Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Helping marketers peek inside consumer minds | Ranjan Kumar @ Entropik  image

Helping marketers peek inside consumer minds | Ranjan Kumar @ Entropik

Founder Thesis
Avatar
276 Plays2 years ago

Entropik is solving a truly hard problem of how to read the mind of a consumer. They have built an AI-powered platform that helps companies decode the emotional state of consumers by using neuroscience and machine learning. This conversation is a masterclass in building deep-tech products and scaling them globally.

Additional links:-

1.StartUp Central: Can AI gauge what you feel? Emotion AI provider Entropik says this

2.How Entropik is solving for the attention economy by humanising experiences with AI

3.A Young Bengaluru Startup Has Built World's 1st AI That Can Detect & Understand Your Emotions

Recommended
Transcript

Introduction to Emotion AI by Ranjan Kumar

00:00:00
Speaker
Hi everyone, I'm Ranjan, founder and CEO of TropicTech. He works in this fabulous space of Emotion AI.
00:00:19
Speaker
You would have often seen in Hollywood movies that they use a polygraph machine during interrogation to judge how truthful the suspect is. Imagine if a marketer could use a polygraph test to judge the response of his audience to his advertisement or product.
00:00:34
Speaker
Entropic is a company that makes this possible without needing an actual physical device. In this episode of the Founder Thesis Podcast, your host Akshay Dutt talks to Ranjan Kumar, the founder of Entropic. Ranjan loves solving heart problems, and Entropic is solving a truly heart problem of how to read the mind of a consumer. They have built an AI-powered platform that helps companies to decode the emotional state of consumers by using neuroscience and machine learning.
00:01:02
Speaker
This conversation is a masterclass in building deep tech products and scaling them globally.

Ranjan Kumar's Entrepreneurial Journey

00:01:07
Speaker
Entropic's most recent fundraise of $25 million is proof of the amazing business that Ranjan has built. Stay tuned and subscribe to the Founder Thesis Podcast and any audio streaming app to learn about how to solve heart problems using technology and first principles thinking.
00:01:22
Speaker
So when we were running oaparty.com, we had a payment gateway platform called Citrus Payments. And back then, we were about probably the first 10 customers of Citrus Payments.
00:01:37
Speaker
So I got an interview and I said that, hey, I probably learned, I knew how to, you know, take the risk and start, but I had no clue how to scale because I had never done sales in my life and sales, one of the most crucial things as a founder that you need to have. My co-founder was a lot more tech savvy, so we knew how to build platforms, we didn't know how to sell pay.
00:02:02
Speaker
So, I started looking out for job and next thing and then my co-founder as well. So, my co-founder landed in Oyo. So, he was one of the first guys in Oyo, took part in the growth function and he later ended up heading the growth function for them. I, on the contrary, joined, so I got a call from Jityan Gupta.
00:02:27
Speaker
He was like, hey, we are just raised from Sequoia and the series A had happened and the scale. I think it was back then, it was still that basement office in Swathi building in Pali. Sort of early start-up vibes, but hearing him was great. We met here in the cafe coffee day in MGA. That was my 30 minute interview with him. Like, when can you start? I said that tomorrow.
00:02:58
Speaker
And the conversation was pretty candid, right? So, I will switch things. And Jitin had a thing about people. I think he was, I still tell him that, super-sap about people, how he selects and how he judges them very quickly. So, I think it was primarily about what I'm here for and what I'm supposed to get.
00:03:22
Speaker
And what I'm here for was very clear that I'm going to go back to my entrepreneurial journey.

Citrus Payments Growth and Lessons Learned

00:03:28
Speaker
There's a part of my journey that I don't know how to scale and how to scale quickly. So I am here to learn that. And then I think Citrus was a fabulous time scaling. So I was the first guy driving the South business, and then I had seen such and such.
00:03:47
Speaker
selling to e-commerce companies the payment gateway service. So payment gateway services to all sort of e-commerce consumer internet essentially and then later also to the quite a bit of B2B as in the small part, credit consumer. So some of the customers over by Journey, so yeah we built out a Bangalore office about how it will
00:04:13
Speaker
the overall company scale to sizable 5-600 people. We got acquired by Naspers, the PayU in 2016. Culture-wise, Citrus was a cool place because for the first time I was having a luxury of actually working with people. Running my first company was you working alone.
00:04:36
Speaker
We have a set of team with whom you are working. And while a lot of people who has come from big companies actually cribbed a lot about like, hey, we don't have enough team and people for me, it was like, hey, this is from 1% to 10 people. This is as it gets. Right. So, uh, and, and I think the, there's no playbook playbook of sales as such back then. It's not that now you have SDR, VDR, MREs and all sorts of things. Right.
00:05:03
Speaker
Jiten was brutal and he was like, boss, this is your account, this is your number. Don't ask me money for going into events. Just sell it. And how do you sell it? Have you seen how bang guys sell it? That's how I sell it. And I think that very old school of, that was my first stint at learning sales.
00:05:28
Speaker
And the only thing Jiten was counting on was, hey, this guy's been on Subunu for two years. He would have learned a thing or two. And I think that was pretty much it. And I was ready to join tomorrow. So kind of started there. I think we did good wins in two years in South and that sort of.
00:05:51
Speaker
Quite a few countrywide accounts, in fact, right from Intras to Flipkarts to Amazon, a bunch of every one which we didn't believe earlier that we can actually win.

Founding of Entropic and Deep Tech Vision

00:06:02
Speaker
And I think one of the greatest realizations was that when you're selling, you're not selling to companies, you're selling to individuals. And if you think that you're selling to Amazon, it's just locked head. But if you think that you're selling to a guy who works in Amazon who lives next door to you, that's far more faster and efficient.
00:06:25
Speaker
I think all those quads and buzz around goes away and then you're just talking to another person who's like you. So, I think those were some of the early, early sort of learning size still carry and...
00:06:38
Speaker
That was my time at Citrus, made amazing friends. And then it was a cash exit. So everyone money. Something you got the equity, like you had ESOP. Yeah. So kind of places where ESOP really worked is now there's a lot more generic concept of it. But earlier ESOP used to be like just another thing in the KTE note valued and
00:07:03
Speaker
So yeah, I think Citrus was fabulous time. And upon exit to Citrus, I realized that, okay, I probably have learned a thing or two about sales. You left Citrus when it got acquired by the you? That's right. That's right. Like you thought that, okay, now I've learned enough time to get back to entrepreneurship. So two things, I think that one and as well as
00:07:28
Speaker
My clock was ticking, right? So this 2016, I've become 30. That's what I have talked about from my Wharton call. I'm five years into it. I think I've learned more than what travel I would have learned going to school, but I probably need to go and reapply. And I could see, I think at a certain point in time, you have too much energy and less wisdom.
00:07:54
Speaker
There is a time when you have too much wisdom but less of energy. So, you have to square off and plunge at a time when you have both had decent proportions. So, it was a good time. And somehow, since my ITC days, this thing of
00:08:11
Speaker
time value of things has always been constant to me, as in time is a crazy variable to deal with. And unless you play right with it, I find myself always losing.

