Trascript
Fryderyk Pryjma (00:18)
Okay, Fede, pleasure to have you here. I'd like maybe from you introducing yourself and telling a bit about your startup, Sofutu. Yeah.
Fede (00:27)
Pleasure to be here, Frederick. I'm Fede Weiner. I'm born in Buenos Aires and I created with some partners and colleagues and friends different technological products and projects lately. And you mentioned Sofuto, which is one of them. It's our startup in the European, Estonian ecosystem mostly, but we also operate in South America.
So our vision is quite global. are very, it's a mixed background and not only in terms of professional qualifications, but also nationalities. So this is a bit our trade off on this.
Fryderyk Pryjma (01:10)
Okay. So who you are partnered with doing that startup? So if we do, there are some other like as well products in the meantime, you do, right?
Fede (01:18)
Yeah, correct. So I can mention Julio Naya. I can mention some Estonian people like Auntie Reintal and Timo and Lisbet. They are all from from Tallinn, Tartu and surroundings. And there is other people that, you know, it's more I would say part-time because when you start something, not everyone started 100 percent. you say they...
join part of the journey or they dedicate partial time and I don't want to let's say forget anyone so I think those are the ones that I am more frequently interacting and all to all the others sorry for now maybe not mentioning but you are you know how important you are on everyday tasks and yeah maybe next time when we have more time we can make a longer list
Fryderyk Pryjma (02:10)
Yeah, okay, good, lovely. So basically here, I'd like to speak about the design operation on your product. So let's start with maybe how the product idea came. I read in interview with you from the Techno poll that you were sort of within this film industry and you've seen the space for creating this, yeah.
tagging, video tagging, automatic tagging software. So if you could say a bit about what your product does and where the idea for it came from.
Fede (02:46)
Yeah, this is the first product we launched and it came up from a project in Bulgaria with one customer, football team called Beroe and actually as every professional sports team they need a video tagging solution to analyze games. When we started talking, they said we realized that there was a space for something new in the market.
because the solutions that these types of teams can achieve, I'm not talking, know, multi-millionaire clubs and players from the Premier League, Bundesliga or similar, they were quite basic and actually they didn't work as well for them. So we decided to make a roadmap and then always in the...
Fryderyk Pryjma (03:19)
Mm-hmm.
Fede (03:28)
say last mile of that roadmap was to integrate AI and a backend that can process video and auto tag. But the first product we launched was autonomous video tagger focused on sports coaches. That's called VisioTag. Visiotag.com is their website. And actually we have, let's say like 80 coaches working with it between Europe and America.
And yeah, that's a manual product, I would say, because you use locally, there is no cloud on it. And it basically helps you to cut any sports video quickly, or at least quicker than what you can do on Sony Vegas Pro or iMovie or any other video editor you can use. Then after we...
completed and launched that, which was one of the commitments we had with Tenopol, which is the largest business and science park in the Nordics based in Tallinn, Estonia. We went for the backend and we developed this in partnership with Amazon Web Services from Helsinki. We decided for Amazon because it's one of the leading platforms, but they also believe in our product from the very early beginning. And they granted us not only
credits but they also invested in our idea. So what we did is to add an algorithm in the back that can analyze objects, transcribe the video and then help the editor to tag quicker and faster. And the target for this product, which complements with the first one because the first one is the front end and the other one is the backend, is any type of video editor that needs to process
hours of shootings. It can be a documentary director, a TV industry, or TV program, or YouTuber that has a lot of archive or just a lot of shooting and then to quickly ennace their process up to five times, they can use this solution.
Fryderyk Pryjma (05:27)
Okay, and like when you started doing this with this sports team from Bulgaria, was it right? ⁓ How did you create the idea for it? Was it replicating of already existing software? Was it in a way like a new solution? And how did you convey from the idea or the problem that the
Fede (05:34)
Mm-hmm.
Fryderyk Pryjma (05:50)
Or was there a problem that you identified that you wanted to solve? Or how the idea came up to solve it in this way? So you mentioned that there is other software, of course, similar to that. So was it more about replicating what was already done or?
