Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
The Martech Masterclass | Srikrishna Swaminathan @ Factors.AI
318 Plays2 years ago
Martech refers to startups disrupting traditional marketing tools and channels with technology. Srikrishna has had a front-view seat in the Martech revolution. Prior to starting up, he was part of the global leadership at India’s first unicorn, InMobi. With Factors.AI, he plans to help B2B marketers make easier, wiser data-driven decisions.
Know about:-
- Experience at InMobi
- Product Features
- The product’s wow factor
- Privacy concerns and regulatory trends
Recommended
Transcript
Introduction to the Founder Thesis Podcast
00:00:00
Speaker
Hi everyone, this is Sri Trisha here. I'm one of the co-founders and CEO at factors.ie. Very excited to be here. Take me on a tour, I'll take you on a tour.
Theme of the Podcast: Learning from Founders
00:00:11
Speaker
This could be a great intro. Hi, I'm Akshay. Hi, this is Aurob. And you are listening to the Founder Thesis Podcast. We meet some of the most celebrated sort of founders in the country. And we want to learn how to build a unicorn
The Disruption of MarTech
00:00:34
Speaker
Just like everything else, the field of marketing has been massively disrupted by technology. In fact, the term MarTech refers to startups that are disrupting traditional marketing tools and channels with technology. And this is a massive opportunity.
Sri Krishna's Background and Mission
00:00:50
Speaker
Two of the biggest companies in the world, Google and Meta, are essentially conduits for marketers to reach consumers.
00:00:57
Speaker
Shri Krishna Swaminath, the founder of martechstartupfactors.ai, has had a front-view seat in the martech revolution. He was an early member and a key leader at India's first unicorn in Moby, where he was responsible for setting up a division generating more than $100 million in annual revenue.
00:01:17
Speaker
at factors.ai, he is on a mission to make it easier for marketers to consume data and make smarter decisions about how to spend their marketing dollars.
Building Vodobo at InMobi
00:01:26
Speaker
Listen on to this masterclass on all things smart tech.
00:01:30
Speaker
So I chose to more to get other startups at that point in time. And Inmobi was one of the companies just growing really well out of Bangalore at that point. So in Bangalore, there were two big startups then, Inmobi and Flipkart, both were just becoming a unicorn that pointed time in early 2013. So I chose to join Inmobi and then spent close to seven years at Inmobi. That was a fantastic start at the same. I started as a sales store, but eventually went into more general management, building a business unit internally.
00:01:57
Speaker
What was the business unit that you built? Internally. Internally, it was a business unit called Vodobo, the whole affiliate business of Inmobi. Later, I also handled the device monetization and the agency business at Inmobi. I started that up also. The affiliate business grew to more than $100 million plus PM. What is an affiliate business?
00:02:18
Speaker
Yeah, that's interesting. So, in any...
Understanding Affiliate Business Models
00:02:21
Speaker
In movie, it's fundamentally like a marketplace. So, any marketplace, there are two-sided marketplaces. So, there is demand, which is called advertisers, and then the other side is publishers, which is where you place ads and other side. Actually, it is basically creating another third tier in between, because you can't get all the supply yourself from the publishers, where you can put the SDK, but you can work with other
00:02:41
Speaker
partners or athletes who would actually help you get the major supply, whether it's in South America, whether it's in North America, whether it's Japan Korea or China or so many other things. And this is a more easier way to expand your supply possibility. The margins would be lower, but the turnaround would also be much higher and you'd also be able to scale up the business very, very fast. And this thing, that's something. What do you mean, turnaround will be higher? What does that mean? The thing is like the scale of the revenue which you can get would also be very high.
00:03:10
Speaker
from the initial times itself. And more importantly, you would also get multiple deals. That is, let's say, as a simple level, it's like you have an advertiser, which is mind-thrower, and you want to go and tell them, I'll find inventory for you, because our SDK is there, and let's say, Crick & 4 or NDTV or wherever it is. That's one level business. It's scalable only to a certain level. It depends on how fast you are able to get the publishers like NDTV or Crick & 4.
00:03:32
Speaker
On the other hand, if you work with third party publishers, you'll get access to thousands of other publishers who they have worked hard to get it while you are able to bring in the advertiser from your site. And other things, so suddenly, instead of miners spending $1,000 with you, they'll start spending $10,000 with you because of that. So the margins could be the same.
00:03:47
Speaker
Okay. And the ads which go through publishers, ads which go through the other networks, the aggregators in a way, those ads would give you less margin because you will share that margin with that aggregator. You would have to change the margin, but at scale, once you expand the scale and you are also able to get more variety in terms of advertisements and in terms of publisher grouping and other things. So you also will be able to increase the pricing. You'll be able to optimize the pricing data.
00:04:15
Speaker
And you would also be able to expand the quantum of money flowing through you.
Monetization and Optimization Strategies
00:04:20
Speaker
That's the goal of any marketplace. At the fundamental level, you focus on increasing the quantum while pricing margins and other. This thing is more of a secondary choice which you have to trigger on. So for example, the first two quarters, it wasn't profitable. It was actually net gross margin zero. And this thing, then we started tweaking the pricing and eventually we started taking it up pretty well. At peak, it was more closer to 40% plus in margins.
00:04:44
Speaker
And it was $100 million plus business. It was one of the most profitable units with an in-movie as well. I thought pricing was through a bidding system, right? How did you fix the pricing? Because I have an outsider's knowledge of ad networks. So therefore, I'm asking you to explain it to me as a layman.
00:05:03
Speaker
So it is a bidding system, but it's also the bid and last sprint also, you can keep changing by publisher and by bot saying that, okay, this is the price which I'm going to pay whoever wants to pick at this price, you can pick what the advertiser pays to me is more overall return on performance metrics. So that's what it's called a performance update network.
00:05:22
Speaker
where advertisers pay you on final performance in terms of number of orders you get or number of installs you are able to generate another thing while the publishers would be basically bidding on trying to optimize on this is a set of impressions clicks we can build. We will also be able to get to so much conversions, we will also hold them accountable to conversions but we basically go back and put it up in the platform. There is an affiliate platform first we are using a third party platform then we build it internally also saying that this is the price which we paid you won't pay anything more than that. So that's why between the bid and us
00:05:51
Speaker
between advertisers we can increase the margins over a period of time and also it's also scalability for example any publisher would once you start increasing from 10 advertisers let's say you have 100 campaigns working with you once you have 100 campaigns almost everything the publisher wants to work with you because you have variety of campaigns I can go to some other
00:06:10
Speaker
advertiser or someone or even directly and get that particular campaign from them or something of this or but I won't get the hundred others. I'd rather work with them. Second thing is payment cycles and other the same. Most of the payment cycles, if it's managed by a very large company, one, they trust the money will come in the right time. Second, they have a standard payment metric. If they have to independently either work directly with the advertiser or any other affiliate, which is smaller, there's always the risk of payment default or payment delays.
00:06:36
Speaker
So payment delays is going to cripple anyone. That's why you understand the concept of how working capital can cripple anyone from actually scaling up the business and that becomes tougher and tougher for any other person to do. That's why the moat in the marketplace also, once you expand the business and you're able to guarantee a certain level of standard working capital practices.
00:06:54
Speaker
and also very clear clarity on like you get this confirmation of payment as long as we confirm that this is very clean inventory and the order has been performed as per whatever the requirement, you'll get the payment done also very clearly. So it works very finely. Okay, so just to recap. So a publisher is not going to get paid based on the outcomes, but he's going to get paid based on in a way the input, which is the number of views
Ad Network Performance Management
00:07:20
Speaker
It's similar to like you employ both. We actually can work on both sides because like you were sales team. So you're paying the sales team on number of mandates work, but you are earning based on number of items sold.
00:07:35
Speaker
So it is something similar with a publisher where they are getting paid on views and you get paid on the throughput, like how many views led to a purchase or an app install. And therefore, if you are able to run some algorithms to optimize which campaigns should run on what kind of publisher to give you the best throughput, it increases your margin.
00:07:57
Speaker
to overall increase your margin plus you can also then variant like which kind of campaigns for which kind of publisher sub IDs you are the tracking mechanisms is pretty clear you can track from impression clicks to number of insurance to downstream metrics and all those things and you can keep playing around like which downstream metrics work you will also then see what is the quantity if there is any kind of clicks timing and all those things so there was there was a lot of investment around both the optimization algorithms and the fraud detection of the algorithms also because
00:08:24
Speaker
There is a very easy way to generate installs through backend API calls and other things. But then you look at bot metrics and clicks spamming metrics and all those things. So the standard things. And this is at more than billion clicks a day. And why do these supply aggregators exist? Wouldn't everyone be by default going to say a Google? Everyone knows that if I have inventory, I can sell through Google. So why would a supply aggregator exist?
00:08:47
Speaker
Yeah, of course. So Google's budget should be more like, so people would work with Google. People do work with Google. So actually in the ad tech as an industry, there would be a pareto rule where like there'll be people go to Google, Facebook also for their exchanges, which can give the low tail and long tail and they can also get 80% of that. The remaining 20% then they go for the other set of person who's one for better pricing, second also for more variety.
00:09:11
Speaker
And also the thing is today's ad requests, it's like time I can't sell it tomorrow. So let's say I have 100 ad requests. If Google is able to buy only 80, then I'm going to send the remaining 20 to someone else. If I don't send, it's going to basically go waste. So I have to have multiple connections and there is a mediation order around that.
00:09:29
Speaker
Otherwise, it's not going to set, so that's why the market sells. Okay, okay. And these supply aggregators, give me some examples. Who are they? The affiliates that... So, across the board. So, I used to play both sides of the market, and in movie also, the shrink particle would also be a supply, can also be a demand, because what's your marketplace, you can twist the...