Exploring Deep Tech and Human Emotions

00:08:22
Speaker
So I went back to starting my next company after Citrus again.
00:08:28
Speaker
And did you have an idea in mind, or you wanted to spend a few months thinking? I met this person, Mr. Dilip Hart. He was a president of operations at TSW Group and a person in his 50s and with wisdom and looking to invest in young entrepreneurs.
00:08:54
Speaker
and was wanting to be part of it more than anything. And I've never met a believer like him. So this was 2016, early when the exit information was out and the process was going on. So I met him and he was like, we talked about various things and various kind of ideas.
00:09:13
Speaker
things. I think it was very exciting the way he chose to believe in me over a set of conversations of various ideas and things. And I think I earlier thought that probably he doesn't know that much about space and all and I got him tricked. But later I realized that he had this insane ability to read through people.
00:09:41
Speaker
And there's nothing that I was thinking that he was not aware of.
00:09:46
Speaker
He sort of, he understood that it's fairly near age, you'll probably think like that. But I think what was fascinating to him was I was thinking of building something which is fundamentally different and very unique and very IP based and very tech tech heavy. And there was a burn about my first company, which was primarily not being able to find differentiator early on.
00:10:14
Speaker
So, I was clear that I will die fighting a problem statement than a competition. That's a nice approach. Very nice. And it was something that came out of, you know, from the days of my JE preparation IIT exams, whenever there used to be this very hard papers, right, I was super happy because my margin of error was very high. You know, a hard paper where your cut-offs are 10 marks out of 100.
00:10:44
Speaker
you I can I will do lot better in those than the exams which were easier because here and there is small plus one marks minus one marks and you are out of the game right so I hence always learned that pick a super hard problem even if you are solving 40% of it you are based out there
00:11:03
Speaker
And that was very counterintuitive to how businesses were done because consumer internet among my IT peers, the consumer internet was the easiest thing to get started with, right? Or even B2B if you're looking at the CRMs were the easiest to start most known.
00:11:21
Speaker
And, uh, here I was thinking of going something deep, take whatever, right. And then, um, kind of, so I met him and I think very inspiring set of conversation over some of science. Uh, are we still fancy his balcony where we have the sense and we talked about things and then we bought a whiteboard. And one day I was whiteboarding, what three hours to him about things that the way this business will grow and what it can potentially be. And then he said that rather than whatever you will do, I will put in money.
00:11:51
Speaker
I'm not questioning anymore about what you want to build. That's your call. And he asked me that, why should I invest money with you? And I said that, sir, I'm about 25 years richer than you. And he got aligned with the time value of things. And I said that,
00:12:16
Speaker
Given 25 more years to me, I'll be at a different place and given 25 years past if you have to go back, then you would do things very differently. So here's your chance and here's my chance. And that's 90 instead of conversation.
00:12:33
Speaker
And I think that led to me starting this um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um, um,
00:12:55
Speaker
So we got to pay you back and you're like, let it be. I know that has become this, that amazing, amazing person. So I think his confidence in me was very pathbreaking at that time. Because I think you always have that sense of doubt, right? And then you believe that somebody who has seen it so much is trusting you with their money means something probably. And yeah, that's how instruction started.
00:13:23
Speaker
But knowing that you want to do something in deep tech is still quite a distance away from an actual business. How did that happen? So there were a few philosophies around what I want to do next and it was more philosophical than actually in the shape of business. But that actually still stays as the boundary conditions around which we built and traffic.
00:13:47
Speaker
So for me it was we to be because that's what I want. I knew how to sell to enterprise because they've got money and I don't want to sweat my blood where there's no money. The third was super heavy differentiation and the best way to institutionalize it was we decided we'll only build things which has got IP value.
00:14:17
Speaker
it holds no IP value, then we'll not build that. So my how part was always very clear that if there's a problem, I'm not solving it through which is substantially differentiated. I'll not do it. And the last was that it should be large enough problem and sizable enough impact to generate. And hence, choose a problem which is not a market use problem or a product guys problem, but choose something which is a human's problem.
00:14:46
Speaker
And that is as wide as your market has to be. So that you never walk into a fund risk conversation being asked, hey, what's your market size? And you don't have a slide which is building up. And then your top-downs and bottoms up and all those kind of things. And looking at all of it and thinking a lot more first principally, what I realized is that the world is going to go digital. And the nature of conversation is going to change.
00:15:17
Speaker
which means the way we have conversations one-on-one and otherwise, we measured a lot of gestures, we measured a lot of, not just about what people said, that's probably about 5% of information, but mostly about how they said it, what they said, what was the gesture, where they are attentive, all that stuff, right? And then it stuck to me that as we go digital and everything goes digital, this is gonna change forever. How can people still have that
00:15:46
Speaker
intelligence in place and by that I mean not really businesses have that intelligence but just me and you having that conversation can have that intelligence and and then it can obviously transfer into various form factors of business and all those will come together. So just thinking at this level and saying that hence there is a need and there's a dire need and likely need of
00:16:08
Speaker
being able to understand human behavior and emotions on its first principle basis at scale. And can we do this? So just this much was known. It was fascinating. It was inspiring. It was profound to me. And I felt that I want to commit something like this for a good amount of time. And I researched about this space, looked at how many guys I've ever practiced, and found my hard problem statement with 10 marks cut off.
00:16:39
Speaker
Okay. So a couple of thoughts I have here. One thought is I'm getting, you know, you in citrus, you realize that you're not selling to organizations, you're selling to human beings. So, so that, that would have like the mindset that it's important to be able to read human beings and their behaviors, because that will help you do sales better. So that might have been the trigger for this.
00:17:03
Speaker
Yeah, so essentially when I walked into sales journey of mine, my sales rate was very high and I used to close about 80% of the deal that I'll walk into and against about 30-40% conversion otherwise. I think what I realized is that my success, if I'm getting a meeting,
00:17:23
Speaker
If I get to meet someone, I'll definitely end up closing. And somehow that conviction kept on building over two years. And I realized a few things where I kept... I used to predict whether this guy is going to close or not and see whether this guy close or not. And then I learned that what are the triggers through which I realized that it is going to close or not.
00:17:43
Speaker
And those never was things that he wrote in mail or things that he said was always about that odd questions which he had answered in a certain way. Give you examples of this. So you had this long chat about pricing and all things and otherwise
00:18:05
Speaker
And then you have this weird question. And a classic example, we were doing this with Commonflow. And that was about an 80 crore transaction volume deal that we were trying to do. And we were having this conversation with a guy who was a CDO product manager. And pretty much the conversation was written to various things of product and stuff. And then my question was, how do you think the startup is going to do this whole startup space, the way it is happening, is going to do
00:18:34
Speaker
down the road and he jumped off the table and he talked about everything entrepreneurship around and everything out of the space is growing. What is going to happen? What kind of products is going to come? Is it going to be? Is it going to be this that all sort of stuff he talked about and he spoke 30 minutes. And then we talked about it as in what I did prior to this and he talked about what he has done prior to turn out. Both of us were entrepreneurs who turn into jobs. You know, uh,
00:19:04
Speaker
That one odd conversation, you knew heel convert. Yeah, I knew heel convert. And then there were conversations like, I think this guy was here for soon car and we were having a conversation.
00:19:19
Speaker
And his question was, I used to always have this, you know, towards the end of the conversation, some trick pointers to see whether he nudges, does not nudge. And to get responses of it with facial gestures and all that coming together was always like, okay, this guy's sorted now. I felt like, okay. There are some other things that always felt that selling is a very emotional process. In fact, decision making is emotional.
00:19:47
Speaker
We have made decisions long before we knew we have made decisions. Oh yeah, that's so true. And then the rest of the time we are trying to find out our true points. And those who are very intuitive and left-brained, they are always parallel thinkers and very very quick at this. And you see that with limited information, the way they'll be able to make decisions is crazy.
00:20:09
Speaker
And, um, I, I felt that that's fascinating. That's, that can accelerate everything. And that's very profound. So, so, you know, those were the things sort of, when I started, I felt that, you know, understanding behavior is going to be so massive, so massive. And did you, uh,
00:20:29
Speaker
Think of the use case for this as people doing sales over Zoom calls. Therefore, because the video is being recorded, a software could give nudges to the salesperson. Did you think of that as a use case for it? Obviously, this will only work on video. You need it to capture video.
00:20:51
Speaker
No, I didn't have the understanding of form factors back then, right? So I was just very super convinced by the fact that the technology is going to be massive and it doesn't matter what form factor. And I was 2000% sure that I will build a business out of it. So I went after tech.
00:21:08
Speaker
And to start with, I started looking at, so this is the landscape of emotion or behavior,

