#123 Why teams are turning to warehouse-native analytics (with István Mészáros at Mitzu)

[00:00:00] Dara: Hello and welcome back to The Measure Pod. I’m Dara, joined as always by Matthew. Hey Matthew, how are you doing? 

[00:00:20] Matt: Yeah, I’m alright. Thank you. Yeah, how are you? 

[00:00:21] Dara: Yeah, I’m good. I’m looking forward to diving into more exciting news and, and a good topic. So we’ll kick off with the news. so we’ve got a few, few 

[00:00:31] Matt: bits to go through.

[00:00:33] Matt: The first thing I had was a lot of people probably would’ve noticed things like Spotify and a few other big services going down, last week on the 12th of June, I believe, which was all related to a Google Cloud outage. Because naughty Google didn’t stick to their stringent policies of releasing and testing features.

[00:00:58] Matt: So from what I can see, there was some, some new feature they released that didn’t go through the normal checks on their internal systems and had the right tags and flags on it, and it kind of propagated through all the various different regions and caused massive outages. So yeah, Google Cloud went down.

[00:01:16] Dara: A bit of a rookie mistake really. But you wouldn’t find us making mistakes like that, would you? 

[00:01:21] Matt: No. Especially when it’s all the time about security and all of those kinds of things. So yeah, a bit of a drop ball from Google there. Do as we say, say, and that as far as I’m aware, yeah, as far as I’m aware it’s all back up and running.

[00:01:35] Matt: but just make sure you have backed up your data and have it sitting across multiple regions to deal with Google’s problems when they, when they drop the ball in this way. I think a couple of years ago, right on the right smack banger around Black Friday, the US Central one. Region for App Engine went down and everyone who Google had just been saying move to service side, GGM, move to service-side, GTM, they all lost.

[00:02:04] Matt: They had a big black hole on their Black Friday day. Yeah, because they were all in a single region in the US Central Wall and it went down. So yeah, there are ways you can, you can wear ways, you can mitigate it by making sure you’re not just reliant on one region, one zone in the cloud, but yeah. Yeah, keep an eye on it.

[00:02:20] Matt: Anyway, my second piece of news, this is, there’s been a lot of, we found recently at Measurelab, there’s been loads of new Google scary sounding emails coming through about this new feature or that new feature around, oh, you need to update your service accounts, sorry, you need to update your schedule queries to be working on service accounts because of whatever.

[00:02:42] Matt: And they’re, they’re a bit opaque and a bit difficult to sort of pick apart and understand what they’re, they’re talking about, one of which. This week was about GA four and BigQuery, a data transfer service. So essentially saying there was, it just sounded a little bit like they were making changes and you maybe you had to update some settings in the data transfer from GA four to, to BigQuery.

[00:03:05] Matt: So, so essentially they’d released a load of new tables, but they’d released them into preview in the first instance. And then as they kept on working on them, they took them outta preview, tinker, boom in the background. And then they re-released these tables and fixed various books and stuff to general availability.

[00:03:22] Matt: So it may be that you’ve got multiple tables or slightly different name tables and slightly different schemas within them. I think, I think basically it’s not going to affect, if you already had it set up, it’ll just create new tables for the, the setup you previously had and, and fix the issues.

[00:03:37] Matt: You’ve never set it up before. It’s nothing. So it kind of seems like you can pretty much just. Carry on as is. but it may have sounded a bit alarmist and scary in the email when they, when they sent that out earlier in the week. 

[00:03:49] Dara: They always do, don’t they? It’s always fairly like it, it’s always that kind of like red text and I dunno if this one was, but you know, it’s usually like red, big, red, bold text and you need to take action.

[00:04:00] Dara: So actually fine. Yeah. Action required. Yeah. Yeah, so to worry about, there wasn’t too much to worry about. Is there anyone you think who, is there any kinda edge case where you would be caught out by that? Or is it just not likely to be much of a buffer? 

[00:04:15] Matt: I don’t know. Unless maybe you’d, you’d, you’d, you’d set up some sort of reporting on the existing tables perhaps, and then they’ve created new tables and you, and they’re not going to be updated anymore.

[00:04:27] Matt: like some, but apparently the original tables were a bit unusable. So like the landing page table, the landing page dimension was missing. So you probably, people aren’t building reporting off of that. Useless table. But yeah, I suppose it’s just worth just checking what’s there, checking what’s coming through the, with the new tables and making sure that you’re not, you’re not building, reporting off it or something.