00:09:46
Speaker
It's like a forex marketplace. But let's say a supplier aggregator would be like a kind of a web publisher in the US or South Africa, which aggregates all their sports websites or news websites or any other kind of particular sort of gaming platforms and other things. Or there are platforms like Mobvista, there are platforms like few others. It can be even a OEM provider like Xiaomi or the thing which you use to sell their VM slots such as that. There is a few other companies which is coming into place like Apple and other things. So there are multiple kinds of that. And each of them would have
00:10:16
Speaker
maybe the same set of supplies or different set of supplies coming in from different parts and other things. And you keep scaling that over a period of time, depending on how it works. How does the pipe work between the customer who sees an ad and the advertiser, say, a mintra ad, which I'm seeing on my Xiaomi phone? How does that pipe work? What does the pipe look like? Xiaomi will be connected to an aggregator. That aggregator will be connected to Inmobi or
00:10:41
Speaker
No shame, we can get no aggregator or directly to an advertiser itself. So the way things get tracked is there is a tracking URL. So tracking URL is as soon as you click both, whether it's an impression view through or when you do a click through, the server side click gets fired or the direct client side click gets fired. And it goes through the each and every hop sync. This happened, then this happened, then happened, then the post back happens to the eventual server and then it gets managed. So these tracking links are basically something. What is post back?
00:11:09
Speaker
Postpack is like when a conversion happened or when a goal event happened. So it can be a click conversion, it can be any of installed, any of those things. And this gets tracked. Like what happened after the person clicked on an ad? Did he actually install? Did he just browse away from there? So that information goes back.
00:11:27
Speaker
everything, the timestamp, which device they got installed from, whether it's from a Samsung phone, whether it's kind of whether he was walking or not, what are the kind of details. All this is captured by a tracking URL which is supplied by mobile app trackers, which is like an adjust.io, which is there, or a singular is there, or an apps flyer is there.
Sri Krishna's Leadership at InMobi
00:11:44
Speaker
Some of these are then self-billion dollar plus companies. So these are the companies that show all these things.
00:11:48
Speaker
So this mobile app tracker is, it's like a add-on player in this transaction. Like this is something which an advertiser is spending for this. Advertiser spends for it. Okay. Advertiser spends for it. And this is essentially able to give back information to the advertiser on the demographics of the people who saw or clicked. Never have a game company installed including everything, whatever the details which they can, they're able to get, they'll get and send it across.
00:12:17
Speaker
And how does this mobile app tracker become able to look at Xiaomi's consumer base? So Xiaomi has to work with them? So it won't be demographic data, it won't be demographic data, it'll be device data. So it'll be device ID, it'll be device maker, device whatever. Okay, so Xiaomi sends this data back in a standard format and that is just passed by the mobile app tracking company.
00:12:44
Speaker
Same thing, not just showing me. Even we know whether it's an iPhone 6 or iPhone 10 or whatever it is, we know so many other things. Who used to add in all those things. All this comes from the world. So basically, when a publisher wants to monetize their eyeballs, then they have to install some SDK. That SDK would include stuff which is sending back this information like the phone, the
00:13:05
Speaker
So that SDK would send the ad request. When the ad response comes, it will have all the things. SDK is more from, let's say, they would have an in-movie SDK or a Google AdMob SDK or someone else. So then they'll send back all the details.
00:13:18
Speaker
And this SDK, what does this do? It resides within the app software, and it has allocated a space to present that, like, basically. So every single publisher, whether you use an NDTV or the thing, if you have an Android phone, there are Android markers. You can actually see what on SDKs are there inside your phone.
00:13:37
Speaker
Apps. Okay. Inside the apps. Okay. So you built up this affiliate business, which meant that you did the partnerships with these like Xiaomi or Verizon. These partnerships were what you were building up. And then you were ensuring that the pipes are built to transmit the ad and get the tracking data back. So multiple things. One is internally convincing the team saying that one, we have to sell this inventory. So get advertisers on board. Second is getting publishers on board, then getting the whole
00:14:06
Speaker
piping and system and technology, whatever the parts hold. Third is like making sure the finance, working capital, everything of that gets on hold. Then eventually building the team to manage customer support and other things so that the business runs. So at the end of it, it was more than a 60, 70 people team.
00:14:22
Speaker
It was close to $100 million plus business as I was mentioning, making more than $40 million in profit and also working across at peak, it was like at least 600, 700 campaigns on a daily basis. And it was pretty large pipe. So as a comparison,
00:14:37
Speaker
Couple of other companies which used to do the same set of business and which is also listed and maybe the same size of it correctly and other the thing is AFL, which is almost a $2 billion company in the Indian market for Iron Source used to do very similar business. It's also listed now Iron Source has even pivoted into using the money. They also started building a gaming platform. They also have a gaming studio and others. Now it's close to $8 billion company in NASDAQ and other things. So that's how the business emerged.
00:15:03
Speaker
This one-to-go business is essentially a business of serving supply to advertisers. You have to go aggregate a lot of sources so that your advertisers are able to advertise beyond the publishers who are directly linked to it. Every ad tech is serving supply only. So whether it's even whether it's you work with Facebook or Google or whatever it is, fundamentally you have supply. Either it's your own supply or it's someone else's supply. Its ability for you to serve impressions, views, eyeballs to advertisers.
Founding Factors.ai
00:15:30
Speaker
So that's all. That's our ad tech is fundamental.
00:15:32
Speaker
So, then what next after you built, what was this? So, I built that. I was part of the exec team at Inmovie. I was reporting to the CEO there, Naveen, and other things. So, it was pretty good. But then, I met a college mate of mine, Aravind. Aravind and I went back, go way long back, close to almost 20 years. We were in the St. Commage together, RV college. Hosted, Aravind had joined Google, IAC, and then Google. And she was there in Google for five years. And then he started up another company called Chatimity, and that got acquired by Freshworks.
00:16:02
Speaker
And that product eventually became one of the first versions of fresh chat. And after a couple of years at fresh shops, he wanted to come out and start up again. He also wanted to be back in Bangalore, so he came back from Chennai to Bangalore. And there we met and he had another idea of analytics. How to make analytics simple and easy and why given the number of data silos is increasing like
00:16:22
Speaker
exponential rate in the current world rather than it wasn't basically a very simple thing like Google Ads and then you wrote website form full and CRM there is the amount of data silos have actually increased and the need for analytics and how we actually look at data metrics and after the thing also needs to be more AI driven and done differently is the top process of the idea and it also had genesis on what Freshworks was struggling on and at that point in time so he came back with him to me with that idea
00:16:48
Speaker
And we started talking and I also had another colleague of mine at Inmovie, Praveen Das, who was part of the product team at Inmovie. He used to run the whole data product set in movie, build the customer data platform at Inmovie and audience's product set in movie and the same. So both of us were also, I and Praveen were also looking to start up. Praveen also came and said, let's start up. So I thought let's, three is always better company than two. And we wanted to get together and build.
00:17:12
Speaker
And that's when we formulated the idea and there was also the, we also saw the need on how Freshworks came in with. So the idea was very, very simple. So sales and marketing and even any website data silos have been increasing. And also the amount of quantum of data is also increasing very heavily. So for example, for any ad to flow through an app, that is Google, that is Facebook, that is, let's say you would have any other kind of LinkedIn or Jito code and other ads, this is, I'm talking about B2B.
00:17:37
Speaker
You're talking about online sales here. Online sales, yeah. Any set of digital marketing, the number of channels, and there are also offline events. If you do B2B, there will be webinars, there will be some kind of events and all that. But the data is recorded in terms of these emails and all those things coming. But the number of channels is increasing. Then you have the website.
00:17:55
Speaker
So like for example, like a company say a D2C, say like a mama are selling cosmetics and a wellness product.
Challenges in Data Management
00:18:04
Speaker
So one is they would have data coming in from their own website of the consumers who log in and what stage they reach, how many checked out, how many did not check out. Then they would be running these email campaigns sending out
00:18:16
Speaker
new product releases, they would be getting some data from that. Then they would be running these ads on Facebook and Google and in Moby and other such platforms and they'd be getting some data on what kind of customer is clicking on their ads. Probably the marketplaces themselves would also be giving them some data like Amazon and Flipkart.
00:18:31
Speaker
They would also do that. They would also do that. Marketplaces don't share much data on who the user is and all those things. They just give reporting data of cost and revenue. But most of the other data, which is they want to do it on their own website because they don't want to prefer to be part of Marketplaces. There'll be website data, Google, Facebook data, and there'll also be a CRM data, which is better. I don't think they'll be using Shopify, Mama Earth is very big.
00:18:55
Speaker
They generally use Shopify or one of their own CRM data that's also there. All this data is in different places. You spend and cost in one level. You track the user journey all through the website. Amount of time spent on the website. There is also a lot of content marketing. Amount of time people look at a particular content log in the website or whatever it is. And then there is also, once they sign up, amount of time they take to sign up, which products they eat, how much they bunch with the products at some time, how many products do they add and other other things. And eventually they become part of the CRM also when they close the deal and other the thing and they pay the money.
00:19:25
Speaker
So you have data from cost, which is channel metric to revenue. You need to integrate all this data and stitch the user journey into it. Currently, that kind of a user journey into and switching wasn't there at that point in time. There were multiple point solutions which were trying to do part of it. Or if you say the same thing for Freshworks. Freshworks is a huge SaaS company. They were doing more than 100 million revenue and now it's more closer to 400 million.
00:19:47
Speaker
They were spending close to 25-30% of their revenue on marketing, digital marketing alone, because digital marketing matters. They'll be spending on channels, they'll be spending on content, they'll be spending on a lot of webinars and other activities. Then people will come sign up. Eventually, there will be a sales representative assigned and they'll also be marketing email newsletters and all those things going. All this data is spread in multiple places. And that problem was, how do I get predictive pipeline?
00:20:13
Speaker
To get predictable pipelines, fundamentally, you need to invest in marketing. Marketing is not a kind of a cost that's basically invested. Because you put in money in marketing, you get output and revenue. Very, very simple. But how do I know that if I put $1 million in marketing, I'll get so much pipeline, I'll get so much revenue in a predictable, clean way. To do that, you have to bring together all the data, stitch the data into it, and then see it saying that if I spend in this, by this cut, by this geography, by this product, by this thing, this is the kind of BCD and this is the set of customers I get. So use them.