Fede (05:59)
I think it's a good, good.
I think it's a good question, Friderik, because this is also like in arts, how a musician creates a song. He might get inspired by many different interactions in his personal life, in this artistic life, in where the music style comes from and so on. I think in software it's exactly the same. Of course, normally there is something out there that resembles to what you want to do. And it's good to...
experience it and see if you see a gap that you can improve and that's exactly what we did we tested the software out there and let's say for this level of analysis that they needed and the budget they can pay we compared it to those tier levels and we thought hey we can improve that we can do it better so you don't start coding till you interact with as many users as possible to shape the idea and
let's say, make it resource efficient because it starts writing one line of code. So we did that with the First Division Coaching Team of Beroe. We also interacted with other coaches out of this team just to see their, let's say, their taste for it. So yeah, we did that, let's say, a video tagger, but it's not another video tagger. It's easier and simpler to use than other solutions out there. It simplifies...
the UX UI to a very minimalistic approach with enough flexibility to run it on Mac or iPad and let's say an efficiency cost that is unbeatable to others. Of course, that doesn't mean that the product will be 100 % successful from the early beginning as any other product can be. If we think about the first iPhone and how many sales Apple had with that and how criticized it was,
then definitely you cannot imagine that, you know, a decade after is the dominant player in the market and still today is the most appreciated, let's say, device for people in tech. So it takes time that it might be successful or not from a commercial perspective, but it's already successful for the fact that in three months we achieve it, let's say, to generate the idea, to code it and to make it to the Apple marketplace.
Fryderyk Pryjma (08:22)
And if we think about this very early moment, so when you started comparing those other solutions and then talking to the coaches, how did you do that? Did you do that remotely, in place? Did you draw something on a physical board? What was the interaction with the coaches and people that probably were not so good at tech? How did you manage to sort of pull out the most important data for you?
to create a solution that would be useful for them because now you say that it's around 80 coaches. So I can believe that there are some other teams using this as well.
Fede (08:57)
Yes, you know, I believe that for anything you want to create in life, you need to have teams with people that is not from that field. So you really get feedback, another point of view. So you are not biased only by coaches or only by software engineers. If I had my mom at that moment, she was a psychiatrist, I would have invited her to the design thinking workshops we did.
Fryderyk Pryjma (09:14)
Mm.
Okay.
Fede (09:22)
from a user perspective just to see how she can interact with it because the more you can get on data collection, normally the better. Of course, you don't need to arrive to saturation. It's the same than university research. There are some limits out there. You know when you have had enough normally. Experts in qualitative research mentioned between 20 to 35 cases depending on the topic.
So would say that's a good number to follow. And we did most of them, let's say, hybrid. So maybe first session was face-to-face and second session remotely. We tried to do kind of focus group approach. So you also learn on those interactions between the users, but it's not so easy when everyone is working and particularly in sports, people travel a lot away home, there are trainings. So you need to be a bit flexible on it and try to make the best based on your resources. So...
Fryderyk Pryjma (10:11)
Mm-hmm.
Fede (10:15)
If anyone listening wants to create some piece of software, I recommend this exercise, I recommend design thinking workshops, because those are, when you have multifaceted people in there, is the most, let's say, valuable asset you can have before start coding.
Fryderyk Pryjma (10:23)
Mm-hmm.
Okay, and then how did the project during the time iterate? Did you develop it further or did you focus on these other applications of this technology and developing the backend you referred to previously? Because I know that you have lots of pilots like in different spaces. You told me about the pilot in one clinic, right? For the police, was it for the fire department as well?
This is a lot of players, like, of course, as you say, like, and this really cool, what do you refer to? So if your mom could use it, everybody could probably use that. So that's a good way of thinking. But yeah, I just wanted to understand more, like, how do you collect user feedback? During the time, how do you interact with your customers, with your pilots? What sort of data did you manage to pull out of it? What are you happy with?
what you maybe would have corrected. Yeah.
