: Competitor Analysis with AI

11:35talk about competitors because I think this is pretty important in the space how many of you know what type of ads your competitors are running I was going to say most most of you probably do how many of you know your competitors's pricing strategies some of you that's awesome how many of you know which emotions they're targeting on their pdps you oh my god wow that's amazing fantastic everyone should know which emotions your customer which emotions your competitors are targeting on the PDP I think if you could take kind of of the guesswork out of analyzing all of this you could compete on a level that no one's going to be able to catch up with if you knew your competitors inside and out and knew exactly what they were doing and which way they were going to go next you could absolutely dominate the entire Market so I created kind of a custom GPT that studies the landscape the actual competitor base the industry itself and all of the emotions therein mostly

12:21because I wanted to identify the same exact stuff that liquid death has identified they came into the market a brand new player in the space with leg Legacy Brands Aquafina danani they came in here and decided they just wanted to be different and they actually went after a psychological angle and that's how they built this $ 1.4 billion do industry that they've just dominated over here mostly because they wanted to identify that Gap in the market from their competitors and they're I love that brand I Can't Get Enough it so I recently ran like a full psychological panel for a brand and within minutes we can identify kind of all of the things you would need to know to actually overtake the market just B based upon like psychological data so pricing structure core messages bestselling products all those things AI can do right now based on just scraping data off of freely available information using these kind of elements my time to Market decreases right so I know exactly what my competitors are

13:17doing and how I can differentiate myself so I can get ads out the door within minutes I can get Landers up that I know are going to hit I can pull emails I can do SMS I can do like full-on industry events based upon what we know about the market and it can drastically change the direction of an of an entire brand because what you know once you know what consumers are saying and what they're actually seeing in the market what they're interacting with then you can easily draft anything you need marketing wise that will fit that so this these are some of the ads that I generate for some of my here's how to ID gaps you take a picture of this if you want open chat explore gpts this is where you're going to find all of your custom GPT like models basically and connect with Access Link so Access Link is a custom GPT that I use that basically scrapes any link I put in there so I'll

14:04take all of my competitors links websites landing pages anything that I want to analyze drop them in there and then ask it to analyze it for psychological positioning Market position USPS value props like give me the just the down and dirty details of what my competitors are doing and that way I'll have this information on hand I I don't have to like worry about it the other fun thing about this is you can actually have chat send you a report every single week of like here's what the competitors are doing every single week so super fun stuff I Can't Get Enough it okay so last little bit here

The Importance of Context in Emotional Marketing

14:36we're using kind of AI to uncover like emotional sentiment and competitor stuff but that's just one piece of it there's a deeper use case for this that I think is massively beneficial one of the most difficult pieces of data to get on a customer right now is not necessarily emotional data it's contextual data context is very important for humans most of the because the situations we find ourselves in have a direct effect on our Behavior right what you interact with as a human on a daily basis actually changes how you consume things what you consume who you consume it with how much you consume and where you get it from this is very very important I think obviously text is messed up but Behavior I think is what we're all going after as marketers right and a little bit of emotion but there's more things beneath it that the actual behavior is built on so behavior is built on emotion emotion is buil built on your perspective of the world

15:30which comes down to how were you parented what did you experience as a child what were the first seven years of your life like which generation were you a part of which world events did you experience and how did you feel about that time period as a human perspective is based on that all of those experiences that your brain logs away all day every day and then experience is based on context meaning what what were you experiencing and how was the world framed to you so I think to truly understand and kind of reach the humans that we want to reach on that deeper level where they will never ever forget something that they experienced with us it's pretty important to to actually analyze the context right that they actually experiened that in CU we understand that she was feeling nervous scared excited whatever it was but it was because the poor woman had to flee her home country that's a lot of Terror and a lot of fear and a lot that resides underneath just one experience that we

16:27need to know so really good illustration of this this is a book that I currently read my children it's called that's good that's bad my mom used to read me this book I don't know if you guys have heard this one such a fantastic book really kind of illustrates how context works so it's a story about this little boy that basically had a really weird day at the zoo it details kind of some strange things that happened to him after kind of this balloon whisks him away and it kind of goes through the storyline of like it gets weirder and weirder and he's interacting with all these animals and he's doing things with them he's trying to find his way back to his parents and each moment that you actually experience in the book it provides more and more context for what's happening along his journey we are journeying with him through the book seeing all of these weird things that are happening to him so we're kind of privy to the journey right the entire

17:13experience and some things turn out to be good some things turn out to be bad but towards the end of the book we kind of find out right like how this kid is feeling about his experience and one thing I want to point out is throughout this we already know what happened we were there with him while he went through this right and that's massively beneficial for marketers in particular if you can put your brain inside the brain of your customer and watch what happens to them as they're going through their problem you can do some things that are just amazingly like powerful for their particular life so when he finally gets back to his parents we feel just as relieved as he does because we actually felt it with him and this is what I really want to hone in on because we actually have direct access to where this kid ended up right so so in the end he finds it back with his parents he's obviously sense of Happiness relief understanding this is

18:04kind of empathy at its finest and this is what I think AI is is best used for is understanding the entire Journey that our customers are going through not just the problems that they're facing not just like the jobs to be done I think very very important so context is everything obviously when it comes to emotional marketing we need to find a better way to figure out how to uncover emotional context so this is how I am currently doing that I've been working on something that I think is going to be pretty fun just to work with with the brands it's a custom trained chatbot that's trained on customer data to act like a customer and I I can't get enough of this stuff just because based on how much data you run through the actual chat you can basically generate a a customer in real time it's a customer in your pocket and it's fully trained on customer data so what it comes back and what it gives to you is not going to

