Aaron Conant  0:18  

Happy Thursday everybody. My name is Aaron Conant, I'm the co founder Managing Director here at BWG Connect. We're a giant networking knowledge sharing group of 1000s of brands. We kicked this off about six years ago, as a small group, where we were working together to tackle the biggest issues that's presented across the digital landscape, but also help each other with service providers selection as a whole, who's doing the best things out there, and who's working out the best for everybody. So we've grown over 9000 brands now, thanks to everybody who's participated over the past six years, we'll do close to 100 dinners and a couple 100 virtual events. So if you have any, you know, topics or ideas or you ever want to chat, digital strategy, just reach out to me Aaron@bwgconnect.com. As we kind of get started here, we're kicking it off three to four minutes after the hour, we'll wrap this up with a few minutes to go as well, just to make sure everybody has time to get on with their next meeting without being late. And the other thing is, we want this to be as educational and informational as possible. So along the way, if you have any questions, drop them into the chat, drop them into the q&a, or you can always email me Aaron@bwgconnect.com. And for the email, it's literally during this call tomorrow next week, whenever you want to chat. You have any questions? Shoot, shoot me over an email, we usually can kick that out to the group and get your response pretty quick. So now is we jump into the topic for today rethinking search marketing? Well, we've got some great friends, partners, supporters, the network over at NetElixir Udayan And it's, you know, just absolutely fantastic. You know, the insights that he's able to share with us Udayan, and I'll kind of kick it over to you. If you want to do a brief intro on yourself and NetElixir. That'd be awesome. And then we can kind of jump into the content sound good?

Udayan Bose  2:00  

Absolutely. Aaron, thank you so much. I don't know if you wouldn't mind giving me the screen sharing access. Oh, for sure.

Aaron Conant  2:08  

Yeah.

Udayan Bose  2:08  

Amazing. Thank you. Yeah. should be all set. Perfect. Amazing. Thank you. Thanks, again, for having me on this session. I mean, this is always a privilege from the entire NetElixir app team is essentially representing all the good stuff that the team has been doing. And today's presentation, I think, is the third add on if I'm not wrong in this series that we have been running over the last about, what four months, five months? Yeah, since our world was taken over by GPT. So it's I mean, I really, it has, it has been exciting. It has been crazy. And it has been just just a different world altogether that we are living in. I still remember the first time we did this presentation, which was I think in the month of Feb or March, probably March actually, when we have been charged GPT and Microsoft announced their their sort of a partnership and so on. And so much has changed since then. So I fairly quickly do an introduction about the company, many of you have been regular. So great. Great to see you all again. Thanks for joining in, I'll do a little bit in terms of generative AI and search, try to just break it down in the key bit as well. Winning in the new search landscape, and just end with the concluding remarks as well. So very quickly about an introduction. I mean, we have been around for Netflix has been around. And it has been just just an amazing time, almost 20 years where we have literally sort of built ourselves into from a startup to the one of the top 27 agencies for Google in North America. And it has been just, Justin. Just a very humbling journey. Overall, we about five years back decided to invest in our own building our own technology platform, which use vertical AI vertically is basically we wanted to keep all the data country confidential. And essentially try to ensure that this technology works very well for our customers and only our customers and we're not really sharing the data with any external AI platforms like NSA now discharge GPT, and so on and so forth. So that's been driving a significant chunk of our growth as well, which has really as I mentioned, landowners in the top 27 agencies. For Google, we are also a certified minority business enterprise. We also have a program or initiative it basically a a wholly owned subsidiary called the Quran Foundation, where we try to support some girls from the underprivileged background in India, pursue their dreams of a better life through education and eventually employment. We currently are supporting about 24 Girls, most of them come from the slum areas in the city of Hyderabad and many of them are single parents. Families and there is no way that their parents can afford their education beyond the 10th grade until 10th grade the basically the extra subsidized education there were to study. And as a result of which we found out something which was really concerning and deeply concerning. And that's that stat really just blew me off that every seven seconds, seven seconds, I repeat that one girl under the age of 15, gets married away. In the developing world, seven seconds, a girl under the age of 15, gets married. So we want to do something about it, right. And this is our very, very small way of really trying to make a difference for some of these girls. So we are perpetually sort of really going ahead and moving forward, because a lot of the support from many of our clients, many of our partners, all of our team members, and will be shipped. So if you are interested in contributing, you can scan, scan this QR code, directly take you to our website, where you can see the work that we are doing and feel free if you're if this is something that you are passionate about helping these girls at least pursue their dreams, right, they need to get a chance, I really appreciate that. I'll keep it about me, I found it Netflix, almost 20 years back along with my wife. So it's a husband and wife combination, running this company for almost 20 years, you can connect with me on LinkedIn, that's my email address, feel free to send me an email. I'm also very privileged to be an Innovation Fellow at the Columbia Business School. And I also do a fair amount of lecturing at the Johnson School of Management at Cornell, the City University of New York and Canadian school of business as well. So a lot of the stuff that you're going to hear is going to be very professorial in nature, there is a lot of research that I'm going to share with you. Again, if you have any questions, please feel free to add to the q&a, or the chat box, we also have two polls where we are going to be giving out ear tags for whoever sort of answers those poll questions, the first or the cut gives the correct answers the first, please feel free to add those responses in the chat box, specifically, not the q&a box, but chat box because it becomes really difficult to toggle between both of these lessons here. About generative AI and search. And again, I don't tell you that I want you, I want you that I'm as a professor, I'm just using going to use that hat, and really just start with a poll and then a model. So my first question to you, I mean, I we have all been talking about AI and Adam requested that can we can we really then make it something that people are able to walk away with some tangible to dues, right? So I just wanted to start with this one, right. So as prediction becomes cheaper, prediction becomes cheaper as basically AI becomes prevalent, right? Anything which is launched initially becomes expensive. And then slowly the price comes down, whether it is a Tesla a few years back when it is launched to Tesla ADHF variations now. So that's normally I think the case like any technology over time sort of becomes cheaper. The value of what out of these four, in your opinion increases. I know it's a bit of a twisted question, and I'll leave it to a model as well. Do you think that as predictions become cheaper, Is it human judgment, which becomes valued even more emotional intelligence instinct, domain knowledge? Any guesses? Any guesses at all? So I see the B, A, B, B as such. So let me let me get to this, I think we already have the winner. The answer is a so whoever answered a first RM Angela, let you keep a track? Yeah. Awesome. Amazing, perfect. So this is a model which I really have, I think introduced earlier in some of our previous conversations as well. This I think, is fundamentally for anyone, anyone who wants to do get into the basic economics of AI. So what this is from a book called prediction machines are simple economix of artificial intelligence. And what they do is at the end of it, AI is accomplishing some tasks for you or helping you in doing some task as such. So when you talk about tasks, to understand as to where exactly AI is helping, you need to understand and break down the task into different parts. Elon Musk has become very famous for his first principle thinking so think of it almost like a first principle thinking breaking down a task, a task can be, I don't know, maybe picking up top to writing something to speaking. Everything can be broken down into different bits, right, you have to start with an input, there is a certain reason there is a certain data file which you are essentially getting access to. Based on the input, there are three parts. The first one is prediction. So if you do this, this will happen right? So you're able to predict and that's where the prediction part comes in.