Fede (11:27)
Okay, great. Look, when we started, we made the roadmap that in three months, our product should be in the Apple Store. And this product will be for sports coaches. And that's what we did and we achieved on time. And that's not easy to do in summertime, I would say. And actually with few hands, we were only three people working at that moment on it. It started with basic sketches on paper.
Fryderyk Pryjma (11:38)
Mm-hmm.
Yeah.
Mm-hmm.
Fede (11:51)
That's, you know, it's the basics. Let's get back to the basics. There are so many tools out there, Mido and dashboards of whatever you want. And actually, I think I still believe that the best is to start with a pen and paper because it's the way human brain can, let's say, be more focused and direct to what you're thinking. So we draw in there the first windows from for for the Mac version. Then after all, we do the same for the iPad version and
Fryderyk Pryjma (11:51)
Mm-hmm. Mm-hmm.
Mm.
Fede (12:18)
And then we started interacting with the users to see how they react to those drawings. And you can also discuss where the button should be, because we might be biased. We might think he's here, but for the coach he's there. And that's exactly the approach we took. after this process, you gain insights. It always can improve. I mean, it's not a process that you do once and that's it.
Fryderyk Pryjma (12:24)
Mm-hmm.
Fede (12:45)
Ideally, you iterate as much as possible and when they became real users, then you have real feedback because it's also the pain point of them being pain users that they want solutions yesterday. So then you can see this pressure as well and it's good for both parties because it obligates you to put the bar one level up.
Fryderyk Pryjma (12:48)
Yeah. Mm. Mm.
Mm.
Yeah, yeah.
Mm-hmm.
Fede (13:09)
you always, like in Olympics, you want to always try to jump a bit higher just to beat the others. And the most important thing is always to find a focus where you can make some real life solution with your early stage MVP demo prototype, whatever you want to call. And then it comes the second part, which is the algorithm that does the, let's say, automatic...
Fryderyk Pryjma (13:13)
Mm-hmm.
Mm-hmm. Yeah.
Fede (13:36)
identification of those tags, more focused again on TV, film, advertising, YouTube. And in here, we believe that any video is subjective, a potential client when it's long shots, when it's over 20, 30 minutes, you know, and you need to fill up for your social media or for your manual editor afterwards, like Premiere or Final Cut.
Fryderyk Pryjma (13:37)
Mm-hmm.
Okay.
Mm.
Maybe do some shorts later on, right? To upload something
for the social media and stuff. Yeah, yeah.
Fede (14:07)
Yeah, but we've been working with a filmmaker from Peru recently and she explained a case. She was with 500 hours of shooting in the mountain and she was looking one specific shot with one car. She couldn't find it for two days because they didn't target it properly. with our solution, she could have prompted, show me a shot where there is a car.
Fryderyk Pryjma (14:10)
Mm-hmm.
Yeah. Okay.
Mm-hmm.
Okay.
Fede (14:31)
And then she gets the shortlist, right? And then she can select where the location of the reel is. So this is a bit the use case. But then, of course, when you are in a startup mode, you try to take as many fishes you can to survive, because then otherwise you starve. So when we saw opportunities in the security business, like police and city security and firefighters,
Fryderyk Pryjma (14:35)
Hmm.
Mm-hmm. Yeah.
Mm-hmm.
Fede (14:56)
we can detect patterns out of this algorithm on the video. So we can say, OK, let's see patterns of whatever you want to track. And the police asked us, this is actually in Buenos Aires, there is a crime wave with motorcycle crime. basically two people ride a motorcycle. They run on the street. They see someone talking by phone. They take the phone or something else like a backpack. And then they
Fryderyk Pryjma (15:00)
Okay.
Yeah, yeah.
Mm-hmm.
Mm-hmm.
Mm
Fede (15:24)
go straight away out. So the police ask, hey, can you identify these type of patterns? So we did a demo for them. And actually, we can we can see when there is a motorcycle with two riders. And maybe if that motorcycle is five times in the same camera over X amount of time, we can send just another to the patrol in the area saying, hey, it's better if you check this.