Innovation with EEG Technology

00:21:14
Speaker
right? There's brain, and you have sense organs, so brain, face, voice, eye, touch, and last is text. So these are the seven ways in which you can never understand behavior, because there are only seven things through which we express.
00:21:30
Speaker
Just go through the list again. Brain. Brain, its face, its eye, its voice, its touch and its text. So these are seven things through which we are ever going to express.
00:21:45
Speaker
I looked at how people are expressing and what kind of intelligence is built into it. So you have NLP for text. So you understand the text, how people have written and what they have written and semantics and things built around it. So a lot of chatbot companies were working on that and that was already evolving space. And I said that I'm not going to approach from that side.
00:22:06
Speaker
There's a bunch of things about voice also that was happening. And I said that I'm not going to approach things from that side. The most scalable and the easiest and probably the more done dusted space was that much, the voice and the text into the scene. And I said that I want a hard problem. I want to solve it in a much more IP capture way. So I started from the other end. So I looked at brain because that's the most accurate form factor and more.
00:22:33
Speaker
defined and proven sort of a thing and truly what people behave, can we measure that. So, in that I looked at this technology called as EG, electroecmephilogra. It has a EG headband that one can wear, it tracks real pulses, how your neurons are firing.
00:22:51
Speaker
And basically it gives you a bunch of those the raw brainwave data called as frequency ranges and alpha beta gamma wave. So when you're sleeping your alpha is very high, when you're calm your alpha is very high, when you're anxious your beta is very bad. Deltas are bad.
00:23:06
Speaker
There's quite some research work that has happened into it but that was never productized. So the first thing we went about building was this software which plucks when like this is like your headphone like that it's a back and it plucks with like a Bluetooth with all the EJ devices.
00:23:24
Speaker
sort of tracks on a second-by-second basis, data points like attention, engagement, emotions, happy, sad, excited, relaxed, bored. And then we validated each of it. So let's say I'm watching and I try to be purposefully very focused and see whether the dialer moves or not. If I'm in a happy mode, I do that. In control testing, I get five people
00:23:47
Speaker
disgruntled, create those scenarios and then sort of let them hear that. So kind of a lot of women's learning on neuroscience is where we went. And the first ever small tool that we created was brain brain mapping based consumer studies. And Nielsen had a neuroscience lab back then. And Nielsen had a company they acquired called Neurofocus.
00:24:16
Speaker
and used to use EEG devices, but they used to have manually analyzed all this data. So we had a software which gives you end outcome like attention engagement emotions, which can be locked more easily applied.
00:24:27
Speaker
Instead of customer and we got using that we got part of VSTART program, Viacom 18 had accelerated program and we got into that and we started testing content for them to tell them whether your movie trailers are people are at in attention levels are high at a certain point and low at a certain point or where they are excited, where they're happy.
00:24:50
Speaker
that changed the way content testing was ever done for them. So we started doing that for their trailers, their episodes, their movies, their short form. So we invite the same focus group studies that they used to do, right? But essentially, instead of surveys, we said that, hey, why don't you wear this and do this in a central location kind of environment.
00:25:11
Speaker
We did this for Netflix, we did this for Amazon Prime. These were decent clients and Nielsen was outsourcing it to you, something like that. No, no, no. We were directly working with these guys. So, we were part of Accelerator, so we got a chance to do a couple of projects and then that led to doing a lot of business with them. Same thing we scaled to other ones because you had a case study from Viacom. And this Accelerator was in the US.
00:25:36
Speaker
No, this is, this was in India. This is Bombay. Okay. Okay. Okay. Okay. And then, then in the same year we got, we became part of SAP accelerator, Accenture accelerator, target retail, and pretty much about eight or 10 accelerator programs we were part of. So a lot of giant corporates where the, the verticals and the horizontals were there. And I think we had a depth and uniqueness of the deck.
00:26:02
Speaker
And those two came beautifully to get my first 20 customers.
00:26:09
Speaker
because they had ready customers who were having these problem statements and these were large. All we had to do this is, you know, this fancy magneto stuff you wear and you talk about it and you show the demo and then everyone loved it, right? And then tried and then just had to believe the power of tech and it happens everyone loved it and a lot of them very fascinated. Coke CEO, we met, I think, a bunch of companies, CEOs who used to come in
00:26:36
Speaker
Accenture, you're the poster guys to actually go out and you know, view the headset to this, right? So this, everyone, it is very exponential in nature. So kind of that happened. And as we started doing that, what was the timeline for this? Like when you got into all these was, this was 17, 2017. So 16, we started renting this one. And you were using your cash out from citrus for funding this.
00:27:03
Speaker
Yeah, so all money that I had got, including what my wife had got. She was like, you took out my ice of money, you also took out your ice of money. One more question I want to ask.
00:27:23
Speaker
EEG is like a very much in use medical device or is it a niche device? If I walk into a hospital, for example, would I find an EEG then? What's the use case for an EEG? EEG has got various grades and the rate of it is something that is being used by most of the neuroscience departments in the hospitals.
00:27:47
Speaker
Basically, it's a scanner to your mental activity. And you have a clinical rate 512 channel EEG, which is like a cap with lobes and wires, right? Okay, there's a lot of connections that tend to do that. Okay, so that is basically to do yours. It's like effective scanner. So MRI is the most, is much more accurate version, EEG is a lot more portable version of that.
00:28:14
Speaker
That's been used clinically and in the cadmium. And what does a doctor want from an EEG test? Like what does it help him diagnose? So what data does it give to a doctor? So it gives doctor a data sense of like which parts of your brain is a lot more triggered. If there's any, then measure anxiety, measure epilepsy, measure dyslexia. Those are some of the things that they use it for.
00:28:42
Speaker
But there is a lot more port table. This was a 500 channel kind of an easier device, which is a cap and makes you look like a lab rat. But there is a three channel version of it, which has neural sky had launched, which is just like a band what can be in there, like a headphone, right? So it's a lot of port where people did find it difficult.
00:29:04
Speaker
Also clinically what is being used is wet electrode. So it essentially they'll put in glues like they put in for ECG like that. This dry version was like very simple band and totally, totally safe and all things, right? But no contact with the skin needed. Like with skin in it and those kind of stuff. So it was super powerful. And then I think building a software stack on it, that was immensely valuable.
00:29:33
Speaker
And how much does this device cost? This new device, which keeps? It's $100 anyway. It was very cost-effective, right? So $1 is the cost of hardware, and you can eat directly over any number of sessions.
00:29:51
Speaker
And Viacom will play you bomb for doing this kind of testing with 100 people. So what we started doing with Viacom was on content testing. What we started doing with Accenture was a lot of UX testing on the website, mobile apps. Okay, so somebody would wear that band and browse a website. So you can see if he's getting frustrated by the experience or if he's getting delighted.
00:30:16
Speaker
We also understood at what point in time they were frustrated and excited because they were having a screen recording going on. So I know that at a point where you have to make a purchase or make this search, you are not able to find out, right? And those are the points where the peaks are coming. So it was a lot more pinpointed, actionable to UX guys. Identify as a creative director, it was very pinpointed, okay, whenever this character comes in, this person goes out of box, right? It is going searching all the brainwave data.
00:30:43
Speaker
The actionability of it was super critical to them. Because doing every testing and test, you get like A is greater than B. These guys said, okay, what is not working? So they got a lot more precise value. And then we worked with Tata consumer care using AG. And so today there's a brand called Tata Sampan.
00:31:04
Speaker
they were trying to launch it and they were struggling with all things right from low vote to back as designed to all sorts of stuff. So we did a testing where we created this 360 degree virtual environment and you're wearing this Oculus and then you're wearing that Engie and you're walking in the store as you're doing so. We are tracking your neural response as you're walking into the aisles of the store. So if you are looking at a certain product and your neural senses are going up, which means that you're noticing it,