In many cases, the prediction becomes simpler. As you get trained, so as you have done this, I know if I want to pick up the cup, this is how it will be and so on and so forth, it easily comes easy, it becomes almost commonplace. And then there is a third part, which is the judgment. Judgment is think of it in a way, there is an alternate way or an alternate solution to this particular path in many cases. So, what the author's the professor authors essentially argue, they say that as predictions become cheap, right prediction machines or AI becomes cheap and the prediction the value of predictions essentially goes down. Because the AI is able to point out and really map out all those known known cases, it is able to even point out even in the case of generative AI, as we have seen known unknown cases as well, right. So, that's where it just becomes easy. And then once we're able to predict it takes an action or assists to take an action, there is an outcome and then the AI that it probably will become smarter over time. And the professor's initially countered that there is no way that in the long run, human beings can really compete in that cycle, right, because just given the number of variants and the number of data points that the AI can access, specifically in the open web, I think it's really we're going to compete on the prediction action outcome and training mode. But what they also feel is, by doing letting AI really take care of this and making peace with the part that the prediction part, let the AI take care of it, you can really focus on the judgment component. And that's an important component judgment is the new path. Judgment is a new path, which you probably did not have enough time to focus on. Since you're spending most of the time in terms of this prediction, action outcome and training loop. Right now that is taken care of by AI, it frees up time for you to focus on the judgment. So the professors actually countered that AI, as predictions really become cheaper on AI, the value of human judgment actually increases. And that's a very important component for you to think of. And no, I know there are a lot of doomsday theories aren't in terms of how AI will sort of take all the jobs and cetera, et cetera, et cetera. It not take all the jobs let me as a state explaining as to what my thought process is, it will probably change as to what are the things we can important, what are the new tasks will emerge, and new jobs will emerge, it's going to do a shift, I don't think it is going to take anything, if at all, I firmly believe that AI will make us even more productive, it will really, really help our human ingenuity part really flourish even more. And I firmly believe that it would really lead to an economy explosion, I think McKenzie came up with this lovely paper last week, which says that generative AI actually substantially will lead to an economic boom, by really getting onto those judgmental areas, those paths, which are not being really taken at this point in time by anyone pretty much. So that's where I wanted to start with this just to quell any any conversations of any doomsday theories. And this is just purely

Aaron Conant  13:04  

in the matrix, right? We've all seen The Matrix, right? Like there's going to be directed energy. Nice. So it's a

