Alex
00:00:00.000
that's why I'm using the word balance. because, it's not as easy as just go buy everything, the stellar business is the balance, everywhere, including this, buy versus build, dilemma
Darin
00:01:15.501
You may be sitting around some days and you're thinking to yourself, wow, it seems like I could. Just go and build this myself. How long could it take? Viktor? Have you ever thought that to yourself?
Viktor
00:01:28.520
Oh, every second day. But the answer to that question depends on what that something is.
Darin
00:01:33.936
Now, let me give you an, a raw example. a bookshelf. I could go and build a bookshelf or I could go and buy a bookshelf.
Viktor
00:01:45.945
In my case, there is only one option. I cannot even change the light bulb in my apartment.
Darin
00:01:57.128
Good. Uh, my example was probably a bit contrived, especially since we live in the technology world, but would you ever go and build a bookcase or would you buy a bookcase?
Alex
00:02:06.471
like 20 years ago, I built, an, uh, electronic library, I mean, for the eBooks. during that time there were no clouds, there were no, services like upload care. So we build a lot ourselves. And, spend a lot of resources, focusing on not on the product. but it was also happy times too.
Darin
00:02:28.830
so you were an overachiever, you built an electronic bookshelf instead of just building a regular bookshelf, so, okay. You were, you were just overachieving. That's, that's good to hear. Boy, you, you really took it and ran with it. I'm, I'm really thankful for that. You said a couple things there though that were interesting. there was more to it than you thought. I'm using my words, not your words. Did that really catch you off guard because you didn't have something else that you could use? You had to build it all from scratch.
Alex
00:02:55.658
Yeah, so basically, at that time there were no quality, reader for mobile devices. And, we like really wanted to have a very nice, experience of reading, because we like books a lot of things that are automated today, we did ourselves.
Viktor
00:03:27.315
I dare anybody to tell me, to find me the case where somebody only built or only bought something. Let me give you an example, right? And when I say buy, I mean use existing product, right? You use Linux on your servers. That would be buy, but you always build something on top of it. You always install something on top of it. You use Kubernetes. That would be buy. Uh, no matter whether it's managed service or no, and then you're going to install some stuff and configure it in what's or not, right? So it's always both. The question is whether, whether you put 30% into buy an 80% into build or vice versa. I never saw anybody buying a solution and saying, okay, this is it. Kind of like, I'm not going to tweak it. I'm not going to configure it. I'm not going to do anything with it. I'm not going to put anything on top of it. I'm just going to use it.
Alex
00:04:21.852
Definitely. I completely agree. I just wanted to say that, currently it's so much easier to build a business because, it's easier to buy, so much, like pieces of it, to focus on, uh, what's needed. But it's definitely a balance between buying and, uh, build. my opinion is, How do you maintain this balance is how, you run the business?
Darin
00:04:44.943
Well, but running the business shouldn't really make any difference in how I build my business. Should it?
Alex
00:04:52.086
For me, it's like a nature of, business. Like today's business in the internet. you should pick the game, that you want to play and focus and buy everything else, that you are not building.
Darin
00:05:06.670
I understand your argument. I semi agree with it. I don't fully agree with it because I picked the game, but the problem is, is the rules keep changing of the game.
Viktor
00:05:17.707
But it's also, the business you're having is going to affect what you can and cannot buy. at certain level, right? You can say, Hey, uh, my business is this application that is designed to run in WebSphere. Can I buy Kubernetes managed service? Well, no, because that thing does not run there, so it. I agree with you, Alex, that you should be buying everything that is not your business. It's just that it's not such a clear cut, because many, much of the things that is your business, or more importantly, things that are not your business, but you already somehow have and dictate some other things end up messing up the what should be a relatively easy solution, right?
Alex
00:06:02.240
that's why I'm using the word balance. because, it's not as easy as just go buy everything, the stellar business is the balance, everywhere, including this, buy versus build, dilemma
Darin
00:06:16.457
Well, if we're trying to work through that dilemma, where do we start? I mean, we're talking about, you know, focusing on the things that are differentiated, like build those things and buy everything else. Ignoring the, okay, email services and internet service, all those things. We got that. And where should we, if we're building a business, where should we today consider breaking off between the two? let me give a hypothetical example. Let's say that I am building a AI startup to help developers write code. What would I build and what would I buy?