00:20:43
Speaker
ACB, what is ACB? Average customer value. So, for example, let's say Freshworks, FreshChat, if you see, it's like per customer per seat, they sell $20 per average seats is $5. It's $250 per month or $3,000 a year. For something like Gong.io or something, it'll be $50,000 a year. But what is the
00:21:02
Speaker
customer value per segment and what how you need to convert or even for a mother that something like average basket value will be something like a thousand rupees or thousand five rupees average mother in Delhi is going to buy so much or something like that they would have a value but you need to interact so much I invest in marketing so much I get so that you'll get the ROAS but you'll also know exactly by with slice and cut you'll get
00:21:23
Speaker
Currently, that kind of data for even something like Threshworks, when they were trying to get it was impossible since they put a data engineering team. And they also had a lot of funnels by pulling together data using ETL routes, dumping all the data on data warehouse, then building a Power BI or a business intelligence platform on top of it to see how the data worked and other the sync, or which slice and dice of the data were. And this brings it to a couple of problems. One is you need to invest in internal data engineering to build ours.
00:21:50
Speaker
Second is to switch together data and write specific queries and size and size of the data, you have to have an SQL query right at each and every time. And then you have to visualize it. As the data grows, the visualization also becomes very, very clunky, tab-do or a follow VI or wherever it is. So this is three-set problems. Give me an example of this traditional way which you're talking about, which is a heavy investment approach for a company. What does it look like? They put this data into a very large table, just break it down for a weird person.
00:22:19
Speaker
Data warehouse. They'll put it into a data warehouse. So what they'll use is they'll use ETL tools like a five-tran or a heaver. What is ETL? ETL is basically a piping tool. That's like if a data is there in Google channel and data is also there in my CRM, data is also there in my website, how do I take all the data and pipe it into my data warehouse? So how do you push data from your other sources into your data warehouse is fundamentally what an ETL tool does. So then all those ones
00:22:44
Speaker
Yeah, it is like it would do an API connect between the... It will do an API dynamic, transform and load the data. That's what is that. That's what... Like a very simple consumer version of this would be like a ZAPIR. That's correct. These are... ZAPIR does it without full context, but this is a little more advanced forms of ZAPIR also there. There are also stores that are the same which are sold. So there's a reverse ETL also which will move from data variables to a salesforce also, those kind of things are also there.
00:23:12
Speaker
But at a fundamental level, you have to use piping, or you can build this pipeline. You also can build this pipeline. And other than that, ABA is exposed by either of these connections, and then you build it. Once you put the data into data warehouse, and data warehouse is... What is a data warehouse tool? Give me an example of a name.
00:23:29
Speaker
Snowflake, BigQuery. Snowflake is there. BigQuery is there. Redshift is there. Amazon Redshift and all those things. It's dependent on the kind of Azure data warehouse and others are also there. But the key thing here is data warehouse is like a store room. You can dump all the data. It's something like Srikrishna clicked on an ad, you can dump it. Srikrishna was on the website, you can dump it. Srikrishna, full form, you can dump it.
00:23:52
Speaker
It won't have context on like all these records of Srikrishna is the same thing and you have to stitch that. Then you have to, how do you stitch that? Either you write queries just like, okay, I want to look at all these people and I want to arrange the data and other things. Or when you write a general queries, it will find this slice, this slice, this thing. That's why it becomes very clunky for you to manage. Or you have standard queries you have to return and then you have to upload the data. If you change the date or if you change the thing, the query runs and takes so much time for the data to come back.
Solutions by Factors.ai
00:24:21
Speaker
So there are two parts to it. One is you need to write SQL queries on top of the data once it goes into the data warehouse. Second is you have to bring in all the data in the right way and then you have to switch the data and other things and dump it into the data warehouse. So this is what a lot of companies would have to do if they have to do the end-to-end view of all the data from channel all the way up to the site.
00:24:41
Speaker
With our product, we bring together data, we stitch the data. On top of it, we also enable analytics on the data. So what you do is sometimes you put a data warehouse, how you put this thing, then you sit on data with a mix panel or heap analytics, just as analytics. What are these mix panel heap? You'll have to really break it up for layman.
00:25:01
Speaker
As I said, one is bringing data, so there are ideal tools. There's a data warehouse, which is Snowflake or Redshift. Then on top of it, you have to do analytics. That is, you have a dump of all the data, but then you have to do analytics. Analytics, either you can write your own query, or you can use a clone, which makes it very easy to write a query, something like, pull this data, pull this data out, something like filter and all the other things. Fire a query, you get it.
00:25:24
Speaker
So these are essentially like making the process of data queries into no code, what say a bubble is doing like no code development. So that data querying is easy for you to query the data unless you have the data underlying data in a clean way.
00:25:41
Speaker
But that stitching together, you still need to do. For example, that these three different records all belong to one street Krishna, that you would still need to. So you have a customer data platform, like a segment and others, which will help you route the data in another way.
00:25:55
Speaker
So you pipe it into a customer data platform, and then you do it. Or when you put it down into a data warehouse, the index will sit on top of it. But there, you won't be able to stitch it. But using analytics queries, you can slice and dice and the individual user level and put unique identifier codes and all of the things. And then you can build a report and analytics query, and then you have to visualize the data.
00:26:15
Speaker
Okay. So this is like an immensely high, like bandwidth hogging task to really be able to. So there are two, three problems in it. So none of this, what I'm talking about is like impossible. And this is, you have to one push in at the end of the day, it's like dying and people.
00:26:32
Speaker
Do you want your best of your data engineers and your data analysts and other the same working on your own product and other the same or do you want them to one working on getting together data and then seeing the data. Second is you want your marketers or revenue teams focusing on
00:26:48
Speaker
strategy that is like what would be the right way to position and then spend on rather than spending a lot of time just to making sense of the data itself in the first place because once you if you automate the hard work of the grunt work I would say like pulling together all the data and visualizing and seeing the data and analyze the data very very easy then the machines can also be more iterative and faster.
00:27:11
Speaker
If it is not there, getting through the data itself would be the dashboard is not analytics. Dashboard is not insights. Dashboard is like whatever garbage you're going to be missing, but you have to prepare, putting and then you have to see the dashboard. Dashboard is also limiting in terms of like how much you can see the people and other the things.
00:27:27
Speaker
But it reduces work with a proper end-to-end tool where it combines all the data, switches the data with context and then gives you analytics on top of it and also visualization together. So essentially, you've bundled together a lot of discrete processes, services into one one-stop shop solution for a marketer with intuitive UI that they don't need to know coding to use it.
00:27:50
Speaker
They don't need to know coding. They don't need to depend on data engineering teams. They don't need to work with the kind of sprints separately to make things happen and other the same. They can just directly start working on it, integrating all their data so within 15 minutes, it's just a OAuth-based authorization or an ABI key.
00:28:06
Speaker
Immediately, as soon as they start with it, the data comes through and they can start. Standard templates actually showcase how the data works and what are the data points which are there. And then if they want to build any additional data dashboards also, they can build it pretty easily using the navigator. And it's within a week or two, you can immediately start getting immediate value out of
Onboarding and Integration with Factors.ai
00:28:26
Speaker
the product.
00:28:26
Speaker
tell me about the onboarding journey of a customer. So this would essentially, again, let's stick with the Bama Earth example. Let's say the Bama Earth's marketing team was to use this. So what would be the onboarding journey for them?
00:28:38
Speaker
Simple. So one is like how we actually have ourselves kept our hold onboarding is both the product-led or a demo motion, depending on what the user is. You can just immediately get started where all you need to go is into our website. First name, last name, email ID. Immediately you can get assigned them into the project. The project will automatically get created. Then
00:28:58
Speaker
For them to, they will be taken through a settings page, which is like for them to integrate all the details. What we have kept is totally OAuth-based no code, which is like putting the JavaScript SDK on the web pages, like a simple GTM tag, which they need to put in the GTM header. That's one thing which they have to.
00:29:13
Speaker
What is GTM? Google Tag Manager. They have to put it on the header, which is there of the Google Tag Manager, which any started web admin will understand. Second, apart from that, they have... All websites come with a Google Analytics it built for you. Google Tag Manager, not a Google Analytics. We do our own analytics capture. It's a tagging. Tag Manager is like, how's your website structure? Analytics is Google Analytics's own analytics SDK, which is there. We replace Google Analytics completely for any of the companies that you work with.
00:29:41
Speaker
Okay. So Google tag manager is like a standard feature. Every website must we have or every publisher. For UTM tags, unified tag manager. And what does this do? It helps in ranking. No, no, no. It doesn't help in ranking. It basically tags on a website, tags all your data or whatever the structure of the website very, very clearly. Where is your JavaScript codes? And you can place code within that, or you can also even say that this is the place where the form fill is. This is the place where the kind of click buttons are. These are the places where let's say blogs, K-students, everything.
00:30:11
Speaker
So it's something which allows the browser to orchestrate your code into the UX. It's a blueprint of the website in the back end. So I'll put it back.
00:30:23
Speaker
in a very seemingly easy way. So you just place it there. Once you do that, all the channel data, whether it's Google, Facebook, and all those things, there is a simple OAuth-based authorization. You can just click on that and automatically the data would start coming in. Because we have built the pipes underlying for, let's say, Facebook, Google integration. All you need to do is authorize your Google account to send data. Like, Mamas would have a Google account where they would be spending money for ads. Similarly, they would have a Facebook ad. So any platform with
00:30:52
Speaker
local account will have the cost data, will have the campaign data, will have the keyword data, would have each and every data would work. Similarly, Facebook would have that. LinkedIn would also have that. If we are able to bring in all the data from our side, it's just like a simple OAuth-based authorization and that will come in. Finally is the CRM. OAuth is essentially when you just sign in with your username password and click on the allow button basically when it's correct. That's all. As simple as that. So we just allow the data to come in, we'll be able to get in the trick.
00:31:20
Speaker
And what about data from, say, email campaigns, if you're also using MailChimp? If you have MailChimp or any specific marketing automation tool, like a HubSpot or a Pardot or a Marketo, you also get data from them. How does that happen? Again, there's like a OAuth. So it's like a HubSpot has an API key, so you just have to place an API key and get it. Marketo would also have a little more than an API key and a form sign-in. Salesforce has a login kind of thing, so we are able to get that.