Udayan Bose  13:11  

it's a different world. And we really have to just have that mindset to accept this difference. So but we have seen that artificial intelligence, as we used to call it has really evolved as well. Right. So I think the AI keeps on evolving also. And what really changed with generative AI was substantially bigger than what we had experienced anytime earlier. I mean, in fact, Microsoft CEO Satya Nadella said that the AI that we are used to let us say an Alex or CD seems to be done by obviously there is a Microsoft angle to that. But I think it's more of a generative AI and to this entire thing as well. So effectively, I think open AI, at least in my humble opinion, has been has been the biggest change force that our world may have experienced in a very long time. Earlier in my previous conversation without an ad shave that probably can be compared with Netscape browser, which sort of essentially democratized this entire Internet. It will be bigger, just because if you look, you'll notice I mean, within the first two months, 100 billion people had used activity, right. So we are talking about abnormal and explosive growth. So again, the question out here is, I mean, effectively, this is what it is, it has taken to a different level. How can we really capitalize on this? And I think companies are showing the way I mean, this, this was the the most pivotal moment on the seventh of Feb. This year, when Microsoft essentially announced its investment that they have been invest in GPT and open AI. I mean, it was very clear our words have changed forever from that day. I mean, we know as to I think Google tried to counter that with BB and unfortunately didn't really go very well. Bardot has come a very long way now, and so on and so forth. But I think that that was the day when That was essentially changed. The reason why, for our space and getting the conversation back to search now, it really led to the emergence of a conversational search funnel. Right? So this is essentially the new Google what they call the search generic search experience. Right? This is the new Google page. And a lot of this, they were almost like forced to introduce this third dimension. The reason I'm using third dimension, you normally in the Google page, you are used to scrolling down and really checking out all the links. But in many cases, in this case, there is a chat bot, which is assisting you to answer your questions. Now, from that perspective, it really became a hard dimension. And there were some very specific implications, there were some very specific implications. And I think we will realize more and more the fact of some of these implications as well. But let me share as to some of these things with you as well. So you can get a bit of a context as to what you can expect to change in the world of you does a Google search on Bing search for Microsoft search and so on. Let me start with this fact, this was a BrightEdge research from 2019. And which showed that 68% Of all the website traffic or at least trackable website traffic is sourced from the organic and the paid search, the search engine still drives a significant chunk of the website traffic, right. The number may have come down a little bit, but I think still it is more than 65%. So two thirds of the traffic has been driven by search engines. So the website, right? That is how dependent your business is, as an eCommerce company, direct to consumer to get whatever business you are in, in terms of Google, primarily Google for traffic search traffic. Notice, consider this. I'll present to you two steps, right, the first one, pre generated VI, so let us say last year, at this time, there was no chat, GPT, nothing, et cetera. And it was, I think, a known fact that only about 18% of all the traffic, which was searching for anything was really going to the second page of Google, right, only 18%. So that means that 82% of the people did not really go beyond the first page of Google. Right? Who goes to second page. And effectively, I think there was the same we sort of came up the best place to hide a dead body is a page two of Google search results later, but that same has been prevalent now. Now consider this, you have this chat bot assisting to answer many of your questions. I don't think it really takes a rocket scientist to understand the application, this 18%, unfortunately, will drop by I mean, I would, I would probably place a bet that maybe in the next year or so this percentage of people going to the second page of Google may come down to even 5% or lower right 5% from 18%. And that's because

Aaron Conant  17:48  

they're going to be interacting in real time, there's no reason to go to the next

Udayan Bose  17:54  

page, there is no reason to go to the next page, right? So you go to the next page, mostly in terms of research cues, you don't really go it as much to second ways to buy stuff, because all the ads are essentially thrown at you on page one, right? So you do more of research. And that research provides Google an opportunity to monetize the page two, page three, and so on and so forth. Right. And I think if you really, I sort of read this somewhere, I'm unable to quote exactly the source. But roughly about 25% of Google's revenues came from page two onwards, right. So page, one Street drove about 75%. But page two, page three, page four till page N, effectively drove about 25% of the region. Now just a bit of a thought experiment, right? So what in your opinion would happen if suddenly the Chatbot comes in, and only 5%? I don't even know as to who these 5% would be 5% are going to the second page. So understandably, the monetizable space declines quite dramatically for Google. What does it mean for you as an advertiser? Now, when we did a quick analysis, and this is probably actually from 2021, for a word, like buy flowers online, or buy gifts online, there were about 4600 businesses competing on a keyword like buy flowers online 4600 advertisers, and Google's big thing is, by establishing a bit of a long tail, I will say longtail page 2345, they were able to monetize all these pages and make money. Now, what is the point? I mean, why would you even go on to the second page? Or why would you want to be on the second page of Google at all right, because it wouldn't really get you traffic. So just think of the impact it would have in terms of inflating the cost per clicks or CPCs. Unless there is obviously another school of thought which many of the search experts are saying that Google will find out a way of monetizing the the entire conversation search process, but I doubt that because that itself would defeat the purpose. It's almost like me and Aaron not having a condom. Session, suddenly an ad keeps popping up in between. Right, that will leave, that would lead to people leaving Google very clear. So I think it's just too big of a risk at this point in time, in my opinion for them to really start monetizing, monetizing effectively the the solely, I mean, as it is with the paid ads, etc, etc, the search fine. So it's an interesting situation. So again, what if about Robin monetizable real estate leads to significant CPC escalation. Thereby, I think, think of every search page as your, whatever the the the first point of contact for your customer. I mean, Google has this concept of zero moment of truth, right, where even before the customer comes to your website, they actually interact with your ad on Google, right. So that zero moment of truth. That's the reason Google search for search engine results page was so important. Now understandable, the CPC goes up. I mean, it'll, you'll really be bidding a lot higher on the Google keywords, which again, have gotten AI. So you don't really have too much of an option there control there is that, and I'll tell you as to how we are trying to approach that. But the value of that search pacer page, may essentially our search engine results page may actually get impacted because of this change. The second part is maybe this drop in terms of the monetizable real estate, so page 234, etc, suddenly, there is no point in advertising there. Make search marketing, simply unviable for small and midsize businesses. Maybe they see your listings get pushed to becoming even more inconspicuous. Google has been anywhere for the last seven years have been done to constantly push down SEO listings. And definitely the rules for SEO will have to be rewritten. So there is a lot which is going on right at this point in time. But all of these are good things to think about. Specifically, because 65% of your website traffic is driven by Google right to third. So your business is literally dependent on Google. And that the situation and then Bill Gates made the statement last month, on exactly a month back in London, that he says that eventually a top AI agent will replace certain shopping sites. And he says the currency is unable to do that because it is working on data, which is still September 2021. Or at least GPT is but eventually once we are able to match this real time and this batch data, there is no reason why you really need to go to a second date, right? Because maybe a mania, as simple app, maybe an agent on your phone that you just tap ask a question, whatever it is get them to buy whatever it takes, right? It only has your system. So it's a it's a fundamental, fundamental shift in terms of how you people search, and also for you as advertisers. How exactly and what will you really advertise on as well. These are important questions to be in and specifically mentioned, the reason two thirds of your website traffic is coming from search engines, right? So you might as well do not mean you're

