How Factmata CEO Dhruv Ghulati grew to $52.8K revenue and 3 customers in 2024.

hello everyone my guest today is roof gulotti he's a forbes 30 under 30 leader in technology from uh for europe has built startups at on at entrepreneur first and tech stars london having started his career in finance at bank of america merrill lynch she trans uh transitioned into being a product leader engineer and scientist in the space of artificial intelligence and data science drew you're ready to take us to the top yeah let's go all right so you're building fact mata now tell us what the company does and what's the revenue model how do you guys make money yeah so so um fat matters vision is is that um we have a proliferation of online information um content coming from social media news sites and we've been focusing in the last kind of few years to how how to analyze that data how to make sense of it um what we've been missing on the internet a little bit is a sense of sort of a quality control layer on that content uh telling you if that content is credible safe uh and actually what's what it's saying from a deeper perspective so um what we're trying to build is a ranking engine for online content that understands content for its quality credibility and safety and who would pay for this kind of thing yeah so so what we're doing is in terms of monetizing that ranking engine is we take components of that ranking engine and essentially they are algorithms that analyze the tone of voice of language as well as they detect claims within that content um and we have two three main products today so one of them is a moderation service that can detect if a piece of content is is framed in a sexist racist or propagandist way and so we help platforms to moderate their content trust and safety teams um we have an api um that we monetize essentially any platform that wants to have tags on its content or filters on its content that tell you if that content's opinionated or propagandist or hate speech and so on um and finally we've been building a product this year that is essentially a new way of doing media intelligence and social listening so we scan the internet so that's twitter reddit facebook posts um news websites niche blogs we essentially within those extract the claims that people are making um the actual individual statements that people are saying so that we can provide analytics to brands government agencies consultancy firms and so on okay so you're upselling against those three product suites do you also have some kind of upsell based off number of seats or data usage maybe number of api calls per month yeah so on our api it's a number of api calls per month uh or based on volume uh in terms of our moderation service it's also based on volume so how many pieces of content we flag is hateful that were successfully taken down we get paid for that and in terms of our our media intelligence service uh we charge on a per topic basis so if you want to analyze claims that are made about coca-cola if you want to know those claims that made about the anti-vaxx movement uh the claims that are made about christchurch massacre johnson and johnson's talcum powder scandal will charge on a per topic basis interesting okay that's helpful to understand so give me we don't have time to go on every customer cohort down every single product line you have but give me a sweet spot on average what's the company or customer or brand gonna pay you per month or per year to use this technology yeah so obviously it differs but i'm gonna focus on where our main kind of revenue stream is right now which is around our expert media intelligence product um so um that ranges essentially um based on the type of topic and the complexity of topic that you're giving us but typically it ranges from uh two and a half thousand dollars per month um going up to even ten thousand dollars per month okay you say a fair average though might be like three grand a month something like that yeah something like that and paint so let's go deep on that so if i pay you 300 a month today about how many topics are you probably covering for me we're covering one topic for that okay covering one for that and typically what we have is we have clients wanting to you know you can imagine a big uh brand or a big pr agency or a big government they don't want to just analyze anti-backs content they might be thinking about all sorts of other types of fake news that they want to track uh or they might want to think about all sorts of topics that are relevant to their brand that they want to track so they'll take a bundle of let's say five or ten and then we'll charge on a discounted rate per topic um so per customer it ranges you know got it so it's not three grand per topic it's just your and your average is about three grand and then you know someone paying 10 grand a month might you might be covering 10 topics for them at a grand per topic because it's volume discount exactly i see okay put this stuff on a timeline for me when'd you launch the company uh so we launched the company in um so we raised our first check of funding and actually got started probably in november 2017. when was the first line of code first line of code probably november 2017. okay same and then and then you raised how much capital today uh we've raised two and a half million dollars today okay and how much was that first round do you remember uh that was a million dollars did you need that capital or would you do you wish you didn't take that dilution so early [Music] that's a great question um i would say um we in the nature of what we're doing right now started tackling a big problem like misinformation disinformation um i think we're lucky in a sense that we would raise