00:31:48
Speaker
If you like to hear stories of founders, then we have tons of great stories from entrepreneurs who have built billion-dollar businesses. Just search for the founder thesis podcast on any audio streaming app like Spotify, Ghana, Apple Podcasts, and subscribe to the show.
00:32:08
Speaker
So you've made the process for a customer essentially pretty no code. He just has to do that one time, sign in an authorization and on the GTM he has to place your code, which allows you to replace Google Analytics and then start looking at all the website analytics. What kind of analytics do you look at on a website like who are
00:32:27
Speaker
So select time spent on the page. Let's say what is the standard website journey that something which comes to before every single market. I have people coming in from a Google campaign or organic campaigns or I write blogs and I distribute the blogs through multiple channels and other things. How do people come in? Where do they come in from? So that's the first question. Once they come in, what do they do?
00:32:47
Speaker
Or does this behavior differ from people coming in from within India? For example, Delhi versus Bombay versus Bangalore, for example. Does it differ from people coming in from, let's say, a Windows desktop versus an Apple desktop? Something. Does it differ for people using Chrome versus Safari? Does it differ for people who are, when they actually log in through a Facebook account versus they coming through a Google account or inbound sign-on? Then there will also be other form of data which they do.
00:33:17
Speaker
Do people who actually first time when they come in and when they sign up in this thing, do they buy a fine Ruby item or do they buy a 3,000 Ruby item or rather the same? Or does it, what kind of behavior does it change? Do people actually spend more time on the blogs which I written? Let's say for Mama, I write what is a happy motherhood or any of the other details and other things. Does it have any impact in the path of use journey? How much time does it take to actually select a product and go to Add to Cart?
00:33:42
Speaker
and after that to cut how much time does it take for him to let's say do the checkout and other this thing is there any other thing which actually does this time difference change by terms of kind of customer any slice and dice of the customer it does or people coming in from within Delhi versus Bangalore does it change because of let's say IP address or any of the other this thing does it change by tapping payment mode or payment mode if you're paying a UPI versus credit
Advanced Analytics in Factors.ai
00:34:09
Speaker
card
00:34:09
Speaker
your VI versus credit card or all those things or cash on delivery versus normal the same. What are the things which are changing so that you also know how do you customize your website and customize your own kind of path flow within the website. The point is like you can also do the same thing with something like Heatmax and other things. Heatmax is recording of all the web sessions but let's say if there are for a mama kind of company it'll have 1 million visitors on a monthly basis. You can't look at 1 million videos right.
00:34:35
Speaker
What is a heat map? This is, again, like a SaaS product. You can just buy this and plug it into you. Heat map is something like you can buy the SaaS product. You will get, basically, you'll have recorded sessions of the user journeys put in the website. So let's say I copy.
00:34:51
Speaker
Each and every user. Each and every user, depending on the plan, if you have each and every user, then it's going to be pretty hard because you can't record every one, but it'll show what's apart. But the thing is, you can't look at every single video, right? It's impossible. But on the other hand, when you do analytics, when you capture all the data with your JavaScript testing, whatever it does, then it's easier for you to one, write queries on the data, like how many people looked at my blog and then went to checkout.
00:35:14
Speaker
how many people looked at this and then did add to cart or what are the people who did add to cart check go to the checkout page and then drop off what were the things they did for which slides and nights of the people which set of audience did it and did do it or something should I send them a nudge those things can be better done through analytics there through something like a heat map kind of a tool second slide you can also the heat map also doing analytics or is it just storing videos it's heat map
00:35:41
Speaker
A sheet map captures videos. A sheet map just captures videos. Analytics is tough. Because once you capture all the data, then you have to store the data and link it, link the data, and then you have to enable sign-up and the same. If you just show what is happening, then you have to do the analytics in your brain. He spends time. He's going there. Why did he go there? So if you take a random video, he might go just for the sake of the thing. It'll be like an internet monkey there. So anyone will be randomly doing something.
00:36:04
Speaker
How do you analyze the pattern versus noise? That can happen only when you have the data stitched and then you can query the data. Without doing that, then everything would look like a pattern. And how would you do the linkages? How would you know that this person has also seen an ad on Facebook? That's why the JavaScript SDK is there and we also have the Facebook integration. So when people
00:36:25
Speaker
come from a Facebook particular campaign and how much they have paid to the cost campaign that we get from Facebook. Once they click on the link wherever it is through the UDM and whatever the cookie browser data, it comes to the website. He has done this. He has spent so much time and our JavaScript SDK is on the web page which basically drops a cookie. It's a first party cookie which we drop and then we are able to activate each and every sign in data or like even form full data and etc. We are able to capture that and then we are able to link it to the web.
00:36:53
Speaker
So one is when someone clicks on a Facebook ad and then comes, but second could also be that someone saw a Facebook ad, he didn't click on it maybe, but the next day he typed in mama and then came. Would you still come to know?
00:37:06
Speaker
We'll look at it more as an organic data point. You won't be able to capture when the user saw a Facebook account. That's more like an influencer. For example, you tell someone else, you check this out. I actually bought some merch. You should also do it. Then it's more like an influencer. But what you can also know is what is the ratio of organic to non-organic data. Non-organic is more like inorganic from campaign and other things, instantaneous bike. Organic is when people do this thing. This is basically the yin and yang on the browser. Yeah.
00:37:35
Speaker
No, it's basically the Indian Yangtas marketing itself. This one part is brand marketing, which is like, you do dot leadership, you do more blogs, you do just brand ads, you do YouTube ads, you do maybe even TV ads and other things. It doesn't mean to see. You basically build recall so that people are... Build recall? Yeah, when the need arises, they know what to type. They know what to do. But whenever the need arises, you should also do...
00:37:59
Speaker
ads, right? You have to also kind of like do ads, Google search on Facebook, so that they click and come rather than they go anywhere. Because if they need it, they can go to hundreds of millions. The classic example is Coca-Cola is a well-known brand across the world. They still spend a lot in ads, right? So because they have to keep re-hydrating your ad message and also make sure that when people think of it there, or they feel thirsty, they think about Coca-Cola.
00:38:24
Speaker
In a sense, the same way when people would know Marvel Earth completely, they might have even bought about it. But whenever they think about, let's say, any baby cat product or a mother product or something, they have to think about Marvel Earth. They have to build that thing. Or like in a B2B SaaS kind of a mode like Thresholds and other, the thing is Thresholds, everyone knows, and it's also the same. But whenever you look at a chat product for your website or whenever you are looking at, let's say, a service product for your ITSM or something, you have to look at Thresholds.
00:38:48
Speaker
or a charge we would put it in this thing. So B2B versus B2C, everywhere there is a brand plus performance marketing. We measure whatever that is there available in performance marketing. Wherever it's not available in terms of just view throw or influencer and or something of the sort, it'll come as an organic direct traffic.
00:39:06
Speaker
And you can also see how much your organic direct is increasing and how it's changing. That would give you a very good impact on, okay, this month we had done a lot of brand or Munich activities, which is just a lot of you, oh God, just generally, let's say like we released a rap video on our product or we released some kind of a PR, a lot of PR got done because of hunting news or something of that sort. That has increased a lot of organic because people see PR on Twitter or all those things or someone talking about it, or let's say Navel Dravigan talked about your product or something, suddenly there's a lot of interest in your product.
00:39:36
Speaker
These are all things which is basically would be captured indirectly in order that not directly but you still know in terms of these paths and all those things and you would have a kind of weightage around it like some kind of organic or inorganic part of it and also then even if people come in organically, people might come to your website in organically but after that how much time does it take to do
Complexities of Ad Targeting
00:39:57
Speaker
it? Once he comes in organically and later he comes later through an ad and then he comes back you can switch both of them.
00:40:02
Speaker
This guy actually took two months back but suddenly when he saw this ad and he came in and now he signed up for the demo and eventually he did a checkout. You can see the whole journey into it and that's where an analytics tool also helps. That's where the context within an analytics tool is not within a session or a lifetime. It's actually the whole user journey into it.
00:40:21
Speaker
But how does the website and Facebook talk to each other? So the Facebook website has a website, JavaScript SDK, and it drops cookie. Basically, Facebook channel data, everything is also linked to us so we know what's the cost. And Facebook also gives the channel URL, which is the same thing which we talked about in the mobile site.
00:40:38
Speaker
website also and the same thing. Whenever a campaign turns, there'll be a campaign URL and you would be setting up the UTM parameters there. When someone clicks on, let's say, Facebook or Instagram or something, they come to the website. So you link, this is the user who actually clicked on this usage and came to your, just saying. How does the reverse happen? If I searched on mama's for, let's say, Allergen free shampoo, chances are I will see an Allergen free shampoo on Facebook. Like for the next two, three days, I will see those ads on Facebook.
00:41:08
Speaker
So that's a retargeting cookie. So there are some tools where you can use basically like what are the SDKs the mama earth is dropping on you as well. But that's basically a retargeting cookie, which is you have to install a Facebook pixel on the website. Similarly, you would also put a LinkedIn pixel or something and you would also put a Google pixel. So when you had search, let's say mama earth, something you had normally going to mama earth. Wherever, let's say if you're reading NDTV or you are in Crecganford or something, the mama earth would add with a solid.
00:41:35
Speaker
So that's basically a Google display network or same thing when you're on Instagram also that things keeps following you. And that's basically like a retargeting cookie. So there are two parts to it. One is ads, initial direct targeting, then there is retargeting, then there is brand targeting, all three. And that's how it becomes so complex. You can't target, track everything and keep in mind everything. That's why a product like ours also makes sense.
00:41:56
Speaker
So many things happen, you have to know what is the kind of a journey? Do people come back most to retargeting ads or do they come back from the first club ad? Do people come and drop off very easily after retargeting also? Or how does it change by campaign, by creative, by channel? So you also help in retargeting like in terms of dropping that retargeting of it?