Alex
00:06:50.042
it's always, a dilemma about the resources, when people. decide to build something. They see a very, clear path to building and shipping it. but, it's not the, whole picture because, you will need to maintain it. you will need to, update what you, built. you need to run compliance, and, Keep everything clean. So, I heard, the other week that, some, uh, CEO of some companies said that we've, built, JIRA in two weeks, with ai and we are going to teach Jira because we are like just built it. that's, actually what, Currently happens with the AI startups or startups, with the help of ai, but, I think it's, can be, shortsighted in, some cases because first you need resources to build something and you can underestimate them. And second, as a business owner or CTO or some manager, you easily, forget about maintenance costs, usually because it's, very hard to estimate them if you have, no experience with dealing, with, self-built,
Viktor
00:08:01.350
In atop that the. Ridiculous thinking behind that example is, okay, so if you're unhappy with the price, the, the money you're giving to Atlassian for Jira, right? And you want to get away from that, why still? Why build it if there are dozens of open source projects that will give you better solution than what you are about to build anyways.
Alex
00:08:26.898
it's a very attractive idea, idea for many, managers, that their company has unique, management, structure and, unique processes. there is no tool in the world that can, support their processes. So they need to build, their own, process management and task management, service. I think it's because, like the feeling of uniqueness of, themselves that, keeps them, wanting to do this. So one of examples this person said, uh, is that we will have Kanban boards finally, I'm not a fan of Jira two, but I just see what will happen to this project in, uh, half of a year in a year.
Viktor
00:09:12.496
That's so, it's so funny. There is a person who says, we need Kaban, which is supported by Jira, but we don't want to pay to Jira and we need to build it ourselves because nobody else in this whole wild world came up with idea to have a ticketing system for free that includes company. We invented Kaban Kind. It's our thing. Nobody else used Kaban before us.
Alex
00:09:48.661
yeah, don't want to blame, uh, him or her so much. but yeah, I know the feeling because I was there, when I was younger.
Viktor
00:09:57.515
know when? When companies think like that, they better be Google. Or similar company because there there is a very few number of companies on this planet that are so big and facing challenges that nobody else faced. Right? Because you can say Google. Yeah, we need to build our stuff because actually we are doing things that nobody else is doing yet. The rest of the industry will be doing it in a few years. But then you have everybody else, like 99% of us kind of, we are not doing anything that hasn't been done before. And if we are, that's because we are silly. That's the only explanation because vast majority of people is not on a Google level or Netflix level or what not.
Alex
00:10:44.501
Yes, this, this is a very good point because, it's also a part of the answer of, uh, initial question about, what do you need to consider when you build versus buy the size of your company, is one of those, resources, to consider. there can be definitely cases that, Company will need to build a lot. but, because of this, there is no universal formula of, um, making the decision.
Darin
00:11:10.909
here's the formula, I think, or at least one of the formulas. You gotta have money. If you're gonna build, this is the build conversation. You gotta have money and you have to have people. If you don't have both, actually flip it around. You're either, this is like those, the funny meme curves, right? You ha you have the, the, the tails on the, in the big bubble in the middle. I have no money, I have no people I've gotta build. I've got a ton of money and I've got a ton of people I probably need to build. It's in that middle to where. Buy is the sweet spot. How far off am I on that?
Alex
00:11:44.708
Wow. I never thought about it like this and it's very interesting. But, The first case when, uh, you have no money and you have no people and you are building something. I agree. this is so very risky of course, but, this is how it usually starts. But, many of, if we are, speaking about software, not bookshelves, there are three plans for that, for this specific spot. but other than that, I agree with the middle and, um, the sweet spot. I think this is, uh, where, I want to challenge, the big companies because even, uh, the big companies with a lot of stuff and a lot of money, they desire to build something specifically for them sometimes, just, result of not, built here, like syndrome, and not the real, uh, necessary.