[00:04:48] Dara: I personally, I wouldn’t expect to find the landing page dimension, the landing page reports, but that’s just me. No, no. That’s just a combination of everything. Never assumed landing page report. Never assume. 

[00:04:58] Matt: So they had an IO the other week, IO 25, which is like the software conference I believe, and they announced Google search aid which I think has been in preview in some places for a while.

[00:05:11] Matt: And this is different from the summary text that pretty much everyone will see at the top of every Google search. This is then a completely standalone independent cert function. and Google started to, I think, push people there by default in many cases. So be using this AI search feature for Google and it’s like longer, it’s, it’s better for longer form.

[00:05:34] Matt: Questions and, and long form answers. And it, it kind of feels like a bit of a merging of the, in the, like, the chat interface of Gemini and the search interface of Google kind of being melded into one because 

[00:05:47] Dara: I’m trying, the way you said there, like, it’s, it’s kind of a blending of, you know, Gemini, interface and, and search is like, I’m trying to kind of jump ahead and see where like at some point they’re going to become the same, aren’t they?

[00:06:00] Dara: And it’s like, how are they, how are they, what’s the plan there? Is it just going to be that at some point you’ll just be actually using Gemini, without realizing it, you know, or is there a, yeah, I mean, it’s got to be Googled for the two to still be separate. Because Gemini now as you’d like, like all the Ls, they’re all pulling in like live internet results as well, aren’t they?

[00:06:20] Dara: So, yeah, it kind of, yeah, just trying to see where the, you know, is there going to be a distinction in six months or years time or would it, would it just all be the one maybe? I mean, 

[00:06:31] Matt: definitely, it’s definitely the strongest muscle Google can flex. Because they obviously, I suppose people like open AI starting to insert search into their, into their, into their chat interface, was starting to eat Google’s dinner a little bit.

[00:06:45] Matt: And just the very fact that people were asking open AI questions, full stop was eating Google’s dinner a bit. So this is probably some hope of clawing back some of that, some of that search that is starting to leak away to these other platforms. But yeah, I don’t know. I don’t, it’s hard, so hard to tell what a new interface is and, and there’s going to be a lot of site owners and stuff out there.

[00:07:06] Matt: That’s, that say to Google, well, you’re just taking further eyes of our site and just surfacing more of the information on your own Yeah. I suppose to a certain extent that’s a problem for Google, for their advertising. 

[00:07:21] Dara: Yeah. Because I assume that there’s no ads in any of their, in any of the Gemini, inter, you know, any aspect of the Gemini interface at the moment.

[00:07:29] Matt: Not currently. So, yeah. You feel it’s coming. It feels like it has to be one of them’s going to shift, get open, AI, anthropic, Google are going to start to get adverts in responses and that’s going to be the way to monetize it. Yeah. as depressing as that is. 

[00:07:42] Dara: I, I think we, I think we might have raised the, you know, in a previous episode we might have raised the topic of, you know, LLM optimization, like SEO for chat two basically.

[00:07:55] Dara: Yeah. it’s going to be really interesting to see how that pans out in terms of like, how do you actually, you know, both organically and if, there is some kind of paid model in the future, but that’s going to be quite disruptive. You’re going to have to completely shift your focus from optimizing for, I was going to say web is still web, but you know, for an LLM interface rather than through a standard search engine.

[00:08:19] Matt: Yeah. Because I think about what we said on that last. Call it Shopify and OpenAI. That’s right. I keep making these calls. Podcast was, yeah, it was. It was Shopify and OpenAI and we were a bit unclear like, does that mean, how does that work? Is Shopify going to be in OpenAI or, and I think looking at it since you can do the full sort of shopping journey in theory whenever this comes out within OpenAI.

[00:08:44] Matt: So then it is like you’re completely abstracting away. I suppose you’re selling a product, but at the same time, it’s hard to understand who is buying that product, but maybe it actually gives you more information than you could ever possibly think of. Because everyone’s put their hopes and dreams into the chat interface of OpenAI and you’ve got so much dirt on ’em that you can send that back to your people that are buying and they can target you Because you know you’ve got holes in your socks.

[00:09:12] Matt: Something like that.

[00:09:16] Dara: If that’s, if that’s the worst thing they know about me, I’ll be quite happy. 

[00:09:19] Matt: Yeah. I need to stop using underwear related examples. I think last time it was buying underpants and now it’s holes in socks. 

[00:09:25] Dara: Listen, you’re just using your own, your own buying bag. Yeah. As examples. 