00:42:16
Speaker
We don't do retargeting, but we track everything. We have analytics, attribution, and insights. We capture every single data end-to-end. Our key value prop, which is that, is like everything you need to capture all the data, then sync the data, stitch together the data, and also bring analytics forward to you in an extremely super fast pace. So we have invested in huge amounts of data engineering. The way we prepare the data and assemble the data also makes the data very easily searchable.
00:42:44
Speaker
And the way we survey insights on which are the common parts, which are the more high converting parts. And also, with our product, that's where our AI comes in also. That is, honestly, we bring in data, we visualize the data, then we help you do analytics. The AI part is, you can specify any endpoint at any starting point. We can see all the parts going through that endpoint to starting point and rank them in order of relevance in terms of which are the most well-performing parts and which are not the well-performing parts. That's where our differentiation also comes in.
00:43:13
Speaker
that's where it comes. If you don't have all this infrastructure, one you have to spend on data engineers, everyone to build on this, then you have to have a data analyst to run these queries. Then you also have to optimize your data engineering and your analytics setup and all those things in such a way that where it will be faster also because when you're running management meeting, which is half an hour, you can't keep running queries and then saying that, oh, we're waiting for 10 minutes for the data to come out. So
00:43:36
Speaker
You have to also do it in a very fast way because speed also enables you to actually look very professional and also methodical for you to run experiments and tests around the data and other things, which is where we do. So now you've connected the pipes and all this data is coming to your data warehouse. How do you stitch it together? How do you help the system read that this is all Shri Krishna?
00:43:59
Speaker
So that's why the context in that we have built in the whole data points. We don't let the data be in one of the things. Anyways, we have a channel data connection. We have a website, JavaScript SDK, which is there. There is a cookie data. There is also whenever someone signs up, there is an email data, which is there. And then there is a CRM data, which is where an email record or a phone number and other things. So we take each of those things and we search it internally. And that's where at the user level and in a B2B places, both at an account level also. So for example, let's say you are buying a B2B product like Ring Mask.
00:44:28
Speaker
Which is there. There is a God which is a competitor and Bingman you are doing. But a purchase of something like a Bingman is not a one single person decision. Let's say your SDR guy will first look at the product saying that this is good. Then a CEO would also look at it. Then a CFO would also look at it whether this makes sense. Or a CRO would look at it and other the same. Multiple people who can draw on but it's the same.
00:44:46
Speaker
So we stretch user journeys by user level, how each person looks at it. Then we also stretch it at an account level, seeing all this agreeance to a particular account. So you can do analysis at an account level or at a user level, both sides. And that's the way how the data comes through. And once you stretch the data, analyzing the data becomes more faster and easier. So you have the unique idea and identify it and then you... What is the user journey once the data is staged? How do they run an analysis? Is it typed in regular English sentences? Can I type customers
00:45:15
Speaker
More on this, like in English. Yeah, tell me about how that analytics are. So you have a whole bunch of data in a storeroom which you have stitched together and therefore it is easier to analyze it. But how does the user actually analyze this? Marketing teams come into forms, right? One is consumers of the data.
00:45:34
Speaker
Second is like people who process the data, that is people who run specific analytics. So for example, the head of marketing and also maybe the demand generation manager saying that I need all this data, which is like channel amount of, let's say, signups among people coming in and revenue data all in a one single place. That's all we need. And other the same. One way we do it is we have standard templates that showcase the data.
00:45:54
Speaker
In terms of the set of the standard templates for this industry, we get the standard templates. So automatically, you will start standard templates. Other thing is when you have specific queries you want to look, I want to specifically have a separate LinkedIn query. Or your product marketer in your team would say, I want or a content marketer will know, I want to know very specifically how content impacts the revenue. Very, very simply, and I want to run the query. For them, they will basically come to the analytics dashboard and say that people who come from, let's say, website session,
00:46:20
Speaker
spent time more than five seconds on my particular content and other the thing and then eventually went to a sign up they'll be doing and then they'll be running a funnel and they'll be saving the funnel so that's how the any running new analytics queries and all those things like this you can do any kind of the thing i want to do this but filter by people who came from only your united states or they filter by people who came from Bangalore or break it down by campaign by channel so by google by facebook by linkedin or something like that
00:46:46
Speaker
You can do any of those things, which is basically almost as much as conversation in English, but with the more structured conversation. That is, instead of you writing, knowing SQL, it's more like people who came to this, then did this filter by break down with this. And that's why we'll be writing the whole process.
00:47:00
Speaker
Okay. Okay. So you'll ask them for a starting point, like when the journey started and an ending point, like they bought more than thousand rupees. That's the ending point. The starting point could be they clicked on a Facebook ad and then various filters
AI Algorithms in Marketing
00:47:14
Speaker
can be applied real time to see how that changes. Like how that changes and seeing what's the parts, which are there either you can analyze it at overall impact level or you can.
00:47:23
Speaker
analyze it as a sub-path level also and then run queries. And we'll also give rank and order of alerts. That's something like a search ranking in terms of all the people who came from Facebook campaign and then eventually signed up and became a customer or bought, let's say, a particular form of diaper and all those things. When people come from, let's say, Facebook campaign with this creative hands from, let's say, Hyderabad, the chance of conversion is much higher compared to anyone else, whether it's Bangalore or something. Average is only 2%, but when people come from Hyderabad, they lose 5%.
00:47:51
Speaker
So then maybe you need to double down and say, why is people from Hyderabad doing it? Maybe I have to, if I start spending more on my Hyderabad campaigns, things will work better. Or why people coming in from this, let's say, mothers above 35 are doing something like that. Or why when I see this particular creative, let's say, as I said, X person's creative versus Y person's creative, the chance of conversion better, let's me double down on that. And that's in automated insights. So those are the things which marketers love to do. The data actually serve its sister.
00:48:18
Speaker
How does the visualization happen if you would be showing that journey landed on website and then ran the search? One is like a funnel. Visualization can be in multiple forms. Either it's a funnel data or it's like an event and a standard row and columnar data. Depending on what's the kind of query you want to look at, I want to look for this particular number of people who did.
00:48:41
Speaker
sessions, website, this thing, I just want to know total numbers. By this filter, you'll get those numbers also by a Sparkline chart or something. Or you want to do a funnel, the funnel will also come. So much funnel, then so much drop down, so much time also between that and other thing. And once the funnel will automatically come, then you can save the funnel. You save the funnel and put it up into a specific dashboard. So you want to put, I want to see a separate Akshay's dashboard, which is there. Oh, this is my content marketers dashboard I want to do. It will be saved as a particular dashboard. It will automatically get refreshed every day or every week based on that.
00:49:11
Speaker
based on the data that's getting added to the system. You can also choose different kind of starting points in that analytics, which will tell you about attribution, like whether Facebook works better for users who spend more than 1000 rupees. You have separate attribution queries also, where you say attribution is more about end goal. For this end goal,
00:49:34
Speaker
What is the kind of path? What is the first touch part? What is the multi-touch part? What is the last touch part? If I filter for this angle by campaigns and I compare between first touch to last touch, how do the campaigns perform? In terms of cost per MQL or revenue per MQL or any of those kind of metrics and other things. What is MQL? Cost per MQL? Marketing Qualified Week. Marketing Qualified Week or sign up lead. Cost per sign up, revenue per sign up or cost per user, selection revenue per user.
00:50:02
Speaker
All those things. Lead is a clean metric that can be a sales lead or a qualified lead or a non-qualified lead or any of those metrics. Okay. And so what is the role of AI in this? How does AI make the system better?
00:50:17
Speaker
So this is basically from Aravit's work at the endator, at Google as well. The way he saw that is, you see so many paths coming in there. There would be hundreds of paths, people who went from Facebook campaign and went and bought a particular set of titles. Or it can be like came from a particular LinkedIn campaign and end up buying fresh chat.
00:50:34
Speaker
or threshold, fresh CRM product, or all those. What he had done is you can look at each and every path. Internally, what his algorithm does is it makes a kind of a tree-based design. So people did this and then did this, then did this, and each and every path. And each user would have user attributes. Each event would have event attributes. You can slice and dice through any of that. And then he looks for each tree path, which is there. What does the information gain difference? So whenever there is a higher information gain, which is an entropy metric,
00:51:03
Speaker
those things get ranked higher. So he calculates the me and then looks at wherever there is information gain which is either a up-leveling or low-leveling based on how far it is and then ranks them. So then he comes back and says how the output would work. Can you just simplify this a bit what you were talking about the information gain?
00:51:22
Speaker
So, information gain is more like kind of like deviant state or while average is let's say 2%, whenever people did this particular action, sit or particular event or when people came from a particular geography or a country or a city or something, the chance of that happening became 6% or conversion rate for that became 6%. So, that's like a huge information. But it's also ranked, that's 6%, but in somewhere else it can be even 40%, somewhere else minus 30%.
00:51:51
Speaker
So you see in the basically the expansion and basically you have an average but how does it deviate far away from that average and then you have to also look at significance with the same thing. Significances did 100 people do it or did 5 people do it? 5 people into 40% it's just 7 but 100 people into 10 but 10% is actually 10. So
00:52:11
Speaker
moving for this thing. And using these analytics, then you rank it. So of all the parts which are there, when people came from, so Facebook campaign to converting for a particular diaper is too cursed. But when people came from Delhi for doing this, and so many people did it, the chance of conversion is 10%. That's where it is. But when people came from, let's say Hyderabad,
00:52:33
Speaker
but mothers by 35, but it's only a small subset, but there it's 40%, but the same, that ranks next. But when people came from Chennai, nobody converted. That's very bad, because it's so bad from the deviant and all these things. It was deviant-sorbed up and down, both these bits, and then it ranks it. So in a sense, it's something like a Google News ranking on your data, which on your deviation, how does a Google News or Search ranking work? Do you basically look at something like, this is something which is interesting?
00:53:01
Speaker
how many people are setting it or how many people are clicking.