Aaron Conant  22:54  

not, you're not even going to be typing, right. It's going to be an audio interaction back and forth with a chat bot that's tied into real time web.

Udayan Bose  23:04  

Right? Yeah, yeah, absolutely. So I think it's just a very different world that we are going in. I mean, if you're interested those who are interested in reading this further, I have four articles. And we'll be sharing and if you want, just send me an email, send me an email, my email is my first name will then ethnic mix it up coffee, please send me an email, send me an email, we'd have to share these links. But it can be really helpful for you to really read these articles. I mean, I think one of the best articles explaining generative AI is written by Sequoia Capital way back in Jan. Sam's Altman, again, the founder of open AI, the man behind this chat GPT disruption. Effectively, this is more of a I think the the more human part about them what they are trying to do. Age of AI by Bill Gates is, I think, a phenomenal one. And we have written our thing is how genetic search as well. So feel free to go to our chat, the blog site and just take a look at aspect. So this is an interesting situation. Now effectively, how do you win in this new search landscape? So when we really,

Aaron Conant  24:05  

really quick, where are we at in this? Because I think everybody on the digital advertising side and eCommerce wishes this was pushed out a year, right that it didn't happen so soon, because they have a million other things to they're trying to get on top of right now. Right?

Udayan Bose  24:24  

This list comes in, like four months, five months old. GPT was launched in November 20, November last year, end of November, beginning of December. So it basically has been six months, effectively, right six or seven months. So all this is moving at a work pace. And interestingly, I was listening to the insider intelligence podcast effectively on the way essentially kind of sharing from con the Advertising Festival. And they were saying that there is only two words we are getting up here in town right advertisers ai ai is Artificial Intelligence. That's the only thing there is nothing is getting discussed. There is no discussion about metaverse. There is no discussion. Interestingly about Twitter, despite all the controversies, there is no discussion about anything else. It's only AI. That's it. Right. So that is what I think the power of this rapid explosive, I wouldn't even hockey stick this is almost like a vertical growth. And that we are seeing that is, and it is disrupting every industry. I mean, I can literally guarantee you that your industry is getting disrupted now. So that's the part right. So effectively, we face the same challenge elsewhere. Because as an agency, theoretically, there are two forces that we are dealing with. So already last year, I think this was in the beginning of last year, when Google started introducing its AI modules, right, like the performance Max, I'm pretty sure many of you who advertise on Google are very familiar with performance Max, not everyone may be very happy with performance math, but it is the AI module. So we will set that will will create will make this into our area where you tell us the results that you are expecting, or we realize the goals that you have, you'll give us the audiences that we want us to target, you give us the initial creatives, and we'll do the rest for you, you don't have to bother about human intervention, actually, you shouldn't, because that will disrupt your performance, maximize some of your films, and thereby effectively utilize AI and what has really emerged from this. Now it has been about a year and a half, that we have been working on the performance. But we are one of the early ones, as the as a real agency leadership member to get access, it is pretty much a black box, right? So you can't control anything other than the audience's, you can't control anything around other than the goals that he was setting, and specifically, the initial account setup, and then you really do those capturing that and just measuring that constantly. So it was fundamentally a game changing moment, essentially. And we started working on that at that point in time. And now with generative AI, I think a lot of the stuff can be very assisted, right. So we had two choices. One is we didn't really have very, we didn't have an option. As an agency, I think the classic obsolete agency model was obsolete. So we really had to reimagine this entire agency model. Because the classic search marketing where people have found out keywords, individually change bids is all gone. And there is no coming back, I can guarantee you that. That's all gone. Right. So you really at this point in time didn't have an option. But to really look at it from an AI first perspective, you really need almost like a strategic marketers who can use the value of their judgment, remember the first model, I shared the value of judgment, to identify the new path to grow your business. And what I'm going to share with you some of the stuff I throw share with you one of the very specific cases as well, if you do it well. And if you really have the right partners who have the AI technology, and who really understand the AI first thing, and most importantly, who have been able to get rid of this old classical search marketing baggage, which may be sort of just pulling them down, you are able to get results, which exactly justifies what I mentioned earlier, disproportionate outcomes, it just huge, it's much bigger than what it is, the only thing is only about a decimal percentage at this point in time are able to figure it out. 99.9% probably have not been able to figure it out. Right. And that's where the situation is. So I'm getting to the point. I mean, we're just the inception part. But this is the