00:31:34
Speaker
So we got a very detailed view of how shopper researchers' software study was done. So I had content testing or app testing. I had shopper research and shopper testing. We have done UX testing. And last was we worked with ITC because it was my old customers for product testing as well. So they had Laze and Bingo and a bunch of these ships as you are eating what kind of sensory movements are happening. And then how do you work here? So four use cases were ready.
00:32:04
Speaker
rather than shopping ads.
00:32:07
Speaker
As we did so, we had a decent ad hoc business, which was project by project running by the time we hit 2018. And we did about four or five crores of business. One more question about the product. Were there like a standard set of parameters on which some sort of graph would be built with a time series on it? What were those parameters? It was attention.
00:32:36
Speaker
Engagement, we saw them also call it excitement value as such. And then there was five emotions, happy, sad, excited, relaxed and bored. So out of this five emotions, two are positive, two are negative. So happy and excited are positive. Now happy and relaxed is positive.
00:32:58
Speaker
board and side is negative, and one is neutral. And between happy and relaxed also, one is high intensity motion, second is low intensity motion. So you're looking at valence to intensity, kind of a two by two matrix, and each of this is one point into that. So, I don't understand that word valence. What do you mean by that?
00:33:20
Speaker
Balance is a measure of positive emotions, so high balance is high positive, low back is low positive. And intensity is how intensely you felt it, right? So, happy is a very positive, very high intensity motion. But look at relaxed, relaxed is a very positive motion but low in intensity.
00:33:40
Speaker
Likewise, bored is a negative emotion, but low in intensity. And sad is an extreme end of emotion. So this is called a sub-complex model of emotion. So we started mapping one point in each part. And there would be a time series, and the graph would be up and down. At this point, the happy went up. Or at this point, the sad spiked. And for content testing, even sad would be good. Because often, they want you to have
00:34:09
Speaker
intensive versions, be it positive or negative. So one of the things we actually broke the myth for some of the content companies including when we worked with Star, we worked with Viacom and was that happy people does not lead to conversions, right?
00:34:30
Speaker
It was the happy emotion that leads to conversion. So I think the learning was, and we did a bunch of testing for them, and you know, one of the creative directors and the CMOs were there, and usually creative people are always very egoistic, like, hey, I have building shows for 25 years. Don't bring a technology and tell me this. So you can't overstep onto that feed, but kind of,
00:34:54
Speaker
We got data off about last three projects and we got everyone in the room and see more than everyone else with that. So a project here is like a TV show or just an ad film? Like what were you testing? The full series, a full movie? No, the pilot episodes were testing or if the show is not doing good, last few episodes were testing. You also do trailers or promos.
00:35:17
Speaker
So you started a concept stage when you just have scripts and you are building animatic out of it and from there you test and then you subsequently. Because once you produce something, it's hard to go back. You have gotten paid money to Salman Khan. So we went to that and I heard everyone in the room and we said that here's the most breathtaking thing that you can listen to. And we got in all this data and showed exactly that.
00:35:45
Speaker
It's not the happy motion that's driving the content. It's the swing between happy to negative that is driving the content. So bigger the delta, the more, the better the shows are there for you. And the way we had done is that every time this show shows red air, we actually got the.
00:36:03
Speaker
bark rating of deers and tab rating of deers. So bark has got a rating system every week. Bar rating is like viewership. How many people watched?
00:36:17
Speaker
So you have your viewership data once the show is live. And every week, they get the data from bar or tab. And we got this data together. And we sort of put it together to put a correlation. And it showed that 95% of the time, it is matching with the swing between the emotions. So that's massive. And I think more than anything, I used to be very thrilled at these outcomes, right? Like, oh, I have a part where I come over. This is path breaking. This is disruptive.
00:36:44
Speaker
Oh, I got a meeting with CEO of TAM. TAM is a computer to bark and to TAM CULB Krishna. Just for the listeners to listen, TAM essentially has viewers all over the country, like anonymous viewers, and they installed some device so that they can track what the viewer is watching. And then they create a rating for
00:37:08
Speaker
different shows like this show was viewed how much it or they give some sort of ranking or something like that. Yeah, so I met Elvi Krishna, Tam CEO and I showed the same outcome to him and he was jumping off his chair that
00:37:25
Speaker
Hey, this is massive because this has got ability to, first time being, so we province science-wise kind of positioning operating system that has not happened. And also some of these companies like Tam and Bach, we have put on the scanner for a long time in terms of, are you really getting the rating right, right? Like a broadcast is to continuously put that pressure on there.
00:37:48
Speaker
And while they do their job, it's essentially a sample group data. And they have very wide panel coverage. So what we did a project with Tam was we ran this test of content. So they had a huge repository of content.
00:38:06
Speaker
and they also had a huge number of panel. At TAM is an instant subsidiary. It's an in-user encounter. So, we did with their panel a large set of exercise of benchmarking where with over 100,000 users, we went about testing about hundreds of D or shows with them and then we created a benchmark around which we got the results and 92% time we were able to correlate
00:38:31
Speaker
What leads to what? And we try to reverse engineer that to create a predictive system and get a content in. And can I predict whether it's likely to do good or not? And what are the elements which is not likely to do good? And we could find that too, if there's a good judgment.
00:38:48
Speaker
And you put the device on 100,000 viewers. That is massive. So 100,000 users across a period of a year. So a lot of these were already the panels for these projects that they would do. So you got it, and then for about 150 odd content, we did this. So each content is watched by a certain number of people, 100, 200.
00:39:15
Speaker
So we did this and we got this data together and with that we found it very exciting and we went back to a lot of these broadcasters and they liked what we do and creative directors hated us because you have a critical system of what is likely to work and not work and there's art to it and there's not an only science to it so all this kind of and we took our positioning saying that we are just a voice of customer you're not saying what to do it's just presenting things and
00:39:45
Speaker
So I think that was a great learning. But as we went about this, and I think one of the things that you realize that there's a lot of manual stuff. So you have this was purely science driving value and see as a process of doing this for a year and a half. I think and I was very fascinated about the first principle reasoning. I always felt that scale all things will achieve all that is fine. But at the first principle level,
00:40:09
Speaker
Is the value enough proven? Is it disruptive enough? And is that value substantial for those guys to pay money? And what is that quantum money looks like? And what is the acceptance of the outcome? And is it consistent enough? These are the things that I wanted to check from a use case value proposition and at the highest level to a CMO, to a creative director.
00:40:31
Speaker
Once I had this data point and when we were saying about there's a show for which we did this, the show went live and started seeing 3X viewership than it was in past. There were promos on which a million dollars was spent prior to the launch and they had six versions of the promos, various kinds of trailers that are coming out and we selected that these are the two that you got to bet your money on.
00:40:57
Speaker
And we saved about $350,000 at a scale of 1 million. So 35% kind of ROI left on our decision making by bringing data into the picture. So that is the value that we got sub-stylishated. And it was massive to see. With consumer care, TARA consumer care, we could see them making a decision where the new back has got 47% more pickup lift from the stores.
00:41:21
Speaker
And this is the Starbazaar store for which they had the track of data. The launch subsequently that whole pack design for about 13,000 SKUs. So basically with Tardak consumer care, we could get about package design insights was giving me 47% more lift on pickup rates. The launch rate was 13,000 SKUs.
00:41:44
Speaker
and the value was proven with Accenture. All this process took about three years to us and 2018 was a time when the value proposition was very clear to us. This is strong value. And then for the first time, I downloaded the consumer research market landscape and market size report and I saw that. Then it was just, okay, I don't know enough because I