00:53:04
Speaker
how many people are looking at it, how many people are quantifying what is trending, then they rank that thing. So you can search for business news or something like worldviews. All those things will rank. Of course, now currently a certain Ukraine would be ranked because that's something which is a trending story, which is that then is the thing. How do you find the trend based on a certain kind of multiple metrics which goes into your information gained ranking? How do you look at this part versus another one, which is basically like you build all the tree and then see which tree path where there is a higher variance and then you rank it.
00:53:33
Speaker
The efficiency of this algorithm on the ability, why this makes more sense is like how you want the ability to set up the tree rhyme very fast. You have to generate the tree very fast. This happened, then this happened, then this happened. Somewhere else, something else happened. That is, you started a Facebook campaign. Then we know you came from Delhi. Then we know you bought it. That's like very high.
00:53:54
Speaker
But average, what is it? And then, but if you're same thing, if you came from Chennai, when you did the thing, it changed separately. But same thing, when you came from Hyderabad, but mother, the thing changed differently. All these things would remove certain RAM. But the ability to pass through this in a very fast way, and also probabilistically then, rank it also accordingly, is where the AI and more importantly, the mode also exists in terms of like how fast and how easily you're able to do it. This, when you are confident,
00:54:22
Speaker
What is the end goal of ranking these deviations? Why is that important to do it fast? More than doing it fast, it's also important to rank it very efficiently and in a correct way. The reason is very, very simple. So for example, as a marketer, when you start looking at data in any part, as I mentioned, the problem currently and why is this the trend which is more relevant now rather than maybe five, 10 years back is like,
00:54:48
Speaker
The amount of data points are also increased. So the way we store data is users with user attributes to events with it. For example, Sri Krishna from Bangalore will do certain things. This event time would click so much spent somewhere. Akshay will do something else and all those things each of the users would have. And this is the exponential sets of data because you would have so many event attributes, I would have so many event attributes. Same thing, user attributes also, city, website, browser, everything would change, campaign and all these things. When you have to see through all this data,
00:55:14
Speaker
and then see through all the data, that's basically number of hypothesis explodes. You can't manually figure out, you have to be a very high level of genius saying that of all the things that happen, I think the problem exists in this. I know conversion rates are coming, but I don't want the problem to exist.
00:55:30
Speaker
The fact that you have to be deeply ingrained as a revenue on this, but if you just want to do something like enquire this thing, I want to see what's this thing, you need AI to actually sort and drive this very fast. So for a marketeer, when they do a start point, end point analysis with some filters,
00:55:51
Speaker
In addition to that standard analysis and that average values, the system is also automatically showing them major deviations to explode. Major deviations to explode. Once you have the major deviation, then you'll have to double click on it. That's when the marketers will go back to the analytics thing. What do you call these deviations for a marketeer? Obviously, this term won't be used for the user. What do you call it?
00:56:17
Speaker
No, what we were saying is like what patterns I'm seeing on the data. So fundamentally, my ad to cards are increasing. Let's say anyone on a Monday morning meeting or a daily meeting would be worried about like how much is my signups and ad to cards are increasing? Is it increasing or does it decrease? Or it's remaining flat?
00:56:33
Speaker
If it's remaining flat or whatever it's decreasing, what are the factors affecting it? Is it my Google campaign? Is it my Facebook campaign? Is it people coming from Delhi, people coming from Bombay? What is affecting it? All those things, if you just put the start point and end point immediately, like this is affecting within this, these are the kind of deviation, these are the weak linksets.
Daily Use and Pricing of Factors.ai
00:56:51
Speaker
Once what's happening, then you double-link. Because the things that can affect the campaign are exponentially increasing because there is so many minute data points available.
00:57:02
Speaker
So they wind your data points. That's why you can take out the guesswork out of what hypothesis I want to generate. Otherwise, people will be like, I think.
00:57:10
Speaker
Delhi didn't work. Yeah, you spent a couple of hours in just applying one filter after another filter to see what is the root cause of a certain output. But now the system will automatically generate suggested... How do you brand this out for the user? What we call it as H-plane. We call it internally as the H-plane dashboard, which is there. But basically,
00:57:37
Speaker
Explain. No, no, EX, EX, CNC, IN. Explain is what we call this basically explainer outcome. It can be a single outcome, it can be a journey outcome, it can be any outcome. Can you just say, can you explain to me why MQLs are decreasing? Can you just explain what are the factors affecting explain for this time period? It automatically comes off with an output. Now I can deep dive into it and then figure out. Then I go into metrics, then I go into analytics and other things.
00:58:02
Speaker
So explain is essentially the magic of the product link. That's really the thing which would cause the wow factor for a marketeer. X-Link gives the wow factor but the standard metrics and ability to do analytics makes them keep coming to the product on a daily hourly basis because I need to look at my metrics. I need to look at what change. I need to look at how all the metrics I need to find. I can't keep looking at four or five dashboards.
00:58:26
Speaker
So what we enable is you can look at, we don't need to look at multiple dashboards, but you can also look at which are the key factors affecting your data in a very fast manner. So I think I broadly understand the product now. Let's talk about the economics of it. What is the pricing like and stuff like that? So do you price it per seat or what is it like?
00:58:45
Speaker
Well, we don't price it first. To be very honest, we are still evolving the pricing. How do we price it better and how do we improve pricing? As I mentioned in the earlier, this thing is like what we want to do is scale to a certain level of customers and have multiple case studies and have a lot of customers using the product. What we are optimizing on currently is like increasing the number of customers and improving the product usage metrics.
00:59:06
Speaker
Eventually, we'll come back to the pricing, but currently we are pricing only on monthly tracked users, which is basically the amount of data which we process at a user and an attribute level. That's the only thing which we are pricing on, which is the input cost for us also because we may Google Cloud on amount of data which we process, so we just want to do that.
00:59:23
Speaker
and that's how we are doing the pricing so that's like a very standard linear pricing which is that so we start with we have this thing like starting at $99 and going all the way up to custom like thousand two thousand three thousand dollars a month which is that depending on the amount of data three thousand dollars would buy how many users like tracking
00:59:40
Speaker
$3,000 would buy more closer to something around 4 million users on a monthly basis. $99 would buy something around 10k users on a monthly basis. It's not absolutely linear, but it's basically the amount of data that you have to process as well. And there's no limit on seats, like a team complete, 5 people team, 10 people team. There's no limit on seats. Any number of people within the 5 people team, multiple marketers, content marketers, everyone can use the product.
01:00:07
Speaker
You can set up separate projects within the product. You can visualize everything. It's super flexible in terms of how you want to do the data, how you want to see your metrics, how you want to set up your metrics, and everything else. And when did the product launch officially? So we basically started, I saw last April, that's when we started adding more clients to ChargeVee and other things. And we launched on the product hunt in September. So after September, we started making it more generally available in terms of as the product is going to be more open.
01:00:34
Speaker
Alpha means early customers, beta means testing, alpha is early customers. And as I said, what we are focusing on is now we are at 30 customers, we want to get to 100 customers as soon as possible, get to more importantly product usage metrics, the thing and how they use and how we improve both the UX and also the kind of data destinations and other things which we want to work on.
01:00:56
Speaker
And also more importantly getting into certain level of action. So those are the kind of subsequent things that you are getting to lose. Like you get the data, you analyze the data, you get the output, how the action the data so that it directly ties in with your revenue and outcomes. That's a marketer. How would like actually would be what like offering a HubSpot kind of a service where you can publish content, publish ads and run campaigns or like what? Actioning is a little differently like that.
01:01:23
Speaker
So we see user and journey and account paths, which users work, which paths work. So for example, you take people who have shown interest in your website and other the thing, and you have the set of accounts. Now you want to take that particular audience of people who are visiting and then run those.
01:01:40
Speaker
specific ad campaign for them or a rediving ad for just that set of people that you can do. That's the actioning part. That's what marketers would like to do. That's one thing. Second is you may want to just take off those things and run a specific email campaign for them. That can be another part of action that's the audience pooling or all those things which is there.
01:01:56
Speaker
You can basically then kind of like score these accounts and give it to sales team saying that now you also reach out parallelly or to sales team. Because of all the people who have shown interest, let's say there are 100 accounts who have listed your website, done these activities and other this thing. Of these 100, we think these 20 makes the most ideal customer fit for us and other this thing because they have also shown intent.
01:02:17
Speaker
in terms of using our product and other the same, they have checked on the website pretty well. And these strategy also match your ideal customer profile and the high ACD bracket and other the same. So let's go after that and send it to sales team to actually look at some form of action scoring them and other the same.
01:02:32
Speaker
So either you can run ads, you can send emails, you can start doing sales on top of the data point, which is that, which means whatever the input and insights which we are giving. And once you do ads, also we will also again capture it. You did add to this thing and then what is the result of that? How much you spent on the ads and how much you converted. Or once you do sales with the data and get updated in the CRM, we'll also capture that data also. So we basically complete the feedback loop, but also give them the answer saying that, okay, you focus on these two things.
01:02:59
Speaker
This is where this is there. And also, surface the data up in the systems of communication. That is like the systems of record in the current ages. Basically, there is channel data, there is CRM data, there is marketing automation data in multiple places. How do you assemble all the systems of record into one single Instra journey, which is that? Second,
01:03:20
Speaker
how do you make it more relatable and usable in the systems of communication, whether it's to your own email inbox, whether it's your Slack alert in terms of what you have to do, or in terms of running your specific campaigns and all this thing is bad with intelligence around that. That's why the intelligence, I can't show saying that so many customers signed up just to email blast all the customers. I have to say that now I apply intelligence of these customers, of these emails for this ICP, for this metric, which you have said, these things make more sense for you to actually focus on.
01:03:48
Speaker
These are the places where you have shown highest intent for you to actually focus on. This is where it comes. That's how the actioning part also comes in.
01:03:57
Speaker
And there are two parts to this product. One is marketing is a necessity. It's not a cost which I can cut or it's not something like which I can remove and then something else would happen automatically in losing marketing. It's something which is a consistent daily, continuous activity. That's why you need dashboards. Every day you have to look at dashboards. You need to look at metric. You need to look at attribution and do the strategy. So marketing equals your revenue. Now with marketing, how do you enable revenue and increase revenue?