fantastic time. In fact, I would say that I knew that can't be a better time. If you really have the right partner, your results of the outcome that you're able to get out of Google advertising can be a very different level that you may not have ever experienced earlier. So that's the part which we have seen. But again, what we are trying to do I try to explain to you because we literally have to disrupt our agency model, we have to almost like break down disrupt our model. And this is what we did. We write to use AI to do two things. Only two things. The first is our goal was to do more with less. And the second goal was to boost ad campaign performance. That's it, nothing else. Right, doing more with less. So efficiency play. The second part is boosting ad campaign performance, right, the performance play and performance impact. And I'll share with you very tangible stuff. And I would request you I mean, this is what we have done doesn't mean that you need to exactly follow this. But I think there would be hopefully some cues, which can really help you to think about this utilizing AI as sort of challenge me essentially making it actionable. A lot more. The first thing we started off remember our agents CRM in 2004. Right. And we started with paid search management. We now have really broken down the entire paid back search management process or paid media and account management process. And we are saying that where exactly can we use chat GPT to save on time, or improve the quality of the output there as well. So we literally are just purposefully collectively as a team mapped out the entire account process our account management process and identified all the areas where we can use, I would say generative AI, because my journey is getting used a lot as well. Where can we really utilize this. So effectively, what we have done is we have re engineered our core process based on which our company and as I bragged about a few minutes back with all of you that we are in the top 27, we literally disrupted our entire model, which basically Day Zero for us. And it was extremely important for us to really do that, to just not only survive, but thrive in this new world. So we really re engineered the entire Ad Management process which rgpd My question to you is, all of you are running very successful businesses and you have your own internal processes, I would request you to get the entire team aligned that that's something that I would really emphasize getting everyone aligned everyone in one group, and really get into this process of redefining the process. Because if you are not doing it now, I can guarantee you your competition is or one of your competitors is and you will get left behind, and it will be impossible for you to catch up. Because this is remember, a vertical technology isn't even exponentially the vertical technology right? When the competition has done it before you it will be really difficult for you to compete and sort of catch up. So that's my hope that that that that message, I think I can't really put it even any more directly or emphatically is the key one that I wondered Is this is how we will disrupt our core business model right on which we had built this agency. So I'm pretty sure you can really start with maybe one of your not the core models, and really sort of work into really building those success stories and then extending across the organization as well. Whether it is merchandising, supply chain, every function, engineering, technology, everything web development. So that's my first thing that we have done at Netflix. And this is again, the start of the process. Because the entire process, you really have to ensure that you are masking all the data because you don't want to share anything with let's say, Microsoft or GPT and other stuff. So you have to really invest in all of those technological processes to build that walled garden. But this is really how you really build your own walled garden almost like a data garden in which you can really operate as the first step. The second thing that we have done, we recognize if we are to help our customers compete in the first page of Google, right? Remember, I mean 4600 advertisers or something like buy flowers online, right? And 4600, probably about 10, if you're lucky, 15, maybe 15 advertisers will be there on page one. And as we have seen in the conversation on search in next one year or so maybe six months, no one will go to page two. Right? So what is the point? So effectively, you have to be one of those 15 out of in a cluttered category, like let's say flowers, gifting, like, out of 4600, to really playing the race. So we drew an analogy. And the analogy was, it was almost like competing in a Formula One race. And in the Formula One race, either you are in the race or out of the race, there is no in between you most 100% cannot really ride a bicycle and hope to win the race, I can guarantee you will fail every time. Right. So that's I think what the analogy here was, so either, if you want to play on page one, you really have to be on the Formula One, if you're not, you're not playing on page one period, there is just no discussion at this point in time. And that you can do to really extract the real value from this new AI powered Google and AI tools like performance by writer, you will need the right AI tools. You will we go. So

Aaron Conant  33:29  

I'm just trying to like think about this as well. So as a brand, I'm advertising, I'm slowly going to see my performance go down. This is reality, I can't do it for a little while it's going to be fine. It's gonna get a little bit rocky. But slowly over time, it's going to go down and the conversations gonna have to be with the executive team around why. Right? Because they're going to ask questions. Why is that? Why isn't out in the reality is the tools have changed?