Viktor
00:12:39.039
the talk about the size of the companies in number of engineers, that's completely wrong, completely false, That gives big enterprises that live in a previous century, excuse to think that they're doing something amazing that has never been done before. Right. You take a bank, oh, we have 50,000 people working on this thing, right? And as if that matters, as if that is a reference to, Hey, we built, we have 50,000 people building, so we must need more unique solution than thousand people that happen to work for Google. Again, not true because the number of people is not the reference of how advanced you are. More often than not, it is opposite reference. More often than not, actually great things are done by small number of people. If you take, like today on Tropic for example, that's not a huge Theme. That's by all reference said, best case, medium sized company, probably even considered small, they're building more advanced stuff that do not that, and they cannot rely on existing solutions because existing solutions do not exist for that something. And they're still infinitely smaller than, uh, even an average bank, not to mention the big banks. So the number of people that's, that's such a fallacy is, is, is a point of reference for anything
Alex
00:14:00.378
Yeah, and usually as you said, it's very hard to maintain efficiency in, high populated companies. this is can be a reason why they think they are unique, and need to buy something for themselves, maybe to keep themselves busy too.
Darin
00:14:17.694
Guess we always love CV driven development. That's the, the best development there is out there. I'm questioning though. We're talking about full build or buy and you were talking about not invented here syndrome. A lot of times I think the not invented here kicks in because people don't know that there are options out on the market.
Viktor
00:14:39.177
That's the example of the Kanban story from the very beginning, right? Kind. Oh, but I need Kanban. I just came up with that idea.
Alex
00:14:46.966
this example too because like two years ago I didn't know that companies, like upload care even exist. I, started my career 20 years ago. There were no clouds. no one can, uh, uh, help you to upload file, store them. and up until two years ago, I never knew there is a, solutions to simplify this, part of building, your, business. So. This is definitely, a good point. we live in such, high ratio of, noise versus signal. and, uh, it's very hard for me at that point, even to think about that someone can help me with this, problem.
Darin
00:15:28.967
Well, let's stick with the upload care. We, we won't go down into the all things about upload care, but let's think about it. You know, you're uploading files, right? At a high level that's, forgive me, everybody. Upload care and public relations. The core part of your business is uploading files, correct? At the end of the day, where it's an easy way to upload files, true or false?
Darin
00:15:53.887
Okay. Fair enough. Right. And that's and the perception, but you're, you're hitting on the point that I'm going for and we're talking about signal versus noise. Marketing departments are really great at cranking out a lot of signal, excuse me, the exact opposite. Great at cranking out lots of noise and not a lot of signal. Because if I'm thinking, and I'm a Java Java developer. And I work in the Spring Boot ecosystem, it's like, oh, that's this thing that I use and it's not a big deal. But then I don't understand. It's like, oh, I just need to upload a file. But what I don't realize is, and without pulling everything back right now, that there's all these other things that live in the ecosystem of uploading a file that upload care that probably other people have solved.
Darin
00:16:45.736
because, and this is where these point solution companies, much less, oh, upload here, they've gone not just wide, they've gone deep in doing whatever the thing is that they're doing now. Would I ever build a Kanban company? Yeah. Okay. Whatever. That's, that's too easy to laugh about, I guess. But. How can companies such as upload care, such as other companies that are focused on a one really tight niche thing, how can they let people know about themselves as an option? Because I think that's the key point, because much like you, you said two years ago, you didn't even know who upload care was. Now you're the CTO.
Darin
00:17:29.304
Until I was reached out by y'all, I had never heard of upload care. So what can companies like that or, or if you're, I mean, you're beyond a startup. I do know that much about upload care. But if you're a startup or in that early stage, how do you let people know with quality signal and not just a bunch of noise?
Alex
00:17:49.434
Yeah, in my case, they sent me an offer. This is how I, knew about the upload care. but, for more practical, ways of marketing, if you ask about like specific ways of how to promote, SaaS company, like, applaud here.
Darin
00:18:04.701
Yes. Something like that. I mean, and let's go beyond the noise because ads, I use ad blockers, so I don't see ads. how can a company today really get out?