01:04:22
Speaker
In terms of measurement of all those things, that's where the actioning also comes in. So then the product also, the value prop of the product is not something like, this is something I just show matrix. It's more like I enable you to get more revenue out of it. This would again be like AI driven recommendations and which could then like you could export an email list and run a campaign. AI driven scoring and metrology and other, the thing based on whatever BC of the data of your own data, first party data completely. But you would
01:04:49
Speaker
not, for example, build an email marketing tool of your node. You would just help connect to the email marketing tool. Maybe you can directly feed that data for a campaign or not. So in a way, this becomes like these tools, like the email marketing tool and the Pardo and the HubSpot. These are essentially workflow automation tools. These are not intelligence tools.
Market Focus and Expansion
01:05:12
Speaker
And these tools can be then, in a way, that the brain
01:05:15
Speaker
for running these tools can be factors that it is sending them signals it's like the nervous system of the body where your hand is doing something but it's not it's a it's an execution tool the hand and you get some input back from your hand like the thing you're picking up how heavy it is and then the brain will do some analysis and say okay this muscle over here should move over to
01:05:38
Speaker
Really lift it up. And so how big is the side of the market that you're targeting? You told me that for each of these segments, there are billion dollar companies like the data warehousing, the piping, the analytics tools, the visualization tools like Tableau. So how big do you.
01:05:55
Speaker
So, I will put it this way. So, fundamentally, currently, we are targeting only B2B markets, B2B and SaaS companies, such as all the B2B and SaaS companies, such as that. This chain product can also work very well for B2C. We do have a lot of B2C clients, but from a position and direction focused part of you, we are targeting on B2B because when your integrations become better that way, you focus on the night set of even data input.
01:06:17
Speaker
paths and all those things. And you also become your templates become more and more better and easily usable. And your action use cases and other things also become much better. So one is the focus part. Second, with the B2B itself, there are hundreds of active action tools and a lot of actually data point tools, whether you spend on Google, LinkedIn, Facebook and other things, the data, the user journeys and also timelines also shift and there are
01:06:42
Speaker
close to at least 80,000 V to the SaaS companies of which even if you say the market size of it is the 5,000 to 10,000 companies which are the amount of revenue they keep and the marketing spend is close to 30%. That's where it's like, if you are going to generate so much of revenue,
01:06:57
Speaker
and the thing you need tooling to get to that revenue more efficiently and focusing on for you to measure that revenue and action that revenue also in a more sustainable clear predictable way that's where our intelligence tool like us so our market size the way we look at it the same as a abm tooling marketplace so there are big companies like six cents demand base and others the same we also look at our
01:07:18
Speaker
ABM tooling, account based marketing. So in B2B, the side of companies which help B2B companies do marketing better are ABM companies, which is basically they work with the DSP, they work with the same set of channel partners and other thing to help it. So they are like five to six billion dollar companies, multiple billion.
01:07:37
Speaker
How are they different from what you are doing? What is an account based marketing company? So they do the, they have an intelligence or they don't have the intelligence in certain cases, but they basically enable you to basically run ads later. So we basically set behind that set of EVM tools and all those things is you don't need to running ads is more simpler, but intelligence to run and where to run the ads is the more, more pace this thing. So that's where we see.
01:08:03
Speaker
So at ABM Kapti, we would just look at Google Facebook, LinkedIn Analytics to see which ad. And they also look at DSPs and other things. DSP, what is a DSP? DSP is a demand-side platform. So for example, Google Facebook has explained the same mobile affiliate world earlier, right? So you can run ads on Google search ads directly. But when you're running ads on, let's say, Krikken 4 entity, it would be either from a DSP or it gets
01:08:30
Speaker
made it always be from the Google network, can be from an in-movie network, can be from the same. So how do you get access to all those inventory? You have to work with a demand-side platform like a trade desk and the same thing. So a trade house will also give you access to Google's own property, but it will also give access to other properties, which are their publishers. And in-movie is a DSP, or is it a? In-movie is a DSP app. SSP. Is SSP the right term? Supply-side platforms. Supply-side platforms are also there. Demand-side platforms are there. Both are there.
01:08:57
Speaker
Surprising platforms, aggregates all publishers. In movies, both. In movies, both. In movies, Surprising platforms, aggregates all the publishers, which is, again, the engineering publishers. What you were calling as affiliate. Affiliate and the same demand-side platforms, combines all the advertisers in a single big and address and network connections.
01:09:18
Speaker
So these companies are essentially just looking at ad effectiveness. Did people click on this ad? They look at basically helping you setting up ads and running those ads and other things of them, except for maybe something someone's like a sixth sense and others don't look at the intelligence behind those ad effectiveness also.
01:09:38
Speaker
Like, how do they help you set up an ad? Because they have data across multiple customers, what kind of ad works. And so they're able to... No, they don't do that that way. They basically help you saying, if you want to run a campaign for this particular audience, for this particular this thing, they basically give you the platform.
01:09:54
Speaker
Just like you can log into Google Ads, the thing and say then I want to target people the whole age 50 or age 20 for this particular demographic and this you can run start running Google Ads. The same way demand-side platform will also do it and other the same. But the larger market which you're targeting is all B2B and SaaS companies. So that's the market which we are targeting. Every market where
01:10:14
Speaker
I have one more question on this ABMs. So these ABMs, why would someone use an ABM? Why not build your ad on Google? Does ABM give you additional information? Like, for example, maybe on Google, you can't put a filter of companies with more than 50 headcount, but on ABM platform, you can.
01:10:31
Speaker
See, for example, you can filter a list and then run those ads. Google searches intent-based ads. If Google searches for that particular keyword or something, you want to run an ad. Instagram or Facebook is a property-based ad. You have to put in that property. When people are on that property, when they want to show interest in it, then you do add. A lot of those other display ads is basically non-intent-based and it's basically based on creative. So for example, I want to target people who are looking at a particular CFO magazine.
01:11:00
Speaker
Yeah, yeah, contextual ads, right. For people who are looking at a particular mother's blog, which is there, let's say for your own mother, let's say. For that, because they don't show anything, but if they are looking at a mother's ad, me showing a mother in the ad, then it's going to do it. So you need a platform which access to the publishers to give you access to those ads, or running those ads.
01:11:23
Speaker
But then you need a platform to set up those ads with the right kind of creative and other the same. That's where all these DSP platforms and ABM platforms are the same come to case.
01:11:32
Speaker
So we were talking about the market size. So you told me that B2B SaaS companies are your target. B2B SaaS companies are the same. So this alone for B2B SaaS companies, whole marketing measurement, analytics, attribution and action as an overall or the marketing intelligence platform or the marketing operations platform. So you can call it the same.
01:11:54
Speaker
A billion dollar business market is what we see here. And this is rapidly expanding also for two, three
The Growing SaaS Market
01:12:00
Speaker
times. One is SaaS is becoming a threat. Number of SaaS companies, number of B2B companies is expanding. I see SaaS in a place where what plays true was
01:12:08
Speaker
eight, nine years of parents just emerging that suddenly you have so many apps and other things coming through. SaaS is also coming through and the good part of SaaS is not a winner takes on compared to B2C kind of other things. So there are multiple SaaS companies by geography, by segment for the same activity and other things and each of them would need to spend on marketing to measure their effectiveness and increase their revenue and other things.
01:12:30
Speaker
Second thing is the number of tooling within all these assets, depending on the number of tooling and channels and everything is also expanding. Because there are sales forces, a great CRM, but it's still only 15 or 20% of the market. That is pipeline, that is spot, there is so many other things. Similarly, marketing automation has Marketo part out and let's say HubSpot and so many others marketing automation tools. Similarly, email would have multiple tools. Channels would have not just LinkedIn, Facebook, this thing, there'll be Bing, there'll be Corer, there'll be GetoCrow, so many other things. So how do you look at all this data?
01:12:59
Speaker
And the expansion mode is also more critically one of the very clear renders like almost all sales is having a very clear off online touch point. It's not just an offline only business. Earlier business news week 10 years back used to be like you do online, you do marketing, you do brand marketing. You may also even get inside sales for people signing up. But I also have sales representatives or account executives who will take up the phone, call up and do sales. That's also separate.
01:13:23
Speaker
Now it's basically like you take up the phone and call and do it only after he has actually tried with the product or at least seen the demo of the product or at least read your white paper or newspaper and this thing. So there is so much of online marketing activity which happens before any sales activity which happens. So hence
01:13:39
Speaker
The whole marketing to say alignment and measurement of all these metrics and all those things becomes more important with all these trends. The other previous two trends which I talked about that SaaS companies expanding a number of SaaS tools also becoming more and more complex. So hence the data is also complex for you to get. So these are the broad trends which we bet on. The macro trends compared with the micro market bear make so much revenue and similar companies and with the actioning makes it more useful.
01:14:04
Speaker
But so you're saying this B2B market intelligence platforms catering to B2B SaaS companies is a one billion dollar market. How big do you see it getting? One billion dollars sounds small and not to trivialize one billion is of course a big number. But considering the kind of jobs you have given up, the kind of penny green that all of you come from, you must be pretty ambitious and not satisfied in just a one billion dollar market.
01:14:30
Speaker
What we are looking at is very similar to how charged we were in early 2015 and other things. They were a subscription management platform and subscription painting platform which was there. At that point in time, subscriptions were like, okay, how many people are going to do subscriptions? There are only so many SaaS companies in the world.
01:14:48
Speaker
But only so many tech subscriptions are also unheard of. It's going to be one-off payments, whatever it is, and other the same. Suddenly, in the last, after seven years, in the last two years specifically, a number of SaaS companies doing subscriptions, a number of tech companies doing it. Anyone doing subscription has actually increased very, very heavily and then become more and more interesting. That's basically become the expansion in the SaaS market itself at a fundamental level.