Fryderyk Pryjma (15:38)
Mm-hmm.
Mm-hmm.
Okay. And actually, like in regards to those different sorts of fields of applications that you discussed this solution with and you introduced it, you delivered it. How about like feedback from those sorts of use cases? Did it need you to change slightly your solution or did it work relatively well in all of those cases?
Fede (16:07)
I think the most important thing is to know how to say no or maybe later on because users will always want something. The point is, are they willing to pay for it? If they are willing to pay, then you need to focus on it and decide if it's part of your roadmap and it aligns, let's say, with your vision. If it does, take the money and do it. Make a plan and develop. If it does not,
Fryderyk Pryjma (16:13)
Mm-hmm.
Mm-hmm.
Mm-hmm.
Okay.
Fede (16:35)
be polite to say is not on our focus and try to show them how you can achieve the same with something different, some different work around. And if they are not willing to pay, that means they probably are not so interested in it. And they are just asking because we all want ice cream on Sunday, but you you need to pay for the ice cream.
Fryderyk Pryjma (16:38)
Okay.
Mm-hmm.
Yeah, I really like this. Yeah. How you refer to that? And I think it's very smart and very important to say no many times, because of course everybody thinks best of what they would like the product to look like. But then the truth is whether it actually will help them solve their problems. Because you are essentially a person who is steering this ship and might create an opportunity for them to solve those problems.
And somebody might not be sure whether adding this functionality will actually help them or disturb them from fulfilling their goals. Yeah, that's very good points. Listen, I wanted to touch two more points. One is who actually worked on the product itself? Who was involved in it from the perspective of the, let's say roles. Cause as I get it, you are sort of, if we relate to enterprise style,
of naming convention then you would be some CPO at the same time, so Chief Product Officer, or you leading the product. Who and what sort of roles were involved in creating ideas for the design of it and crafting it well for the users?
Fede (18:03)
Okay, know, titles means nothing. That's my learning from corporate world and I have had different titles across my life and I made tasks which were not necessarily what the title said and it depends what you are willing to do, what you really do at the end or what the company wants you to do. So it's a two way road like any other relationship. Otherwise it doesn't work. But when you're in a startup, I would say those
Fryderyk Pryjma (18:06)
Yeah. Yeah.
Okay.
Yeah. Okay.
Yeah.
Yeah.
Fede (18:31)
Lines are more blurry because everyone needs to be... I compare this to war in the sense of when you're in a big corporation, you are like in an army and everyone, one is a tanker and the other one is infantry and there is one more guy on the Navy. So everyone is divided between real different tasks despite everyone is in the war. When you're in a startup, you're a bit of a partisan. So you do more things than others.
Fryderyk Pryjma (18:32)
Yeah.
Yeah.
Mm-hmm.
Yeah. ⁓
Okay.
Fede (18:57)
mixing those roles just because there is not enough hands and the project probably demands it so when you cannot externalize it for budget reasons because maybe you don't want to externalize it so any other reason you might find then you are more a partisan so you can win the war anyway i mean everyone tries to be a unicorn but it's not the only i would say not the only way to learn is to win sorry is to be a unicorn
Fryderyk Pryjma (19:06)
Yeah. Yeah.
Okay. Yeah. Okay.
Fede (19:23)
⁓ The only other way to win is to learn from the experience, to enjoy it. And then, of course, making it successful is good to have this sharp vision that we are going to make it. But if you didn't make it for any reason, maybe the market was not prepared yet in there. I there are many cases of, let's say, creators across the history, not only in tech, but in many other stuff that we use, that they were too early or too late. And that's OK.
Fryderyk Pryjma (19:24)
Yeah.
Yeah, yeah, yeah.
Mm.
Exactly. Yeah.
Yeah, I actually don't. If I would relate slightly to the interview you said in techno poll as well that actually it's of it will be unique or as well. So I keep cross fingers for that. So yeah, I understand here that everybody who was basically involved in the project from your side and they were. Acting sort of as a UX designer UI designer user researcher like.