18:57just be like I guess this is kind of what the customer feels it's actually going to be focused towards this is what your customers probably do feel we have high confidence level that they do so it's basically a representative customer that I can talk to 24/7 now the nice part about this we're training it on like post-p purchase survey data NPS analysis industry analysis lots of other different things I can actually ask this chat whether or not the marketing materials I'm coming up with are going to work before I run them which is I can't oh I can't get enough of it I just I all I want to do is talk to these show all day long the other thing that you can do though is ask the chat how this person is feeling right where did you come from what did you experience in life what hurts most and how can I help this is kind of the power of of AI and I just I the more I work with it the more

19:44I'm just so kind of amazed as to how deep we can actually go emotionally with something that was not built to be emotional it's it's just insane how much is changing so at its po obviously AI is kind of just making it easier for us to emotionally resonate with people who live on this Earth but it's understanding Human Experience the whole journey not just part of it so the brands that succeed today I truly believe are are kind of listening deeply right they're using AI to really deepen their experience with the customers but they don't just see customers as somebody who purchases from them they're people right with emotions and stories and lives that are completely disconnected from the brands that we're currently trying to build which is the reason why I think it's so important to go deeper here this is kind of the power of these tools and that's how we Foster genuine connection and that's how we change the lives of people like Lori so thank you oh my goodness did it we did

20:38it we got it done now we'll have the slid so there's a lot of screenshots can you go back to that slide for me because there's still people taking pictures thank you yes if you guys want to this is basically free mini training that shows you how I'm generating ads in particular off of the emotional analysis that I'm currently doing so if

Q&A Session: Practical Applications and Insights

20:57you guys would like to take that does anybody have any questions for I would love to yeah I'm going to give it to you state your name so you mentioned that in chat gbt we go to access link right and then I'm I'm trying it right now but we can't actually give it a link it seems like we have do we have to attach actual reviews like export them from somewhere and then give it to it or what is the best way to go the weird part is is it can actually analyze anything that's open so as long as you have a spreadsheet that has all of your reviews in it for Access Link if the Google sheet has open permissions like anybody can access it it'll go ahead and analyze it for you okay so you pull them and then you give yep yep yep so I just put them in a spreadsheet it's the easiest way for it to analyze that's a good one for Access Link you have to

21:40have an open link if it's a privatized one it won't do it obviously cuz the front hey Sarah hey Nate do you remember how much I hated that ad you did I hated that ad it went against they hated this ad like the look of it was wrong the headline was simple and and we paid you a DEC amount of money to write it but then it made us a ton of money yeah but now I hate that ad even more cuz I can't write anything better and we've been running this headline for 2 and a half years now and it's crushing so what do we do next how do we get you back like what's a great kind of like tactical way for a brand that might have some things working for them but they want to kind of yeah you know keep pushing I think you could hire me if you really want to but the other thing that you can do is dedicate somebody dedicate somebody to that job their only job is to analyze

22:32what the customers are doing we don't currently have a role for this in D Toc and I'm kind of shocked that we don't have somebody whose job it is just to analyze consumers and humans so that's the first thing I would do hire somebody or if if you can hire somebody then obviously go get like a consultant or you know there's lots of people that are currently in the AI space that can solve it or you know you can hire me if you really want to hi quick question yes what sources are you getting the most Val bring the bring the mic closer brother what sources are you using to get the most valuable data from customers is it reviews postp your surveys interviews like where are you getting the juice from anywhere you can get large context so I would prefer that we all interview our customers in like real time meaning like get on a zoom call mostly because the way the person's face emotes will tell you a lot about

23:25how they actually feel about what they're saying which changes the concept so I usually use reviews cuz it's the easiest way to grab it but I would prefer fully transcribed customer interviews with at least 7 to 10 customers more if you can get it you mentioned that you even take chats that are happening with the customer service and put them yeah yeah customer service is like massively beneficial cuz they they'll give you a whole bunch of stories and like different things that they want to bring up that they may not say in a review reviews are almost always is like positive unless it's a onear review obviously but yeah yeah that would be a great place to get it there's another really really good person his name is Rory and he has this one way so how Sarah makes those ads where it's like analyzing consent analyzing like some of the best performing images on things there was like another guy he has like a small presentation I think it's on YouTube where you can kind of upload all your

24:21reviews and it just does it all for you in figma and it was like a wild presentation I'll see if I can try to find it on and post it on Twitter but we question yeah yeah I'm coming I thought it was very interesting in one of the slides that you've shown I I took the liberty to build a similar model that you had shown me a year or two ago and what I found very interesting in one of the examples that you gave was that you were able to tie average order value sorry order value of the purchase and also a time stamp of the date of purchase and tie it into the review so just what I'm curious about is that is there any software or is it a very manual task of tying the the review to the purchase and bringing that all into an Excel sheet or into chat gbt or whatever it is because obviously that's another layer of like you said segmentation or psychology that my go I

25:11would love to participate in oh that's a good question I have currently tested this with Access Link a few times where I'm trying to tie customer identity to aov and seasonal purchasing behavior and I've done it with a big brand that gave me like crap tons of data to do this off of they had dates obviously for every purchase that they had aov and they had LTV numbers as well and we were able to very quickly through chat just identify here's all of the data coming out from this specific identity so one identity was baseball lovers and they were able to tie aov baseball lovers in the time period of the year that they were most likely to purchase and the emotion that would probably work that was all through CHF so I don't I mean as far as I know there's nobody like has a definitive SAS product that does this okay so there's nothing out there that T not yet not yet I know it's probably coming but not yet the amount of things that you can

26:04analyze is nuts it's crazy but yeah Sarah thank you very much thank you [Music]