01:15:10
Speaker
and also expansion and the same. Similarly, the same way we bet on the expansion in the SaaS market. So it's like, we do see on a year on year basis, like there is a 30 to 40% expansion in the SaaS market. The good part is SaaS is also not a winner takes hold. It's not something like all the companies are going to be in SaaS is going to die off also. That's the massive trend which I believe. The other massive trend is data itself is exploding. When data is exploding, you have to find
01:15:33
Speaker
more easier ways to bring together data such as data and analyze the data. You have to measure the data well to actually use the data. Otherwise, you'll have so much data and data warehouse, you're not going to use the data in any form and other things. So the data explosion is the thing. Third trend, which is more of a sub-trend, which is that storing data has become easy because data warehouse is a made storing data very easy. It's a constant trend. But analyzing data is very difficult because it's slick.
01:15:57
Speaker
So much data, but how do we analyze another thing? And also data engineering costs also have been increasing. You can't keep throwing people at the problem. Company, like, thresholds earlier or even another, the thing five years back, they would have said, this is the amount of data we don't want to spend on a tool, just throw five engineers, let them do whatever it is. Now the data engineering costs have also increased very, very clearly. So there is a SaaS over on growth trend. There is a data exploration trend and data tooling trend, which is also there, which is also a subject along with data trend and the cost of data engineering trend, which is also there.
01:16:27
Speaker
And how fundamental this has taken off because fundamentally how we are seeing this as there's been this product analytics tools starting to come up in the first half of last decade, which is there, whether it's an amplitude and the same amplitude was now a two or three billion dollar company. It was actually IPO and a seven billion dollar company level. And I thought the same because suddenly product managers became more data driven, which is a large trend, which started a trend because earlier product managers are more like brand based.
01:16:52
Speaker
Like amplitude would tell a product manager what is your user journey and help them improve the UX and stuff like that. UX and product and other things. No, marketers have become data-driven, very clearly. Marketers 10 years back used to be far more brand, storytelling, positioning. Now they are fundamentally keeping the data.
01:17:09
Speaker
They are quizzed very, very clearly within their own teams and their companies and their priorities saying that show the data this works, use the data to actually make decisions. They all want to be data out of the mind. They don't want to be caught off sand because it's fundamentally one means their own job. It means how industry is shaping and the data is also exploding and the number of companies using data is also exploding. That's the broader trend which is there.
01:17:31
Speaker
And that's why when the whole, why the market might be little under with me now, but whether it's you're going to charge me example of how we apply to product marketing data and analytics data to whatever the data solution. Now we are in the golden age of data, data everywhere, but how do you process the data? So I'm a CEO of water, but you don't have the right desalination plans and other things. You're not going to be able to drink anything. Probably put that on your websites, like data, data everywhere.
01:18:01
Speaker
Nice. Okay. So like, I'll put this in a, like a laban analogy, say Zomato's delivery business. See that as an outsider, if I was a VC looking at the business, I would have said, okay, I ordered out once a month or maybe twice a month. This is not a big market. Yeah. I would have probably written that off, which would have been a mistake because the ease of use of ordering on app and getting it within 30 minutes to your home.
01:18:25
Speaker
has changed habits. And the same thing is what will happen here. The ease of use with which marketers can crunch data, get insights and take actions on those insights will change behavior. And that behavior changes anyway, like a long-term trend, as you have more data. Kapsanis who use the data effectively that Darwinism
01:18:47
Speaker
And I have seen this play out multiple times. Bill Gates famously said it's always underestimated trend in the long run and overestimate it in the short run. And it's more like the product market fit is like finding the estimation right in terms of like you find it early, you build to that estimation and then you build it. And actually one of the key, the thing I also remembered as you mentioned is like I was doing the delivery series A right at that part in time. 2013.
01:19:13
Speaker
Almost everyone is like, what is e-commerce delivery going to happen? How many people are going to be in e-commerce deliveries? Or companies on this big e-commerce Amazon will do it themselves. Why would this happen? Now you see almost every single company becoming an e-commerce company. There is Flipkart, there is Amazon, there is also a D2C brand and everything. Everyone has their delivery. And now you have... Even within an apartment earlier, you used to be like one or two deliveries. Now, it has become an easy trend. You can actually transport anything from all the way from, let's say, Jaipur and Rajasthan to Bangalore within one day or two days or something of that sort.
01:19:43
Speaker
It has become such a use case. The pipeline and the whole tools for the whole business, it's like shovels for the gold map, gold running, has become very, very useful. The same thing is like the whole data tools and the data analytics and the data management to get actioning very easy and other things for marketers because you need to generate revenue. To generate revenue, you have to spend on digital marketing.
01:20:03
Speaker
If you're spending on digital marketing and this thing, you need to understand your user journeys and other things. So this becomes a fundamental tool. Are you going to benefit yourself or are you going to just look at multiple tools yourself? And is everyone going to do it or not? Then it becomes an expansion of the market in a very significant way.
01:20:19
Speaker
Even if you're not doing it, chances are if you don't do it, you will lose the market. The companies which use data well are the ones which will thrive. That's across the board how it has come through. There are a big 5x delivery companies for delivery companies earlier in 2015-2016 when we used to see it.
01:20:43
Speaker
Yeah, only a couple of them survived. And there are no amiches also within this, like Threshmino, Pervitinidos or Anish, purefoot Pervitinidos or Anish, and other other things. And then it becomes even more larger market. Got it,
Growth Strategies and Data Privacy
01:20:54
Speaker
got it. So what revenue will you close this year at? So this year our plan
01:20:58
Speaker
to close us closer to one and a half million dollars revenue but more like at a run rate level closer to more like 200k monthly revenue run rate is where we want to close it and the thing that's filling up more like my side to 10x very fast and other the same but that's where is also for us to also the product maturity to marketing maturity which we get in terms of our own go to market and then scaling it up to very very clear we are also experimenting from the pricing plans and other the same
01:21:24
Speaker
As I said, we are not focusing on revenue pricing per user. I'm focusing on getting to 100 customers and then getting to 200 customers. At 200 customers, I'd be far more comfortable with both the product market fit, product usage metrics, and then it's more easier to focus on the metrics of pricing, metrics of ACV, metrics of how you extract more value from the same set of users, how do you do expansion revenue and other things around the site. So when do you think you'll hit 200 customers?
01:21:53
Speaker
So currently I'm at the peak by 200. I want to hit it by end of December. That's fair. Okay. And by, by say 2025, what kind of revenue run rate would you be doing? I think my experience with delivery or the thing, and even with whatever which I had done is delivery people were saying that the series was a hundred crores overall value share. Now they're going to go for IPO at 30,000 crores.
01:22:15
Speaker
See, there was a low air in the wild, this wild imagination, we would imagine this to be a five billion dollar company, four billion dollar company from India. So impossible. Flipkart was a one million dollar company, how is this going to be a four billion dollar company?
01:22:31
Speaker
And is your focus on India or global? Where are you? No, it's fundamentally global. Fundamentally global for us, US, Europe. A lot of our customers are fundamentally US and Europe and other things. That's a lot I'm trying to think of. Fundamentally, we work on first party data.
01:22:46
Speaker
We don't work on third party anonymous data points which are there. Or Google Analytics as an SDK collects third party data as when it shows this thing. It's a free SDK because it helps in Google ad monetization. That's getting a lot of privacy controls. But first party data, whatever the data which you collect only for your own first party as a business which you there, that has itself is exploded and that has a lot of value and that's around the privacy compliance also.
01:23:10
Speaker
That's also another regulatory trend which you need to say like how companies working on first party data are going to actually expand and build out another the sync compared to companies working on third party and other data points and other the same and that would also be like we have to see how Europe and other the same it's like now of course the that is a black swan event in terms of a Ukraine war and other the same also but how is it overall going to our privacy trends within Europe was because early
01:23:34
Speaker
A lot of these Austrian quotes and others were saying that no data should be present in U.S. servers and other things. It should be in Europe. Maybe it should come to an agreement when it can be in either U.S. or Europe because everyone else is covered in the NATO treaty and other things. And then it's going to become more easier.
01:23:51
Speaker
And that's some regulatory trends is also even more difficult to predict, but some of the regulatory trends sort of become tailwinds for the overall market in very different ways. Because of the need for privacy and keeping your data away from Google. So using you basically keeps the data away from Google, like using factors.
01:24:08
Speaker
You don't need to have your own Google Analytics, plus it also keeps it private manually, you can switch all the journeys and you see the data. You also work with very clear norms in terms of we get people to accept cookies, you have the 70 refreshment of cookies and Safari and other the same. And then the whole user journeys and other the same measuring and other the same becomes also more cleaner. Measurement becomes more and more difficult with all the privacy and other things. So hence,
01:24:31
Speaker
the measurement across the board becomes more contextual and cleaner and you still need to measure. You can't work on more kind of like it's a hunch and other this thing. So you need a measurement and platforms which can measure will become more and more important. Got it. And you help companies take up blind also.
01:24:48
Speaker
Yeah, we are completely having a proper data. We are fully stock compliant. We are depending on the geography and other things. It's completely first party data. We don't share one company's data. For MAMA, let's stay down to say that, OK, this is what the trend is now. You can find, for some other advertiser or for similar set of people, you can show ads here and also we don't go any of those things. Got it.
01:25:09
Speaker
But you would, for example, be able to use data across customers to train the AI algorithms better. That's the same problem which Google also has. They use the data across customers to run their ad systems better.
01:25:23
Speaker
And the other thing, we don't plan to use the other thing. And that gets into the huge privacy compliance issues also. You have user-level data, you have PIA data in multiple ways. If you are using it beyond the certain set of businesses, then one, the businesses need to approve. Second, the customers need to approve.
01:25:41
Speaker
And then there is also things about what privacy from one form of the data to use for this thing. So that's where the whole compliance also comes in. Companies have approved to give their access to data only for this particular purpose. Then you use AI and trends to do for something else, then it becomes an issue. But for this particular purpose, for that company, they have to optimize how they have to get the users because that's the company's data roots, which is there.
01:26:02
Speaker
If you like the Found a Thesis podcast, then do check out our other shows on subjects like marketing, technology, career advice, books and drama. Visit the podium.in, that is, T-H-E-P-O-D-I-U-N.I-N for a complete list of all our shows.