capital um to experiment and build stuff even if that stuff was not client ready because we're sitting now like a year in where we've actually got the technology to then apply it to use cases whereas a lot of people who are now starting off having thought about the idea uh then have to go and build like a hp algorithm and they're like already 12 15 months behind so we knew that certain things would be valuable what's the team look like today how many folks yeah so we have 12 people um we have um three software engineers um five machine learning people and rest uh commercials so one salesperson um myself a head of product um and um yeah and a data product manager who helps us get our training data for our algorithms are are you and the other sales person are they quota carrying uh in terms of uh uh targets and so on no no like is it a true sales person do you have is it a true sales motion or there's a quota there's a commission etc yes absolutely yeah absolutely the way that we've done it actually is interesting because um our solution around detecting let's say claims of interest is a step change beyond existing competitors in the market so a lot of it is educating our customers and so the way that i've built our sales teams we have a lot of satellite sales people um in one in new york one in san francisco purely based on commission so we've managed to keep very very lean in that way so there's founder-led sales how do you retain that person so in new york if they're a good salesperson there's no way a small startup like you is going to be able to retain a good salesperson on a commission only structure someone else will come pick them off or there's not a good sales person um it depends because if the salespeople that we've got on on satellite are actually super relevant so they so this particular salesperson was at a company i can't say but basically our competitor right like this i've nabbed three of our competitors like top sales guys who are leaving uh or like um i thought you said you had two i thought you said you had two sales people today you you and one other person so me and other person the full-time guys and then we have these satellites oh they're not full-time okay and so you absolutely right um but but basically they have a black book of people you know 150 warm leads that they were speaking to before and so they get paid and i've done it on a very very high commission like like give me a general sense are you trying like 70 of first year acv 50 50. okay interesting and can i ask you when you add up all of your saddle i'm just curious as a channel if this is effective for you because i haven't heard of this kind of idea before across the entire satellite kind of commission army you've built how much have you paid out total over the past year um total i mean we're just going to market so i can't really say um but um but yeah i mean i mean i would say probably expected in the next month with the deals that were closing i would say probably uh five five out of the eight deals probably coming from from that structure okay i i'm just curious though from a volume so if like if your average monthly is three grand so first recipe is 36 you're basically paying 15k across eight five deals into that system you've set up yes so it's about 100 about 105 000 yes exactly but but for us that's that's it's about long term thinking versus short term right well yeah but the reason i'm asking though is because it drove you out of a cash gap issue right so if you're collecting monthly but you're paying 50 percent of your one questions up front yeah yeah yeah yeah sorry yeah so we the way that we agree our commissions is that they are that um you know we agree a schedule for paying them out i see about six months yeah okay so if someone closes a three grand a month deal you won't pay them 50 of 36 grand up front you'll do 1500 a month yeah yeah and just to be clear these satellite sales people are not they're people that i want to bring on once we raise more funding to actually take on more people no it's by the way i think it's actually it's actually interesting it's it's a smart way to build your sales person pipeline um we built a business actually funny enough fat matter because it's such a it's an idea that really captivates people around misinformation this information it's really in the zeitgeist um of the world um we have a lot of people who want to work here but we're also you know waiting for the right time to bring them on and so you know our engineering team was a satellite of you know 30 different engineers from around the world wanting to tackle this problem and so we've been able to build some really interesting tech um well hold on so how did you drew just get from a legal perspective how do you make sure that all these just passionate people that want to contribute code to what you're building how do you make sure you actually own the code right you're not paying them maybe they have a little bit of good contracts good contracts oh so you are paying them all contractually even though they're not full-time employees you're contributing contract labor well um it depends on contribution as well but this wasn't this was like back in the early days um but but basically you know ensuring that you have ipa assignment um you know in your contract is very very cool yeah but that's my whole point though drew is why so let's go let's let's play for a second you're 2017. you're you just raise a little capital you can't hire 30 engineers you figure you're telling me you figured a way to incentivize 30 engineers to give you rights to code they create their ip for free there's no upside and no cash event from you in the company i'm just questioning how you did that it's the idea