Udayan Bose  33:58  

Completely, completely. And that's about it. I mean, you really need to have AI tools that can extract the value from Google and Microsoft or whatever, anytime you live or Facebook, whatever it is matter. Now Amazon is coming up with its own AI modules resisting entry probably sort of launched out shortly as well. So without that, it's almost like riding a bike and hoping that even the Formula One race you will not

Aaron Conant  34:22  

on the bike was wrong. The bike was great for the longest.

Udayan Bose  34:25  

But the way you're competing, I think you really have to imagine that the space you were not competing earlier in that racetrack. The Formula One racetrack right. You are competing probably in a velodrome where the cyclists are cycling. Right. But at this point in time that that that entire racetrack has changed. It has become a Formula One racetrack now. Right. And that's the key component to compete that you don't have an option but to use AI tools. And that's where I think we we have been very successful. I talk to you about this vertical AI tool that we have built. But now we have some proven success and we are seeing it first tend for our customers, whether we manage as an agency full service agency work with about 120 customers, or our customers are using this as a DIY, they are using this technology because they have their large in house teams. But the impact is huge. If you really haven't vertical AI to the impact, I'll give you an example. This is essentially for a collection of our enterprise customers who are using AI. We, again, a lot of people have been talking about the performance packs doesn't work and all the stuff, the models, etc. We have made it work. And we have generated an extra 500k On an average for the enterprise businesses in eight months, which accounts about 2% of total search advertising revenue, just by using our Alexa in science AI to right, how many can really keep on fighting that part, essentially, no, I can really do it manually. But sorry, you can't you can't compete with Alexa insights, just simple, you can't deduct spirit, and they just know that nothing to do is just change. And we only proven it time. And again, why why you really need to have an AI marketing platform, it can be any it can be. And we would love if you use an excellent site, I'd be happy to really give you a demo and give you a free trial as well. But effectively, I mean, whatever use an AI tool, which can really extract this value from this new Google and Microsoft because you are no longer riding a bike, you are essentially driving a Formula One racecar. And that's what I really want to sort of really urge all of you to do. But there is one more thing that we have realized. And this is the third thing when we had to really go through a massive shift, or a change in terms of how we will operate as an agency, right. This is almost like a rethinking the entire agency model for the future. And that thing was inspired effectively by a very interesting incident. So this was in 2003. And one day in 2003, the fate of British Cycling changed. So what happened was, for the previous 100 years or 110 years, no British writer had ever won anything. I mean, they had won one gold medal in 1908. That's pretty much I think, between that that was I think what their performance was, it had gone so bad that bike manufacturers wouldn't even sell to the British cyclist in the fear of tarnishing their brands, I ended up going so bad. And at that point in time, I think the overall the association there, they hired this gentleman called Dave Brailsford. And he was hired as a performance coach, and he was hired to put the British Cycling on a new trajectory. The interesting part of what made him different from the previous coaches was he was focusing on what he called the aggregation of marginal gains of strategy, right? The philosophy of searching for a tiny margin of improvement in whatever you do. So the whole principle came down to the idea if you really broke down, right, we talked about the anatomy of a task, if you really broke down everything, and just did one person improvement everywhere. He showed that in one year, one person better everyday gets you to 37 times better, while one person works every day gets you to almost zero. And that's what we really have been at NetElixir have been really taken as almost like as our as our formula for success, to really drive this culture of experimentation. So we are lucky again, our technology we currently have four AI models built in all of them are vertically AI models, so you're not only sharing the data with anyone outside your organization. So returning single buyers returning next two months, we're able to predict the purchase value, we are able to predict as to when is your customer likely to churn, right, and we are able to build experiments constantly, not something about experimentation that and we are very big believers in this concept of deliberate experimentation. So when I say deliberate, I'm really focused only on improving those marginal games, right? I'm only doing experimenting just for the fun of it, I am really doing experimentation to help you get that 1% extra gain or benefit. And that's how we have been able to combine this power of AI and the human ingenuity to drive these aggregation of marginal gains to this process, which we call the deliberate experimentation. So rapid, deliberate experimentation as well. So that's where I think how we have been trying to rebuild NetElixir research, primarily because the obsolete the classical paid search model, which we just saw was valid, very valid till about one one year back, or even seven months back is no longer valid. And we are seeing tremendous gain from that I sort of shared with you the enterprise clients 2% Extra revenues, a lot of money 500k in eight months, is a huge amount of money for the same Google App spend without paying up a dime more to Google. So I think I'll just conclude by saying this conversational search is very clearly not really been is altering the search behavior forever. My question is, are you ready? You can't really win in this world unless until you have the right AI tools and the right application methods. But once we are able to crack that formula, you can really have a massive boost in that campaign performance. You should you should as an organization, future proof your marketing investment by remapping each and Every one of your operating processes just to identify as to what are the potential AI application area, we don't mean the right thing, right, just get started with it right, the first draft. And the last part is the AI plus the culture of experimentation, I would say deliberate experimentation really has become the greatest source of competitive advantage in this new world. Because Google and Microsoft and Facebook or Amazon, all of this will become a black box. So how do you really operate? I think you really have to have the right AI tools, and really have to have this culture of experimentation as such. So I wrap up my presentation there, I mean, with the second poll, and again, anyone, please feel free to add to the chat box. So British cyclist performance coach Dave Brailsford coined what what was the term what was the strategic term that he coined to really describe the value of small improvements? Please feel free to sort of add to the chat box. It is critical, I really want to emphasize this part. I see a lot of folks that are putting in perfect the the X The answer is right. It's called the aggregation of marginal gains. And by the way, just one more thing added which I wanted to really highlight with the team when I talk about personal mastery. So this is something which is extremely important. So what happened? What was the outcome? So it was interesting, because during the 10 year span after he implemented this strategy, and this is, by the way, the British Olympic winning team, led by now, Sir Charles White, if you're Chris White, essentially go sort of actually in the 22,008. Olympics, they are considered now as the most successful team in the history of cycling history. So if you really get the formula, right 178 World Championship medals and so on, right. And that's what I really wanted to explain this one, not only to emphasize the importance of experimentation and the deliberate experimentation and marginal games, but also if you get the formula, right, I talked about disproportionate games. And what we have done for our client is just a small thing using Alexa insights. 2%. But if I keep on aggregating that 2% earn every quarter, it adds up to a massive number for me, right, and it becomes absolutely a disproportionate overall impact over the time span. So that's what I would sort of leave all of you with. Happy to answer any questions or if anyone is interested in requesting a demo, while one of our product experts will be more than happy today to schedule a demo for you. There is no cost per demo. It's like complimentary. What Andy's urging you to do is, I think it's high time that you realize that you can't really ride a bike, and still compete in the Formula One race. And you can really scan this QR code right here to sort of participate there. But that's what I really wanted to offer to all of the all of the attendees today. And happy to really take any questions that you may have. At this point in time, we are also have an event coming up next equal experience. We this is our 10th year now. And I think I don't actually discuss this with you. So we bring some pretty reputed professors reclaim Professor this year, we are getting Professor Sheila and get from Columbia Business School. And Professor Adam alter from NYU Stern, both of them came up with their new books last month. And it's a very small, intimate set of award 3540 People 30 of them are probably CEOs of CMOS of eCommerce companies leading eCommerce companies and about 10 from the from the academic world. It's our way of giving back to the world of academics, which honestly speaking has given me a lot. So with that, I thank all of you for joining this. Let me see if we have any time for questions. But