[00:09:28] Matt: These are my problems. Yeah. So it’s, and Google did, did did talk about a shopping feature as well where they had like, it’s almost like an assistant that you can say I built into Chrome browser and on your, on your Android devices where you can set up price alerts and it will tell you when things are at a certain price point and then you can just sort of one click buy and it’ll go off and do all that stuff.

[00:09:48] Matt: So it’s again, another layer on top of that shopping journey. But there you go. Interesting times break the New world. Some measure avenues. I’ve done any measure avenues before. Probably, probably have, but yeah, we went to the hack the future event at Google, last week, which was really, really fun. So myself and the guys in my team, so Victor, Katie.

[00:10:11] Matt: Pana, we all went together and it was this load of different teams from loads of different cloud partners around the world and around Europe. There were sort of 110 teams or so, all competing on completing these various tasks on Google Cloud and working with AI and analytics and data and deployment.

[00:10:30] Matt: So it’s a really, really cool stretchy kind of event and prizes for the top won, which we weren’t, we didn’t win the entire thing, but we did come 18th out of 109. So I’m, I’m calling that respectable, a respectable place. So you didn’t win. We’ve made friends along the way, and that’s really, that’s winning now.

[00:10:53] Dara: Eighteen’s amazing. That’s really, really good. and last but not least, on the, on the news front, we’ve got a little bit of a plug, which we don’t do too often, but we’ve got an upcoming webinar that we wanted to promote, for, for our listeners who might be very interested in this. So if you find yourself feeling overwhelmed by the flood of MarTech tools out there, I mean, who isn’t, let’s be honest.

[00:11:17] Dara: but we’re going to be cutting through the clutter and by we, I mean actually Matthew, so you’ll be, you’ll be part of this webinar. We’re going to have a live session on Thursday, the 3rd of July at 1:00 PM, that’s UK time. So Matthew, do you want to give just a little overview of what we’re going to be covering in the webinar?

[00:11:35] Matt: Yeah, so we, we kind of talk, like you say, we talk about the MarTech stack and how ridiculously large that has become and just sort of the various issues that come with having. That many different platforms, that many different products or claiming to have very similar or close and overlapping, problems and solutions.

[00:11:56] Matt: And the fact that with a lot of these things you also get a lot of solutions looking for problems and you pay, you pay for that. so we talk a little bit about that a bit, a bit, a little bit about how cloud computing in our world, like Google Cloud offers a lot of different services and tools to be able to build this stuff out in a bit more of a composable, composable way without having to lean on these, this plethora of, of, SaaS products that are out there.

[00:12:23] Matt: and we talk about a couple of case studies that we’ve built for our clients at, at Measurelab and how yeah, what, what would’ve been the, what would’ve been the, the off the shelf version and how we handled it, with, Google Cloud and what that difference that made to price and, and agility and all those, all those good things.

[00:12:41] Matt: So yeah, it’s going to be. me, myself, and Irene, no, me, me and, Mark worked for Dara’s co-founder, double Act partner in crime. On the 3rd of July, 

[00:12:56] Dara: I will drop a link, a sign up link in the show notes. So yeah, do make sure if you’re interested, you sign up. It’s free to attend. And feel free to share the link with people in your team as well.

[00:13:06] Dara: Anyone who you think might be interested. But again, yeah, it’s Thursday, the 3rd of July, 1:00 PM UK time. Alright, that’s it for the news for this week. Our guest on the show today is a fan from Mitzu. He’s one of the founders. Mitzu is a warehouse native product analytics platform. So we had a really interesting chat with Isan, including the origin of the name if you’re interested in that and the logo.

[00:13:33] Dara: So do stick with the full episodes to, to, to get the answers on that. but we covered things such as, you know, I don’t want to give too much away, but we cover things like, you know, who would need this approach and, and what does it offer, and who it might be suitable for, what are some of the obstacles? But yeah, it was an interesting chat. 

[00:13:51] Matt: yeah, it was, it was a really interesting conversation, particularly as, you know, we’ve been talking a lot recently about how you get your data into data warehouses and get it into a good shape and all, all those good things. And, and this is an addendum to that, a, a a reward you get for having your data sitting in a, in a data warehouse, in good shape and, and, and good structure. So, a really interesting conversation. 

[00:14:13] Dara: Okay. Enjoy. Okay. A very warm welcome to the Measure pod Istvan. It’s great to have you. 

[00:14:20] István: Yeah, thank you for the invitation. It’s nice to be here. 

[00:14:23] Dara: So we, we always, when we’ve got a guest, we always get them to do the job of introducing themselves. Because it tends to go a little bit better than if I tried to do a very bad intro and then you’d correct me on all the things I got wrong.