that we built the world with it yeah but they have no upside i mean well the upside sometimes people have outside working on a problem that is is really important to them and gives them the upside is you know you worked on something a very challenging problem as an engineer or a machine learning person that no one else is working on you've actually worked on that you can take that into your next job but you keep all their ip and monetize it and they have no equity they get no upside in the enterprise value of the code they created um that depends on i mean i think that's a fair fair point but i think you know we were able to do it because of incentive you know we haven't forced anyone to do that so by the way that's why i'm pushing here i'm trying to decode the pattern you use to drive this incentive structure early on because obviously if it worked there's other people that would want to do the same thing yeah absolutely i think i think for for us um and the way that i care about my business is that if what we're doing works out uh we've had this this thinking across the board so for example one of the ways that we are building our system is trying to build a community layer that helps to train our algorithms to detect fake news hate speech um you know bias and content and we have our own community of journalists that we've been working with now ultimately if we start to monetize the algorithms that are being built in a substantial way because those original people were actually helping to train the algorithm uh we actually want to give them some revenue share from me why you don't need them after you're at scale you know dang that's like uber saying now that we're rich in billionaires we're gonna start giving our drivers more money no it's the opposite it's called network effects you have more leverage in that case and there's no incentive for you to give more money to them uh it depends on if you're trying to build an ecosystem or you're trying to make money right so if you're trying to get you know if we want to solve the problem of and we're seeing this right now in the world around misinformation disinformation and you look at the big platforms for trying to get the same community that i'm trying to get right to train their systems a lot of those people don't want to work for one of those some of those big platforms a lot of it is because they don't feel like they're getting a fair deal right and they feel like that exact thing that you just said where you know you're going to be contributing to work you'll get paid for it short term but then you'll never need what are you giving them there's no deal with you it's a promise of potentially something in the future absolutely absolutely so how's that a better deal than like you're you're saying it as if you have some better deal with them you have no deal with them right now um i have a deal with them absolutely as a promise in the future um and that you can bake into contracts and and that's also an element of trust as well have you done that you've built that under contracts future rev shares for people that help you train your algorithm uh that's it's not something that we we uh but that's it's not something we have had quite frankly time to actually think through the structures of that but that's definitely in our road map to in our api for example thinking about you know how we're going to farm this out but that takes a lot of time to think about penis structures and you know how it's never been done before machine learning really it's like how do you tie together an algorithm to the actual annotators of that algorithm and how much they actually contributed to improving interaction the closest thing the closest thing to this same pattern is because there are many multi-billion dollar companies today that start off on an open source community where there are a bunch of people contributing to code for free they now with their teams contribute x amount of lines of code for free to get back to the community but they obviously are making all that money themselves they're not paying out money to every contributor now there's a lot of companies that are based on blockchain where they are doing this everyone does have upside because they own tokens the more code they contribute the more tokens they own and it really is driving the community yeah and and that's great because tokens you can actually quantify them out and that's based on the quantity of code that you corrected but you can argue that you know is that real again this is a hard problem you know just because you contribute loads of lines of code they might be really shitty and you know um there's an element of provenance basically here that we're trying to solve and in machine learning it's very difficult yeah um how many customers are you serving now today uh we're serving three customers okay good so it's very much an enterprise sales motion then yeah yeah okay so you know yeah so three customers three grand a month they're doing about nine ten grand a month right now in revenue yeah zero a year ago [Music] yes zero year ago okay cool so yeah i was gonna say so you basically between 2017 and today have sustained yourself off the the rounds of funding uh arbitraging labor to incentivize them without having to put up cash via other kind of mechanisms of incentivization this sales commission structure where they're not on your actual you know p l every month it's it's upside only i mean look a very interesting way to build a company how aggressive are you being with burn are we talking like 20 grand a month or 100 grand a month or more uh that's about 100 a month i mean can you sleep okay at night with that are you comfortable there uh it's been challenging um