Aaron Conant  43:40  

if you have questions, drop them in one of the first ones that come in, came in a little bit ago, you mentioned that predictions can be dissected. disaggregated, can you break down disaggregation? disaggregate judgment?

Udayan Bose  43:52  

Great points? So essentially, I think it's just tasks can be disaggregated. So not prediction. So if you really the anatomy of a task, the idea is the task has multiple components, right? One is the input input is the data that is sort of coming in. The second is that there are three broad components. One is the prediction, I already know because I've done this task before I do this, this is the outcome, that iconic step. Right. So that's the prediction component. The second component is a training component. How do I know that if I do something, this will get to a certain result, because I've been trained, my mind has been trained on that. So that's where the training component. The third component is something which gets unfortunately, marginalized many times which is the component of judgment. In many cases, we don't think beyond the box or outside the box, right? The outside the box thinking etc, etc. So that is one of the judgment parts. So I'm essentially trying to say that the entire task can be disaggregated. And you really have to understand as to where exactly AI can function better than you and we have to be humble enough to accept that AI can actually function better than us in the prediction area. While that frees up time to do something which no AI in the world and this paradigm can be which is essentially identifying those new paths, the new opportunities and new ideas, a breakthrough idea, right as Adam, author of NYU Stern calls it. So that's the part that I'm talking about. So it's not disaggregating predictions, it is aggregating the task. And this is how we are trying to do that.

Aaron Conant  45:17  

The other comments around, you know, I think the media spend where I think you can actually get smarter with your spend, and where the money's going. Yeah. Right. As you're doing the analysis of, you know, you can get to a better return on adspend. Based on the categories and what you're going after the customer that you're going after, right? Are they about? Are they're about to churn? Or are they potential to buy or low potential? Can we turn down that spend? You see, am I helping out in that as well, because I can see by helping with all the data crunching, but then the recommendations around, hey, divert two thirds of your budget to this group, because 100%,

