00:29:38 · Matteo Danieli
[00:29:38] S1: Yeah. So we currently have in our portfolio dozens of products. So it would be pointless to list them all, but I could mention some examples, perhaps the most important ones, and go through some of those and try to find these commonalities if they're out there. So in no particular order, WeTransfer, file transfer, file sharing platform with more than 70,000,000, seven-zero, monthly active users. It's basically the de facto standard in Europe, in Europe especially when it comes to sharing large files. Then we have Evernote, a digital notebook with a lot of complex features and a large degree of customization. And there we have more than 1,000,000 paying customers. The latest acquisition is Commute. So it's an app and website where you can plan your outdoor activities, being some hikes or bike rides. And like there's a great community, a lot of user generated content, and there's more than 8,000,000 people using it every month. Once again, it's, I think, a point of reference for this type of things when it comes to Europe, especially. We have Meetup. So it's platform where organizers can organize events and people can join them. And these events are organized around shared interests. And there as well, every month more than 300,000 events are organized. So again, massive scale. Of course, we have Remini, an AI powered photo enhancer that also gives you the ability just to create photos from scratch through generative AI. That's one of our largest products when it comes to user base with more than 100,000,000 users around the world.
00:31:37 · Luca Ferrari, Matteo Danieli, Roman Kirsch
[00:31:37] S2: It's probably there is no official ranking, one of the top most alternative AI products in any category globally, probably top five. It's massive, massive usage. So I think you At some point, sorry, some point we were using so many GPUs for all the inference that was on Google Cloud. We became a kind of an emergency case at Google to help us scale fast enough, just to give a sense of the insane level of usage. [00:32:09] S1: And at peak times, we used to employ something like from 30,000 to 50,000 GPUs. And when I talk to tech people around this, they look at me and they're like, you're either making it up or you got the number wrong. Because that's open AI levels. [00:32:26] S0: Sounds expensive as well. [00:32:27] S1: Sounds expensive [00:32:27] S2: as well. [00:32:28] S1: But we were able in Remedy to nail the right approach to the business model, whereas we would be able to offer AI powered features for free, of course, with some limitations. And we would be, I think, the only consumer app that offered you this ability apart from OpenAI Especially in image generation. In image generation, [00:32:52] S2: Not actual, that's a little bit cheaper. [00:32:53] S1: Which is more expensive. And the way we did that was that we nailed the business model to make this possible so as to have paying customers subsidize the others. And at the same time, we worked a lot on reducing the cost of inference, like making crazy optimizations so as to gain cents and cents to reduce costs and to be able to offer these at a much lower cost. So if you're
00:32:53 · Matteo Danieli, Luca Ferrari
[00:32:53] S1: Which is more expensive. And the way we did that was that we nailed the business model to make this possible so as to have paying customers subsidize the others. And at the same time, we worked a lot on reducing the cost of inference, like making crazy optimizations so as to gain cents and cents to reduce costs and to be able to offer these at a much lower cost. So if you're [00:33:18] S2: able It's very interesting offer engineering problem, by the way. Very fascinating. [00:33:21] S1: If you're able to offer these features that went viral a lot of times for free to this massive user base, out of that comes an insane consumption as well. But luckily, the ROI is positive, we can afford that. Yeah, that was an insane ride. But anyways, so if you try to look for commonalities
00:33:21 · Matteo Danieli, Luca Ferrari
[00:33:21] S1: If you're able to offer these features that went viral a lot of times for free to this massive user base, out of that comes an insane consumption as well. But luckily, the ROI is positive, we can afford that. Yeah, that was an insane ride. But anyways, so if you try to look for commonalities [00:33:42] S2: Can I add one more, I think? Because it's really quite so we also own Brightcove, which was a publicly traded company on Nasdaq. We do it private actually only a few months ago. Fully enterprise video platform, very sophisticated video technology for transcoding, streaming used by massive clients. It's only 2,000 customers, quote unquote only, but the average ticket is almost $100,000 So the biggest names you could imagine I can't name them but you could pretty much all the big companies that need video use it. And that's entirely sales driven, relatively few massive customers. So we really span the gamut from Remini, which is over 100,000,000 people each paying nothing or a relatively small amount of money, maybe $100 a year, to Brightcove, which is relatively small number of enterprise customers each paying $50,000 $100,000 even some a few million dollars a year in big contracts.
00:33:42 · Luca Ferrari, Roman Kirsch
[00:33:42] S2: Can I add one more, I think? Because it's really quite so we also own Brightcove, which was a publicly traded company on Nasdaq. We do it private actually only a few months ago. Fully enterprise video platform, very sophisticated video technology for transcoding, streaming used by massive clients. It's only 2,000 customers, quote unquote only, but the average ticket is almost $100,000 So the biggest names you could imagine I can't name them but you could pretty much all the big companies that need video use it. And that's entirely sales driven, relatively few massive customers. So we really span the gamut from Remini, which is over 100,000,000 people each paying nothing or a relatively small amount of money, maybe $100 a year, to Brightcove, which is relatively small number of enterprise customers each paying $50,000 $100,000 even some a few million dollars a year in big contracts. [00:34:46] S0: Okay, and then so those are two great examples because they're so opposed to each other, right, in terms of just the ARPU and the technical problems that you have, the level of service that you provide and the way how you think about building those companies. So how much of your playbook is actually fixed and how much do you kind of create a new playbook post acquisition for every single company? And what did you learn over the past acquisitions?
00:35:14 · Matteo Danieli
[00:35:14] S1: Yeah, so Well, go. Yeah, would say that, so for us, it's easier, of course, to tackle products that have similarities, either in business model or type of customer big products that we already have in our portfolio. And there perhaps we already created some playbooks that are perfect for that. But we to really find a good trade off between doing some of those acquisitions where we're pretty sure about what we're going to do for many months from when we acquired the product, but also dive into completely new territory where some of the things that we've learned we know we can reapply. But we also know that a lot of it will be kind of learning. And by doing that, we just expand our playbook. And then we also expand our ability to know precisely what we will be doing if we encounter another target like that. So I think it's being this trade off between exploitation of the know how we created until that point, but also exploration in order to expand that know how. And this is how we are able to sort of target and think about acquisitions in fields that we hadn't tackled before. But as a consequence of this, it's really hard to find a common theme. So if anything, as Luke said, we really span the gamut from mainstream B2C to enterprise. If I were to find perhaps one commonality, Luke, correct me if I'm wrong, but I would say at the level that we are now, it's either kind of established brands recognized typically reference points in their industry, in their vertical. Some of them might be a little more niche. But within their vertical, they're very much well known.