Shift to Scalable Software Solutions

00:42:08
Speaker
didn't want to know. I didn't want to know what sales because I just want to figure out myself.
00:42:13
Speaker
And about that, it's an 80 billion dollar market. Consumer research, there are two ways it has been done. One is surveys where I ask you question of A versus B, what you like on a scale of 10. People have always answered 8.5 and you don't get a clear actionable. The second is I do this focus group discussion and interviews where you get a bit of wise, but then people say what you want to hear. So bias is a big problem. Automation is a big problem.
00:42:39
Speaker
So then we looked at more market, we are doing things. Meanwhile, in 1718, we learned that hardware is going to be innovative to scale. So how can I get the same outcome using without the hardware? So I start looking at computer vision, the facial coding. Using camera, can I analyze the face expression? The company's like effective and realized it were tried and passed, looked at the papers and stuff. Understood the apocalypse. All of them had trained on face image datasets.
00:43:08
Speaker
So the face we analyze 58 odd gestures and expression and that is what the model is built on. We had an advantage of brainwave. So we went about cross training face, facial coding data, face expression data with brain data, which means what a face is saying at any point in time and what my brain is saying at any point in time. Does this match? Does this vary?
00:43:33
Speaker
And my training model should comprise of cross training these two. It solved two very straight-of-art problems that no one ever has solved. One is how do you deal with poker face? And we realized that there is a set of micro-expressions for which we had never been able to label because it was not possible to label. There was no reference point source to label. But when you have a brainwave data, you were able to label.
00:43:58
Speaker
Second, we were to deal with cultural nuances a lot better. A Caucasian face versus a different kind of face actually had very different meaning to things. But fortunately, unfortunately, the brain is constant across human race. This is not.
00:44:14
Speaker
So you can require more robustness into your data model. And we run this test and we lead to cross passing some of the accuracies than anyone ever had. So facial coding, this is how we got built. As we read facial coding because we are capturing the face expression.
00:44:35
Speaker
This is machine learning through which you build this. That takes a lot of data, right? So that meant that you were recording videos of people as they were doing the market results, like somebody who is testing the UX.
00:44:53
Speaker
Yeah, what we started with while we were, no, we were just doing brainwave mapping. But what we started with doing is saying that this was a lot more controlled test where we invested to get in users and do these tests where I will grow both brainwave data and phase data, take the user consent to do that both. So this was the process for us to get that. And it was an investment from our side.
00:45:18
Speaker
As we got to that point from facial coding, a certain level of maturity, we realized that from facial coding, what you get to understand is that let's say this video is running at fifth second to 12 second is where your attention is low. But in that fifth to 12 second, there's so many frames and things coming. Can my insight be more graphical? And then we started looking at eye tracking. So eye tracking gives me ability to tell you heat map twice where exactly people were looking at any particular point.
00:45:45
Speaker
So that gives me a sense now that when your attention was low, you were looking at Shahrukh Khan and not the product. And likewise, when you are in UX, when you're looking at the product image or reviews is when your attention was highest. And likewise, in package design, you were actually looking at Sanjeev Kapoor in that Tata Sampan package brand than anything else. So with that, what exactly was driving that is what eye tracking brought in.
00:46:10
Speaker
So that's how facial coding and eye tracking together become the most powerful combo. And with all the customers we replaced brainwave mapping to these two, proving out that this gives the same outcome at scale. The only thing that we had to do is increase the number of respondents a bit. Brainwave mapping used to be with 30 people. This was happening with 150 people to capture a larger frame.
00:46:32
Speaker
We are about 92% accurate in each of these technologies. We launched world's first mobile-based eye tracking. And then power of multimodal emotion, which is brain, face, voice, eye together, I think has always been our winner, gotten quoted as world's first multimodal emotionally eye player. And then some of these happened. And 19 is when we launched a platform called as Effect Lab.
00:47:00
Speaker
which is a more SaaS platform, consumer research, SaaS platform, which takes care of all your content testing, ad testing, shopper research, package testing, all that stuff at one platform. And it's a SaaS. And we said that, hey, head of consumer insights, here's a product which is a one-stop solution, replacing all your surveys that you have been doing for past. And this sells to CM organization within brands. They have the larger appetite. Same to content and broadcasting industry.
00:47:30
Speaker
And the same carve out of that, we call it Effect UX. That is something we sold to all the digital first companies and their product. And to say that you have a user research team and a UX team, this might work whole sides into how you edge. And this becomes my SaaS proposition, so to say. So 19 we launched, a Clap 20 we launched, Effect UX.
00:47:52
Speaker
I want to understand these products a little better. So how it would work is like if I land on a website, then the website will take permission to view my camera feed as in? The way it works is that essentially let's say a consumer insights manager of a PNG wants to test out the ad. They configure all that and launch. The link goes to the users. So we have about 60 million respondents across 120 countries as part of my platform.
00:48:19
Speaker
So now we're dealing with all online users, no offline stuff, right? So these users will get a link. You click on that link. It asks for camera permission. You took the camera, and you watch that. That's it. As you watch that, I track your face expression, eye movement, all that. And I have that all coming into the dashboard. So to brands, they can see sliced and diced data with the board.
00:48:44
Speaker
Likewise, N doesn't matter. It was ad and hints, ad testing, had it been a UX or a package design. The only this we can do is product testing because product testing is like you're eating laser, having COVID. Okay. Right. While consumption is happening, that is hard too. Yeah. Anything audio, visual and stimulus is what we can solve.
00:49:07
Speaker
That is what the SAS platform is about. So Effect Lab, something that we launched in. How did you build a 60 million pool of respondents?
00:49:17
Speaker
We didn't pay. There are companies which are in business of building the panel. So you have Dynital, Lucid, Synth. These are the guys who anyways these brands use. What we did was we natively integrated all the panels in our platform and all taken together. This is the size. For what matters to brand, they're getting all responded at one place. So that's what we do. So like the brand needs to pay these panel companies separately and pay you separately. No, they pay us. We take care of everything.
00:49:47
Speaker
2019 was Effect Lab, launched it as a fast platform, first time hired a sales team. Till then we didn't have a sales team ever. I was the sales guy with few SDRs and such. And then we were not trying to sell, we were trying to see the value. And 19 onwards we started selling and Effect UX was launched in 2020.
00:50:10
Speaker
As we launched these two things and pandemic hit and while this was scaling in its own and we got sales team and GTM motions and all things together. Something very trusting happened during pandemic. In addition to quantitative service that people used to do which got replaced by effect lab and effective x respectively.
00:50:29
Speaker
People were also doing this qualitative research. So your direct interviews and focus group discussions. And also, so that all went remote. There was no focus group discussion which was in person during pandemic. And people could see that it is 50% cheaper, it's a lot more faster, a lot more better. Brands were loving it. And it is coming at such low cost with software. And the company like Suzy and people were using Zoom and Teams and all this stuff that people started using. And that was a massive surge of volumes.
00:50:59
Speaker
So every brand started talking to us that, hey, do you have something around this? And I remember that you're facial coding and eye tracking. Can you analyze this video conversations and stuff? And quickly, we wanted to start investing into voice because voice was the missing component.
00:51:16
Speaker
And we said that we have brain, face, eye, get into voice. So with voice, what we did, speech to text is done, dusted. So we bought third parties. And voice to melody is where we started putting our R&D efforts to say that it is not just what is being said, it is about how it's being said.
00:51:33
Speaker
So, in about six months into pandemic, we were done with voice tonality to a greater bit because the data was already there and we had validation mechanisms already placed.