Udayan Bose  45:58  

let me claim this, Aaron, so many of you would have read that the first party Chrome is going to be duplicating the third party cookies. And the first party would be eventually as to what Google has actually even announced recently, that 1% of all the Chrome browsers would start replicating the third party cookies, which will be only a first party environment from q1 of next year, q1 of 2024. Now, in this environment, I think this is a fascinating opportunity. And probably the best opportunity since the start of internet advertising, that you have gotten to build your own walled garden. And I'm just challenging you with something which is may seem very radical, I think that you, if you are really progressive enough, if you really have the right partners in place, if you really have the right engineering mindset in place, you can actually build your own walled garden. And within that ballpark, you define as to how much or I shouldn't say you define left technology, which is using your first party data, define how much should you be investing on Google? And how much should you investing on Facebook or whatever the other channels be? Right. So it would really lead to a golden age, effectively in terms of how you really do the media mix allocation. The reason I use Golden Age is primarily because currently the media mix modeling is based on very batched data, it is not done on real time data, right. And as a result of it, by the time you are able to implement the recommendations of a media mix, what has been shared that word, I mean, that media mix may be irrelevant. Now we believe in the AI combination, and with the tools like Alexa inside, that's what our holy grail is, we really want to build a smart digital marketing hub. And with that hub, you should have the control no one else, it's your business, it's your money. So you will need to controlling how much money should be really going to flowing into Google advertising at any point in time instantaneously. And how much money should be sort of flowing into, let's say, whatever media advertising, and so on and so forth. And that gives you power, because that gives you power to focus on just one thing. How can you make more money by with from the same investment that you are doing currently. And that's what I was referring to when I talked about getting a lot more, from less, or whatever, spending the same, and getting a lot more, essentially, that's the direction that we are collaborating with our clients, to really get them to that that future and get them be prepared for that first party future, which we believe if you really get the formula, right, very firmly believe. You can really build your own walled garden where you have significantly more control on your destiny, rather than letting Google define whether it will drop you off and base your folks are working hard and suddenly your ad gets dropped off and all this stuff, at least give some control back to the advertiser in this case.

Aaron Conant  48:41  

How should marketers start preparing to show up in AI in terms of site content?

Udayan Bose  48:48  

That's a great question. I'm show up in AI. I think one of the things is AI works on so many data points, the chat GPT for I believe uses about a mind boggling I don't know whatever the number 1 trillion data points is something some some ridiculous number, right? So effectively, can you really influence AI? I don't think much, but you can try. And at this point everyone is figuring out. So in terms of I think there is an opportunity for you to do two things. The first one that I would sort of highlight Korea is you can really use chat GPT to read through your website and give you recommendations. Because it's remember it is a smart assistant, right? And then you can really use the same prompt and say that act as if charge GPT is your customer. And whatever your customer let us say it's it's a male age group 35 to 44 income bracket, this, et cetera, et cetera. So act as if you are this customer persona. And now, how would you really go ahead and interact or engage with the website? He'll be blown away. I can guarantee you And you'll be blown away with the type of response it sort of comes up with. So that's the part which I think is really, really exciting in this one, right, you can literally use it as a very smart assistant, probably the smartest student you can ever imagine to have. That's the first part. The second part is, you can really go to find out what the customers are really looking for. I'll make a very broad statement, if you really look at the eCommerce area, that is a retail eCommerce area. And we have seen some very interesting statistics, let us say we have 100 page website, out of which probably there is a 9010 route, right? So 90% of all the activity happens on the 10% of the pages. It can really help you understand why it is happening on those central pages. But it can help help you to also refine it down to the most fundamental the first principal part, right? Get rid of all of the stuff that you don't need, because no one is going there. Right? Just because your vanity and you're doing this doesn't mean anything, right. So it can really help you get down to that first principles very well, by cutting through all the clutter. And that's I think the beauty and the important part that I really wanted to highlight. So these are two ways Korean companies are going Yeah, yeah.

Aaron Conant  51:11  

Awesome. Last one, I think we'd have time here. How can project manage product managers? Be better prepared, not project but product managers be better prepared to operate optimally in this new AI world?

Udayan Bose  51:24  

And that's, that's a great question again. So in terms of the product management part, if you really look at the fundamental one starts with, remember those user studies that we have those, we have those blind tests, user studies, and user group studies and so on. This can really help you create the first draft for the user studies, in many questions have been a part of some product user studies where in the the questions are very dry time, so bland, pretty, in many cases senseless as well. So I think the first draft can be a lot smarter, you can ask a lot smarter questions in terms of getting the first draft, because chan GPT will sort of preclude and just get to the point very quickly. The second part, which again is important is you can't really draw up the customer persona very well. When there's a person I don't remember only one because no one really has one customer persona. So multiple customer personas, essentially. And we want to really map out and you can go get down to what we call the principle of intersecting circles. So if you really have that Venn diagram, intersecting circles, what is the intersection point, which is the one thing which is common for in terms of the feedback or the preferences of all these different customer groups, and that one thing is something which will really help Product Manager immensely in terms of hone down and build that particular functionality or attribute or maybe double down and amplify that attribute very much in the product research? Yep.

Aaron Conant  52:47  

Yeah. Awesome. Well, I see we're pretty much right at time here. And again, Udayan, and thanks so much for your time today.

Udayan Bose  52:57  

Thanks for having me. I mean, yeah, I

Aaron Conant  52:58  

mean, it's a great partnership. And I encourage everybody, you'd like to know more about what they do more about NetElixir insights. 100% worth a follow up conversation, taking a look at it, this AI space is changing the landscape faster than I think any of us had really hoped. But that doesn't matter at this point in time, we really have to tackle it. And so, Udayan's team are just leaders in this space. And again, thanks again for your time today. And with that, I think we're gonna wrap up. Hope everybody has a fantastic Thursday, everybody, take care, stay safe, and look forward to having you at a future event. Thanks again Udayan. So much, very much. Thanks. We'll see you