Alex
00:18:16.145
Ah, okay. Okay. Okay. Understand. ads are actually working a bit, when, uh, targeted, the right way, but, I don't know much about it. And actually, as a CTO, the promotion and, uh, showing out is not like my strongest part. And I, don't have a lot of experience, about it. But The highest signal marketing that, we had in our history were, Hacker News posts. a bit of red and product hunt, I believe.
Alex
00:18:49.566
Yeah. But, product hunt, became, very AI populated area, where bots, talking with bots, but still, there are so. There is no signal versus noise problem. There is a problem on, receivers and, not much, a live, people, there, but I think, they are, working on it to fixing. So, we get some, audience there, but, For upload care. It's easy and hard to go, signal versus noise because we can't go, very noisy because we are fixing, a very specific problem and, we, should target people with those needs and we can go full buzz about, uh, ai. Well, again, we can do a bit, I using AI and, uh, implementing, in new like AI products, but as you said, applaud care starts with applaud and, most of people, come to us for, this, because of this specific reason.
Darin
00:19:48.431
it's interesting that you have gone through the big three, and one thing you were saying there is you know who to target. You know the companies that are your target market. So to me, your best signal is to reach out to those companies directly. Now, the question is, again, as CTO, you may or may not know the answer to this. It's like, how do you find the right people to talk to them at the right time? That has to be like nearly impossible, a good part of the time.
Alex
00:20:20.783
Yeah, I am not usually starting the conversations. but, believe it or not, usually, the problem is not like how to reach the, person in the company, but how to reach the correct person in the company because, I was surprised. but Yes, companies are, reaching out to us. and, I'm not the sales person or marketing person, but from time to time I'm on these, lead meetings or, prospect meetings. and, currently I believe that our funnel is, organic from, people who are finding out, about us. But, there is, also a competition and some companies, want to change the vendor sometimes, although it's very hard in, this industry sometimes, companies who already, using such services want to find alternatives and reach out to.
Darin
00:21:10.908
vendor Change is always one of those ones that. Makes sense from the user perspective. It's like, I've gotten fed up with Product X, so I'm going to go work with product Y, or Product X is no longer covering all of my needs because my needs have changed, so I need to go find product Y. But you're sitting in the prospect meetings, what are you hearing from those prospects of. Okay. We're trying to buy. We don't want to build or is it We'd prefer to build, but if you can do it cheap enough, we may buy What are you hearing?
Alex
00:21:45.166
Oh, that's very interesting question. And I was like, a lot of insights for me by, visiting those meetings. again, they are different. companies and some of companies, do not consider building at all. But, there is a class, of companies who consider built versus buy and, they are, concerned mostly about the Complexity of integration, because, I believe that applied care solves a lot of small, problems at once. that's why they have this, dilemma and why they are, considering. So usually I hear. we want to build, but let us do POC and come back to you.
Darin
00:22:27.060
And then what do they sometimes see? So you, they come back from the POC and they're going, yeah, don't worry about it. We can take care of this. Or what do they find that it's like, oh no, that's actually, you're doing something that we hadn't thought about doing.
Alex
00:22:42.712
it can sound like bragging, but I will make it to sound like self criticism. so when they come back and say that, oh, well, we actually didn't expect that, integration will be so easy. I blame myself for not, correctly, like, explaining them, uh, at first that it'll be easy. So I need to improve my English and I need to improve my speech. But as a CTO, it's okay, but. I need to do better. That's what I, got from those meetings.
Darin
00:23:11.707
Well, the thing that you have to be careful of is not to create a bunch of noise from that point forward, because you don't wanna say, oh, five seconds. You can have this going. You know, maybe be a bit self-deprecating, but it's like, okay, it won't be five seconds, but you should be able to do it in 10 minutes for first pass, or whatever your real numbers are. Just be careful.
Alex
00:23:33.578
again, I, my self critics is very strong to say things like that, but, I can, uh, provide an example. So, I sometimes, uh, showcase how to build Instagram competitor, with upload care. it's like not real competitor of course, but the thing that has the same interface, not scalable, it'll not, make you money. Just the interface. As a bookend developer, I can do it in 20 minutes, I believe, to do the one page thing. But, again, 20 minutes doesn't mean that you can do it for your production in 20 minutes. but it's, uh, of course way easier than to implement. Accepting file from, user, sending it to server, receiving it to server, putting it in right place, setting up the backup, transforming and delivery via. And, it can be also, be built manually.