because we've also we're not vc backed we're purely angel backed um but i think what we've managed to do is um you know what we're building is is very difficult from even from a technical r d perspective um and so yes you know we're talking acvs and sales right now but it's taken time to actually figure out where we are right now uh we're now in a very big inflection point we started making these first revenues and now it's time to get the sales people in um but but um uh you know that in terms of raising purely angel funding for two years not many people are able to do that and i think it's also the nature of the problem that we're solving is not just about a big commercial business opportunity but it's also an opportunity to start and make a dent in an ecosystem so how much when you look at the total cash you spent building the mvp before your first dollar revenue what would you pin that number at um to build the actual product yeah all the tech expenses up to your first dollar of revenue yeah i would say probably around so not including like marketing and and include everything just everything all expenses before your first dollar of revenue yeah 32 million dollars okay interesting cool so you have plans to raise them now yes how what's the right amount to raise uh we're raising enough to give us 12 months um so uh that's roughly about um sort of 1.5 million dollars yeah you're burning 100 a month right now 12 months right um will you do it on a note or will you price it i will price it okay so what so i mean you have to then tell a good you don't you want you definitely don't want to get valued on a financial model because your two the revenues are too small that wouldn't be what you like so that means you have to tell a really good story to get the valuation you want uh what evaluation do you hope to raise on obviously you have to negotiate it but what's your hope uh we've already set our valuation uh we've set our valuation right now i i got the existing investors to convert a certain amount the way that we thought about that valuation amount was thinking about what we're going to raise in our next round which is our big series a let's say that's five million dollars um typically dilution in the series a is sort of 18 to 20 so let's take 19 as the dilution number um you know that's going to get us to uh 19 million uh you know minus the five so that's 14 you know and assuming that we want to give our existing investors double their money you know you bring that down to seven and yeah so you rate seven was post or pre money on the one five yes uh pre okay so so eight five posts yeah um interesting uh on on basically a hundred grand 120 grand uh kind of a year in ac in revenue right now yeah yeah people must be got to believe in the tech that's good you need more deep fakes you need more big you know crazy scary stories to scare people to give me a higher evaluation um i think i think we have a lot of things wins in our sale i mean the u.s elections are coming up um we you know people are very scared about this problem there's regulation discussions around europe major ones um and you know we've had some big wins we're one of the first government suppliers in the world in this problem um in at least in the uk yeah you know we can now tender and bid for big government contracts and so we're thinking about um essentially getting the base in place to start you know inflection point do you know do you know just quickly because we're out of time do you know what you're spending all in to get a new thirty six thousand dollar a year customer obviously you have 50 commissions if they come in through commission but any other expenses um i would say service fees uh sometimes at the initial part of setting that topic up to then get monitored um it probably takes us a week um of of kind of three or four machine learning engineer three three machine learning agencies you're spending like between 15 and 20 grand to get a new paid customer at 36 grand a year yeah roughly yeah interesting yeah right because you got 15 grand in the 50 commissions right and then your service stuff or whatever else yeah and remember this is the beginning so once we create a topic that's a fantastic thing about this we create a topic let's say i'm analyzing a few anti-vaxxed i content sell that to lots of people yeah that's that's public data economy is the scale it's good stuff let's wrap up with the famous five number one what's your favorite business book [Music] um my favorite business book um don't have one i don't read business books number two is there a ceo you're following or studying uh i would say elon musk number three what's your favorite online tool for building your company uh shift try shift uh it's an email aggregation gets everything all the tools in one place number four how many hours of sleep you get every night um seven and uh age uh 27. and what's your situation married single kids uh girlfriend okay no kids no kids all right last question what do you wish your 20 year old self knew um i wish my 20 year old self uh knew uh to chill out take a chill workout guys fact mata trying to fight misinformation they've got three customers paying three grand a month right now so nine grand a month in revenue from nothing just a year ago uh they've got a structure where there's essentially you know eight full-time people oh sorry 12 full-time people eight engineers two quota carrying folks but they have a outsourced sales team that's commission driven only that he thinks is going to drive their next five-day sales burning 100 grand a month 2.5 million raised to date raising 1.5 million right now on a 7.5 pre as they look to continue to scale drew thank you for taking us to the top thank you so much