Advanced Emotion Metrics and Market Expansion

00:51:43
Speaker
So, first version of voice tonality we got out and 2020 is better. What are the voice tonality also measures those same five parameters that you mentioned? Voice tonality basically measures two things. It's basically to do with confidence score.
00:51:59
Speaker
which means what is the truthness in what you are saying, right? With what as you are saying what you're saying. Yes, I might say I like the Pepsi or I can say I really like the Pepsi. Right? So these are different things to say. But so kind of we analyze voice tonality and then you said two things. One is confidence. What's the second?
00:52:22
Speaker
Second is more of a directional emotion which is positive or negative by voice planning. And this was again machine learning like trained through machine learning. This is again machine learning. How was the data labeled? You need labeling, you need draw data and then you need some labeling for the algorithm to learn.
00:52:47
Speaker
Yeah. So fortunately we had a fair bit of recordings from, from our past, uh, uh, whatever qualitative interview that used to follow after the content testing that we used to do. And, and we had those and we, and we had also corresponding outcomes in terms of emotions. So, uh, by, by brain and face, right? So that became my friends plan to begin with. And that got, and some of these cross training sort of got us faster.
00:53:13
Speaker
than anything. And then subsequently, we obviously went about having our own labeling mechanisms and things, and voice is still a work in progress, but I think we fairly saw it to a certain extent that the speech to text was fairly darned-usted, so we looked at the major stacks and we sort of got them together.
00:53:30
Speaker
So with these together, we said that, okay, all the consumer interviews that's happening is a big upsell for me to all the customers, existing customer brands that you work with. So we launched something called as a decode as a product. It's launched in 2021.
00:53:45
Speaker
What it does is that it allows you to, and it plucks with whether you are using Zoom, Teams, any of the video conversation platform. You can take those interviews there. It essentially streams in all videos from there. All that comes together in this one single library, and it analyzes your face, voice, speech to text, everything.
00:54:06
Speaker
and things which are of relevance, it picks it up, creates highlights. So you have an auto-generated sort of a dashboard, which gives you a quick sense of what actually happened in the conversation. And then it became like a super hit for all the brands for their qualitative research of what they were using. Zoom interviews, then they'll send someone to transcribe.
00:54:26
Speaker
will give them an insight and put a PPT. Now you have a single system of recorded intelligence within your system itself. So one of the things is brand always get from agency PPTs. They were like, how can intelligence sit at my end? Because every next year I do the same project which I did last year.
00:54:43
Speaker
So that's been our event so far. And how does it work? Like, is it that in the Zoom call itself, there is an agent, like an entropic agent, which is participating? Yeah. So there are two ways it works. One is upload.
00:55:02
Speaker
There are two ways it works. One is that you can upload videos on my platform decode and it analyzes everything. Second is you can authenticate your zoom account. We have plugged it already. So you authenticate and any conversation that happens through that agent just comes to our repository over here. So it's a pull basically.
00:55:24
Speaker
And the third way is that we also have our own video conferencing platform. So essentially you can run everything on my platform and it optically sort of streams in everything. In that flexibility to customer, because we realize that you are in some platforms, you are not gonna switch today. So kind of have that choice. Ours is built with a lot more focus group approach with moderation and things great into it. But inside the conference is all the same for all.
00:55:52
Speaker
So this is how this is how decode works and 21 in December we launched about 12% of our revenues actually today done by decode and as we go forward what it looks like is
00:56:11
Speaker
So what we are covering is the CMO and CPU bucket within the brands and the very interesting happening as we launch Decode. This brings back to your first question that you asked. How do you analyze sales conversation? This is a straight fit. And then massive amount of insights into leads, which is asking about, can I use this for sales and contacts into conversation? Because all sales conversations have gone remote.
00:56:36
Speaker
And I want two things out of it. Can you help me inside with training my sales guys? Or can you help with getting the purchase intent of the lead? These are two things can I do? So we are thinking of CMO, CPU, and product and now sales organization. We also get a lot of interest from hiring functions.
00:57:00
Speaker
Can I use this to analyze the video resumes or kind of interviews and all? We're not yet touching it, still a very open kind of topic. But I think sales looks a lot more immediate. The question of why we didn't go to sales first, we went to, we went to businesses where we have a tighter consent, right? If I, through barrel, I have a tighter consent. Sales conversation, we're still not accepted that
00:57:28
Speaker
hey, how come you go about analyzing my facial expressions? Yeah, in 2017, it was unimaginable. Yeah, but post pandemic, I see that moment being there. But while you still stole your contacts into call recordings, people are still fighting for whatever purpose training and purpose that we say. I think that consent is where it was, right? And then we are trying to time our things around it.
00:57:54
Speaker
So yeah, I think that's where we are. I think we have perfected tech to a fair bit. I think last three years has been taking it to various products, shipping up and sort of taking it out of the market. Today we operate in four geographies, US, Europe, South East Asia, India plus Middle East. These are the four pianos for us, 70 people, 80 people sales team.
00:58:18
Speaker
and about 120 people are engineering product functions, about 100 people are company. In the process of doing so, we have about 17 patents, we raise two rounds of capital and then kind of build a fairly
00:58:41
Speaker
a fairly amazing set of leadership team which is driving each of these geographies and growth and sales and success and product. I feel that we are a lot more ready for next week. I want to ask one or two questions on product and then I'll jump to the other area.
00:59:06
Speaker
This is a pure sense of product. What is the customer journey? Is it like a freemium sign up for 15 day trial and then you pay? Tell me about that.
00:59:19
Speaker
So it's not a freemium. It's a very super enterprise product. So it's a very demo leg for subscription. Wherever you do demo and at large enterprise, you might have a small POC to be done and then you sort of get it. But by usage, it is fairly straightforward. You log in and you can do everything together. So it's very simple. You go and create a campaign of your launch it, get all the insights together.
00:59:43
Speaker
create a campaign like say I want to test a piece of content, so then I'll upload that piece of content, I'll choose a demographic. Correct, you will choose the demographic, you want to set out some sort of a post survey also, you want to capture something, you want to capture some code and all that you can do. And you can launch the test and the result will start coming in together.
01:00:03
Speaker
As a result, Sareen, you can dashboard automatically and allows you to analyze data by user types and you can clear between add one versus add two versus add three or pass project or campaign. You also have a benchmark scores of ours. Like we have tested 100,000 odd videos in past. So you will get to within, within let's say personal care as a category, what does benchmark of effect lab score looks like? And then how far or good I am is that?
01:00:33
Speaker
So your sense of how good or bad this is. And then you get a time series graph of deep dive analyzing as in what people actually like this like then I track map of where people were looking, not looking. You want to aggregate by elements within the video you can do so, which means I want to know how many times people look at Shah Rukh Khan can know that. So you can circle the objects and those object files and insights is what you can have. And then this insights, all of that you can
01:01:03
Speaker
bundle and drag and drop, and you can get an insights report generated. You can collaborate with your other teams within the organization, within the platform itself to actually get all of this set up and share these insights.
01:01:19
Speaker
Fair amount of collaborative development baked into it. Subsequently, doing a decode kind of quantitative research. Both the research sits at one end. For you, it is a single source of truth across all customer views and insights. It's a self-serve tool, but I think brands are transitioning from, particularly Indian brands are transitioning from more agency-led models to this.
01:01:44
Speaker
we have a customer customer success team which sort of onboards them for three weeks a little bit of hand-holding is there so not like us fully self-serve but with a fair amount of customer success support they are doing it right you also said that you are looking to judge purchase intent for sales conversations what would be the way to do that like
01:02:10
Speaker