Darin
00:24:23.175
Well, all those points you just called out there, we heard upload file, we heard, put it in the right place. And for most of us, if we're thinking about, again, sticking with the uploading a file example, that's probably all we think about. But then you went on to say backup and a few other things. Again, point solution companies such as upload care have a lot to bring to the table and. You're hiding that complexity of sort of the integration, you're hiding the complexity because people weren't even thinking about all those other things that really should be covered. How can a company, let's go about higher level now. How can a company help? Potential clients, or let's say you're an open source project, how can you help consumers of your project really understand all of the true features or things you should be thinking about if you're trying to solve this problem, whatever the problem is.
Alex
00:25:23.344
you can, uh, create a lot of noise. Uh, not that, we are doing, but, you are completely right that, most people are sticking to us because of some specific feature. Some of them like. Maybe the majority, I should check stats, uh, sticking because of the uploading, interfacing functionality. But, for some of them, delivery is very important because setting up, CDN is not, easy tool. I, created. Probably S3 bucket, 200 times, by my hand. and didn't, do it flawlessly once, it was also some problems with i Amal or something like, that same with the CDN. Um, but with blood care, it's just, plug and play. but again, I'm not answering your question. Your question was, how to sell all of the features that we do to customers. So we are not doing this. we, trying to sell only what's, makes sense for our customers and, our whole tool set is needed only by a small fraction. So we can convert videos, documents, uh, sometimes uh, people just need to, accept or deliver a file, and we are, completely okay with it.
Darin
00:26:34.808
Go a high level because you've probably dealt with this, if you're making the build versus buy decision as a company and you're going to be an international company, and now you're thinking, okay, I'm, I'm gonna build this, I'm gonna buy this. now, because you're international, you gotta deal with GDPR, you have to deal with all sorts of other. Multi character rules and regulations that you've never heard of in your life. If you're going to build that, it could be a bit challenging 'cause you're gonna have to hire consultants to figure out how to do all this, or you potentially go buy. If you weren't at upload care today as the CTO, what would you probably be thinking? What would be your first thing of, okay, we have this new feature, it's gonna be global. And we have to fit into all 170 countries or however many countries there are today. What would you be thinking or would you be thinking, I need to update my CV and get out of here?
Alex
00:27:34.596
there are some requirements for our vendors that we can, we should, have. So as a SOC two, type two, certified company, we should pick vendor with, the same level of, compliance. This is actually like my job and part of my cv, of, picking the vendor, for, international, company. the funny thing that, GDPR and HIPAA sometimes conflict with each other, one, uh, forces you to remove data after certain period and other, forces you to store it for the, certain period of time. this is a, part of my job, basically.
Darin
00:28:09.635
But how do you solve it? I mean, you. Just brought up the conflict between GDPR and hipaa. I mean, what are you going to do?
Alex
00:28:15.415
if a company needs to store or process some, sensitive data, Medical data, sorry, because HIPAA is, about medical data and, GDPR also targets medical data, specifically. So if you are, going to do this, you need to implement, or buy some solution that can, manipulate with data respectively to the, origin of the customer.
Darin
00:28:36.908
But that's the key point there, is you either have to build it or you're gonna have to buy it. Now, the chances of building it yourself as a company is pretty small. You could do it, but it would take time. You'd have to find the right people versus if you went with a solution that already covered all of those things and kept you in compliance, which is going to be cheaper, you can get it to market faster. In theory, you're not having to buy the knowledge of people. You already have bought the knowledge of potentially hundreds of people through doing the buy.
Alex
00:29:08.810
again, back to our initial, uh, question, upload care I think is a small company. when we are considering build wheels versus buy, we are considering our, annual roadmap. And, how can it look like in the end of the year? What can we achieve? can we, build, another, solution to handle. Data somehow, or we can ship a user facing feature. at this point we are making these, decisions. we can spend year on building something for our infrastructure and no person in the world will feel it. Or we can, spend resources on something, meaningful for user. Usually we don't pick any of this. We are finding balance. So something are going to be built. But, we also need to focus on, our product.