Fede (20:01)
hehe
Fryderyk Pryjma (20:12)
whatever it was needed. basically creating the product. I like also when you said about drawing on paper, this is quickest way to iterate, to prototype, to visualize, convey the idea, which is essentially what design is about, to convey the idea. Yeah. Well, so lastly, I would want to ask about B2C. Because when I think about your product and when you speak about it,
I right away think about maybe B2SmallB or B2C model. So focusing on influencers, like, you know, they create heaps of content, you know, filming during the day, their life, creating vlogs, and then being able to talk quickly and then through big data and loads of video they've created at cumulative time.
They can create some sort of new cuts, new shorts, new videos. Have you tried to target this user group because they seem the most approachable at the same time and the most easy to sell instead of like police, fire department and that's that.
Fede (21:23)
Yes, we did and we do. And actually we are now in negotiations with a YouTuber channel, YouTuber person, which has a channel with more than 600,000 subscribers in in Americas. Actually, he's based in Buenos Aires. And the idea is to, let's say, co-develop the product because we need to validate the software.
Fryderyk Pryjma (21:34)
Mm-hmm.
Mm-hmm.
Fede (21:52)
beyond the football teams now because we already have, let's say, the solution almost complete. However, we need real users the same way we had real coaches at the beginning of the, say, front end of the same product but, you know, connected to different targets. And also there is a line in there what the naming will be afterwards and how we separate the projects because...
Fryderyk Pryjma (21:53)
Mm.
Mm-hmm.
Mm.
Mm-hmm.
Fede (22:18)
We have also different stakeholders and shareholders of each of the agreements. So they are separated, but at the end, they should convey some way. And that's a negotiation we still need to do internally and externally. But coming back to the YouTuber, yeah, they should be the first users and they get some benefits out of it. We try to deliver a scheme that if they say bet for us now when the product is in development, then they will recover those investments with us as soon as we upsell.
Fryderyk Pryjma (22:22)
Mm-hmm.
Mm-hmm.
Fede (22:46)
This
is a quite common model in many consultancy companies on software and we believe it's a good model for those who like to bet and be innovative on new things. But hey, we are open to new cases. We are also discussing this in Kazakhstan and Uzbekistan because our solution works in 23, 24 languages at the moment. So we wanted to be prepared to...
Fryderyk Pryjma (22:50)
Mm-hmm.
Mm-hmm.
Mm-hmm.
Okay.
Fede (23:12)
other markets in this regard and you know those are big countries with if I remember well 20 to 40 million between adding them to so it's 60 million people and there's tv channels in russia with that need so we we had some contacts and we are trying them to let's say to use it of course it's a challenge
Fryderyk Pryjma (23:21)
Mm-hmm.
next
So my take on that
is with YouTuber. So you still see some sort of spaces, some features you are missing in order to fully satisfy YouTubers. Now, is that right? Because you say that you discuss with them investing in the product and developing it a bit further. So it gives full use. If you could say what sort of features you are now thinking of developing and designing and like what's on your roadmap, what next?
Fede (23:57)
That's a good question. So we did the same exercise that with the coaches, we sit down with their editors and ask them what do they want. And actually what they want is a quicker way to find footage that they need to put on air again. This is a YouTuber that is kind of TV channel already. I there are 60 people working in the channel between journalists, editors, producers. mean, this is a kind of, I would say bigger than many TV channels out there. And yeah, they are not 24 seven.
Fryderyk Pryjma (24:16)
Hmm.
Okay.
Fede (24:26)
but they generate content kind of 24 seven and they have live stream, I would say at least eight hours a day. So when, when you do that, you generate a lot of content and yeah, he's on YouTube and you can rely on it. But at the same time, when you need to, when you need to find it, they do actually real manual work. They ask some guy in the, in the newsroom, Hey, do you remember when someone say this in this TV program? And there is two guys that normally act as
Fryderyk Pryjma (24:38)
Mm-hmm.
Mm-hmm.
Fede (24:55)
know, Google search inside the stream and they tell to others. But hey, if we can automatize this, they will save time and money.