So primarily the same as what decode is today, right? You have qualitative interviews. Instead of that, you will have sales conversation. So essentially, you stream in all these video conversations that habit sales reps are having and lies on that and basically makes sense out of it in terms of what expressions are the winner expressions and what expressions are not the winner expressions.
01:02:32
Speaker
that gives you a sense of intent and second is if there are 100 sales guys that are using this product essentially what the best sales guy is doing on tier voice modulation and their way of representation and pitching that the other guys not doing.
01:02:49
Speaker
things like talk time, who's speaking how much, are you listening even or not, even to the extent of what was said with what confidence level, probably you're saying a lot of things but not with higher confidence, or things like you're not cheerful, right, people want to see cheerful people, or you were hesitant or sort of distressed at certain time in your conversations.
01:03:14
Speaker
Can you make that right? And what that was exactly leading to? How do you respond to objections that came from a customer call? So I think those are the things that are powerful values. Have you built that algorithm which can predict or is it working for us? I guess currently being trained.
01:03:35
Speaker
No, so there's no, there's a pure analytics, right? So I just want to build a system which provides you insights about where people were hesitant, where they were excited, where they were not, right? And then the next step will come as in work together as characteristics. It means to have a winning conversation. So I think right now we are still at an analytics level that are putting developers together.
01:03:57
Speaker
I want to understand how you scaled it up from the perspective of selling. Sure. You have a global presence now. How would you scale it to different countries? Is it through digital marketing or is it through a salesperson who reaches out and
01:04:21
Speaker
Yeah. So now what we did was we got a first set of team of SDRs in place to begin with. And that was serving across CEOs. And then we were part of about 10 to 12 accelerator programs from where we were getting early customer access. And we were able to demo and pitch to them. So Accenture had a bunch of customers. SAP had a bunch of customers. And that became our primary first point of connecting, initiating, and having conversation with customers.
01:04:49
Speaker
That was the first GTM motion we ever did to launch into newer markets. And as we got a few fully set of customers into each of these geographies, when we actually went about hiring a team into these geographies, and largely these teams have been guys who are selling software to brands in the past. So pretty much guys who have done SaaS and Bribrize and all that, and build a fair amount of team out with each of these guys. So by AEs and sales directors and VPs are all in respective regions,
01:05:19
Speaker
My SDR team is based out of India. And that's how we operate. What is the difference between SDR and E? SDR is the one who likes... SDR is a prospecting team whose goal is to build a pipeline. Yeah, they build a pipeline. The Sains qualified output value is their KPI. And the A team sort of takes it up from there and sort of tries to convert it. And A team has got a number quota.
01:05:47
Speaker
this is your quarterly target system. So, both of them have different goals. I think the skill sets are very different than I learned back from my sales days that prospecting is a very different skill. If you run wider than that person, person who is too close should never prospect, which means that
01:06:08
Speaker
Because if you are prospecting and you are only closing, then you are more interested in closing than prospecting. And you are very likely to ask for marriage in the first date. Nobody likes that.
01:06:23
Speaker
You see this tendency in sales guys that they will get into the first case and is like, hey, here's here's emotion. His facial pudding is the most amazing thing ever happened on the planet. And you would you like to marry me? And they're like, not yet. Not yet. Hold on. Go prospecting is nice and easy. Just go have a conversation. Don't try to close. Okay, prospecting the goal is to say, okay, would you like a demo?
01:06:49
Speaker
That's the goal of prospecting. Okay. You have that and you have the demo and just say that. So I think the difference between the services selling or the selling that we used to do in past was, hey, what is your need? I'm here to solve that versus now saying that, hey, I've built this, do you need this? Then if you are not, then thank you. It was great catching up and more quickly.
01:07:18
Speaker
So, finding the right guy is a lot more valuable than actually going about solving everything that a customer might have. So, the service to product mindset and hence the sales alignment rate is over. Okay. So, the AE's or the account executives are the ones who convert the deal. Then, do you have like a customer success team which does the hand-holding? Yeah.
01:07:40
Speaker
What's the deal is one then there's a customer success does about three weeks of onboarding get a customer line and stuff work along with them with usual cadences that is in place and yeah that's the usual cycle we also a professional services team of consumer insights professionals.
01:07:58
Speaker
who, let's say, out of the platform, if some of these customers say that, hey, I want to support the consulting support from your side of analyzing data much more by doing it in a certain way, which is very customizable. Then this is a team which is basically researchers from the agencies and the brands in the past who understand and know the language of brands, and they are the ones who sort of helps them out.
01:08:22
Speaker
This team is also a knowledge house of domain in our company. So all the sales guys and everyone gets trained by them in terms of what you should speak if you're in France. Okay, got it. And how do you price it? So it's priced on a standard SaaS pricing. It's a one-time onboarding fee. You have a platform subscription fee on top of it, a usage model, and then there's professional services.
01:08:50
Speaker
But what is the, on what matrix is the user's price? Is it per campaign respondent? Per campaign. Per campaign is defined as 150 respondents. That is usually what people use for that. Basically, one stimulus is what a campaign needs. One ID want to test over 150 people. What does it cost like? And what is the ballpark range for a campaign?
01:09:18
Speaker
Yeah, a campaign of 150 respondents is about $750 for you. And you have a one-time platform license fee of about $15,000, which is annual in nature. So the more you use, the better your ROI economics looks like. Okay. I mean, how hard is selling it? That $15,000 one-time cost seems like a hard sell.
01:09:48
Speaker
No, I think it depends upon what you're selling against. And hence we have arrived at pricing. If you look at what a brand spreads today. Let me give you an example. If a brand in India is doing a typical 150-respondent kind of such a project, minus technology and all, just the usual quantitative and qualitative survey, they're spending about 7 to 8 lakhs.
01:10:12
Speaker
are doing this across both quantitative and qualitative interviews. You're spending about seven to eight lakhs testing out one advertisement or one of these things. Let's say a set of five into that maybe. One project, essentially one objective. And that's easily about $12,000, right? And with respondent and all taken together. If you are doing 10 of such, then essentially you're spending $120,000 in a year,
01:10:42
Speaker
What I'm asking is pay $15,000 as this license and then use everything else. My project is 750 a project, which is not even 50,000 rupees.
01:10:55
Speaker
It's close to about 60,000 rupees. So if you are doing now about 10 of such campaigns, then you have a $7,500 on campaigns and you have $15,000 otherwise. That's a $22,000 and $22,500 is about 20 lakhs for you.
01:11:16
Speaker
spend that you are otherwise doing on dead campus. So, like-to-like, where it becomes effectively a 30% kind of a price, 70% kind of saving overall. With a lot more tech, a lot more insights, a lot more everything. I think we never see price being a constraint. It does not look a hard sell because as a scenario is way too costly.
01:11:37
Speaker
So as a founder, do you see yourself as the product guy or the sales guys? What is your role? My role is I have been always a sales guy, but I think these days, I think I'm trying to be a good listener and better judge of people and culture. That's what I've tried to focus on, but always been a sales guy. But I feel that I connect to sales a lot more than anything.
01:12:04
Speaker
And what were some of your top three learnings in this journey of building and trucking? Things which you realized that you needed to change in yourself.
01:12:15
Speaker
Yeah, I think first was letting it go. You can't do everything yourself and hence how so passionate you are about something. You have to have people who do it better than you. So that was first and I think something which is very close to you, letting it go at times becomes hard and that does not scale as well.
01:12:38
Speaker
I think what makes you great in early days does not make you efficient down the road. So I think that was important. I think second was the lighter you are on your feet, the better you are likely to do. So your ability to shift gears in the day
01:12:57
Speaker
is super important on a very double end day. Can you stay calm and then quickly be very, again, back to double ends? Can you handle that? So that was just my personal growth element to be able to maneuver through that, is managing your emotion as a superpower. If you can do that, that's amazing. I don't think I'm great at it yet. No, I haven't consumed your time, so I can start up with consuming your time.