Darin
00:30:00.099
Own more route here. If you're doing build versus buy, you're probably not thinking about ongoing maintenance. Or maybe you are right as if you're building something, you're Java, your go, whatever. You're gonna have to update versions. Well, you should update versions, let's put it that way. Whereas if you're buying, hopefully. Your vendor is taking care of all of those upgrades because they're trying to keep their SOC two compliance in place. They're trying to keep ISO 27 0 0 1 in place. They're trying to keep all the things in place so they have reasons to keep all of the software updated. Uh, can you think of other things around the, sort of the maintenance nightmares that, because I'm assuming you have maintenance nightmares as the vendor.
Alex
00:30:44.937
all my life I believe is a maintenance nightmare. this is my job To speak up when someone is, considered to be built or bought. so what will be the consequences? Because I was, in a lot of, Or very harsh situation like, burning hardware rate controllers, or criminalist attack to data center or. I dunno, failing Postgres database because of, the, transaction wraparound. as a, highly sensitive, anxious person. when I even hearing something about, let's introduce this new, tool to our infrastructure, or let's build this, or let's buy this. Uh, I'm not thinking about the features and, possibilities that will open. I only, think about how can it break, can we afford it to break? can we afford, to maintain it because, Doing something yourself, usually requires tools. That means that you also need not only spend resources on maintenance, but you need to maintain the level of knowledge of this, tool set in your company and the job market also dictates which tools are easier to maintain or not. So the formula in my head is very complicated. I can't put it on the paper, but a lot of variables.
Darin
00:32:05.851
A lot of variables. And even if you're using a vendor, you still have maintenance because the libraries to integrate with that vendor probably upgrade over time as well. So just because you buy, you're not off the hook of maintaining.
Alex
00:32:18.287
Absolutely. And, uh, some, vendors are better than others. And, uh. Can like notify you in advance about, uh, the deprecated versions, of the libraries, for example, or APIs. But, another, can, um, like deprecate some API in, one week. So reputation of the vendor is also important, and, checking not only the compliance side, but the actual engineering culture. inside the vendor, is also important when, actually picking vendor,
Darin
00:32:49.350
Now I've poked a little bit at upload care and I asked for forgiveness now, but upload care does a lot more than just do file uploading. What does it do? I mean, why, why would I choose up something like upload care?
Alex
00:33:02.492
oh, I have favorite features, that is hard for me to do it myself. So again, when I joined, I didn't, understand why it's needed. but now I have an answer. my favorite features of, uh, plot care is, after quality and after format for images. so imagine I'm building my block or site and I am, uh, uploading, images for it, but, upload here and. And I am putting this image, tag on my page with the link to the file on upload care and upload care automatically, serves, Aviv or VP or jpeg uh, version, of the image based on the browser, of the end user. So no programming, it's just a tech, uh, it's all works through our CDN. The different versions of the file are cached, individually, processed individually, basically all modern, uh, devices received AV version of the file, which is significantly lighter of the same quality in jpeg. so the website loads much faster, and, it works automatically. I can build it myself. on my, like VPS, but it won't be so elegant, because a lot of things are, happening on the CDN level to build it yourself really take, a lot of time. And the second option is automatic quality, which will serve this image with the minimum. possible quality, that is not, adding artifact or an image. So the image automatically gets smaller, in quality, without affecting visual quality and, format, which makes it even smaller and the website even faster. I didn't know it's possible I didn't think about, Such a trick, but now for me it's like, must have. this is, uh, my, uh, point of view, like engineering point of view. We have a lot of AI features, Recently, we, uh, introduced, on the fly video, transformations To the HLS format. So when you, for example, upload video to YouTube, you need to wait several minutes or hours until it'll process to all of the resolutions. But in upload care, you just drop a file and it gets, transformation on the fly and can be streamed immediately, with no, delay. I can talk about it, uh, a lot, but my favorite, after format and after quality for images.
Darin
00:35:31.728
Very cool. Well, upload Care can be found@uploadcare.com. All of Alex's information will be down in the episode description. Alex, thanks for coming on today.