Fryderyk Pryjma (25:03)
Okay, so basically in here it's not really about finding some new features or functionalities but just more convincing them of like this product fit. Okay.
Fede (25:15)
At the moment,
yes, because they need to use and validate the MVP. As soon as they use it, we started last week. So as soon as they use it, let's say three months, two months, we will have real feedback of the day to day use. They might ask, okay, we need a new button here. We need this button to be on the left. We need this functionality that is not there. We find out this error on the prompt. Whatever it is, we need to evaluate and take care of and put on the roadmap. Yeah.
Fryderyk Pryjma (25:19)
Okay.
Yeah.
Hmm? Okay.
Mm.
Okay.
So would my summarize of this call be appropriate of this podcast here? So you basically don't believe in creating a roadmap without testing what you already have your prototype with real users. So basically you don't have some features aligned coming up next because you didn't get them yet from the users that would actually be interested in also
paying for that. So they really believe in that helping them. And your whole idea here and concept of developing Sofutu and those other products within this family of products is aligned with feedback of the users. Am I right here or did I misunderstood you? OK.
Fede (26:26)
You're totally right. I mean, if there is no market for it, it's the same,
you know, that that's proverb that says, hey, if there are three goes down on the forest, but no one listen it is making noise or not.
It's the same. mean, are we going to develop something that no one is going to use? What for? So you try to find this niche from the early beginning and the user from the early beginning, ideally on paper before you code anything or you draw anything digitally to get immediate feedback on the real user needs. Let's put one more example. Steve Jobs, great creator, visionary guy. ⁓ No one can question his legacy, no matter if you like Apple products or not. And even when he
Fryderyk Pryjma (26:52)
Yeah.
Hmm.
Fede (27:08)
his team created the iPhone. The first I read his book and one of the first thing they realized is we are using phones that we don't like. Let's create a phone that we need. So they were the users who tested the from the early beginning this and then probably they run 1000 focus groups because Apple has the capacity to more or less do whatever they want. And of course they have an organization and a structure to do it. And then they take the best out of it. And when you see the user interaction in an Apple product,
Fryderyk Pryjma (27:16)
Mm-hmm.
Mm.
Fede (27:37)
is so polished that you realize how they work from the early beginning with this on their mind. actually, I also like the analogy of Ohan razor. This is something coming from long time ago. But, you know, it has the idea that the razor on your skin take out anything that is not needed. So you go to the core functionality, to the minimalism that is so popular in Japan right now and the Asian culture.
Fryderyk Pryjma (27:38)
Mm.
Yeah. Hmm. Yeah. Hmm.
Fede (28:03)
So you really focus on what the people needs, then probably you have more chances to be successful. There may be exceptions to the rule, that's correct, but if you follow the rule on this path, I think it's more efficient for everyone, for your time and the other people's time.
Fryderyk Pryjma (28:06)
Yeah.
Yeah.
I really like this thought, especially there was also a quote by AB Optimizely, I believe, founder. I don't recall his name, but he was working previously for Booking and some other quite big platforms. what they started doing in Booking was running 10,000 tests at the same time of different sorts of functionalities like AB tests.
and comparing multivariant tests, comparing like one to another and seeing what actually works, what gets better feedback. it's sort of aligned also with those algorithms of X or like Instagram that they first run tests on a couple of your closest connections or LinkedIn. And then if the reception is good of the content, if people interact with it, if they like it or whatever, they will get boost. So everywhere we see this testing and
I really like when you as a founder of the startup talked so much about the user and idea here. Yeah, that was very nice talking to you, Fede. And thank you for that. Yeah.
Fede (29:20)
Pleasure is mine, Frederick. Nice and warm salutation to all your audience. think it's going to be a great podcast. So hope we engage soon and we continue the discussion because it's a...
let's say, it's something very productive to discuss tech from different angles, also for people which is not techie, but you can apply those principles in your personal life or in your family life or in your friends life and you're going to get much better results. So thank you.
Fryderyk Pryjma (29:47)
Thanks
for that.