Reflections on Timing and Intuition as a Founder

01:13:28
Speaker
Third thing which I learned is that I think it is not a new learning, but it always stay true that time is a crazy variable. Be very cognizant of timing. The magnitude of disaster is nothing in comparison to timing of disaster. So a small thing, back timing can screw you up so bad as compared to a huge thing on an OK time.
01:13:56
Speaker
So stay very upright and cognizant to timing. That's just something I feel as founder. I got to be doing that very often. Amazing. Like being intuitively aware of whether this is good timing, bad timing of decisions of things happening and so on. That's something which you developed as a founder.
01:14:22
Speaker
Yeah, I feel that most of the decisions that I see I've made or I try to make, you are doing that with very limited information. And I think you can never make a correct decision in its totality. So the only thing you can really have a good gauge on is it the right time to make this decision or not? And is it a delayed time to make the decision or not? If you are 100% on that and if you are even 30% on the
01:14:51
Speaker
quality of decision, you will do good. So time is, most of the time, missed out the mission of decision making. I think, yeah. Fascinating. That's amazing. And that brings us to the end of this conversation. I want to ask you for a favor now. Did you like listening to the show? I'd love to hear your feedback about it. Do you have your own startup ideas? I'd love to hear them. Do you have questions for any of the guests that you heard about in the show? I'd love to get your questions and pass them on to the guests.
01:15:21
Speaker
Write to me at adatthapodium.in. That's adatthapodium.in.