[00:14:34] Dara: so can you, in as much or as little detail as you want, really, can you just give our listeners, a bit about your background, you know, how, how you got started and what led you to where you are today and what you’re doing in your current role? 

[00:14:48] István: Yes, it’s a pleasure. So I’m currently the CEO and founder of Mitzu, which is a warehouse native product analytics platform.

[00:14:56] István: One of the few, warehouse native product analytics platforms. but originally I had the, I had the luck to be in multiple roles. I was a data engineer. In a few companies. I was a product analyst in a few companies, and I have been also a software engineer for quite a while, like backend and frontend as well.

[00:15:16] István: So I got the full package in my history, and basically this gave me, you know, the opportunity to kind of start a company. That’s what I’m doing now, essentially for the last two years. 

[00:15:27] Dara: So I’m guessing in some of those roles you learned what not to do and that’s what you’ve taken into. It’s often the best experience, isn’t it, when you’ve actually, you know, been on that side of, of things.

[00:15:36] Dara: You know, when you’ve been a product analyst and you’ve been an engineer, you kind of learn maybe what doesn’t tend to go right. And, and, and you’ve got, kind of, gives you good insight into what to do and you’re building a platform. 

[00:15:47] István: Yes. Yes, definitely. It’s been a long journey. I did mostly work in e-commerce and travel companies.

[00:15:54] István: I would maybe, I would categorize them as B2C and, the, the number one thing that was kind of felt across all these experiences. If you have too much data, traditional solutions might break down and you need to be clever to, you know, overcome the issues with too much data. And essentially that’s what led me to start this whole business.

[00:16:21] Matt: Do you want to tell us a little bit about Mitzu and, and what that does? 

[00:16:25] István: I think the easiest way to describe to the audience if they are familiar with these tools is that it is a tool that you build, like Amplitude or Mix Funnel, but it is a warehouse native. The warehouse native aspect means that we don’t copy the data from the data warehouse.

[00:16:40] István: We let the client, the customer store, store all their data themself in their data lake or data warehouse, and we just automatically generate s QL rates on top of the data warehouse and, and execute them and get the results and show like a funnel or retention analytics, chart in the tool. So maybe even like a shorter summary, it is like a template or mix funnel, but it writes SQL on top of your own data.

[00:17:06] Matt: And what are the, what, what were the reasons that led you to that, to that starting a company in that space, what kind of pain points were you seeing in the industry that made you think, right, let’s move away from this middleware layer of data storage to Nat na native Warehouse analytics. 

[00:17:24] István: Yes. so it’s quite a long story. As I mentioned, I was working mostly for, you know, trial industry, e-commerce, B2C, B2C SaaS companies. And the recurring pattern was that at one point in the lifecycle of the company, you read, like, a volume of data that you’ve collected. And let’s say you send that volume of data to amplitude, and basically the whole thing becomes unsustainable.

[00:17:53] István: Either the economics of it, just collecting the data to amplitude, is becoming more and more expensive. also you need, after a while in a company, after let’s say a seed or seed, series, a, startup should or typically dust collected to the VA warehouse as well. And basically having your data in two places and having that mass amount of data in, in two or, or even more places is just not feasible.

[00:18:19] István: And during my work, it was a continuous struggle to manage the synchronization between tools. Basically this is what led me to basically start like a Mitzu first as a hobby project, side, side project on the weekends. Then it became open source and then it’s now venture funded, a startup.

[00:18:41] Matt: And I really thought about that in terms of the, because you’re going to have a whole data storage layer that you’re going to have to pay the, you know, your amplitudes, et cetera, that you are, I suppose, swerving around in not having to store the data for them. 

[00:18:57] István: Yes. Yes, exactly. So in very short, it totally makes sense to have, you know, third party analytics and third party customer engagement tools and all these types of third party tools for marketing your, your, you know, analytics and, and everything up until a scale.

[00:19:14] István: And at one point it just starts to break down. Either it is way too expensive to maintain it or it’s impossible to maintain it. And this is basically, we mostly see that some companies hit that level, that it is impossible to maintain data, sync data in sync in multiple tools. And the natural answer to that is, let’s centralize everything to a data lake, data warehouse, and let’s just connect the tools to it and let’s let the tools take care of things like, queries and everything. But let’s not copy all the data just because of some use case that the company has. 

[00:19:53] Dara: So what, just, just to. Well, I, I don’t know if this is going back a step or not, but just to, to, to ask a question around what happens below that threshold. So what, why would, why would it not make sense for a company to use, warehouse native, platform, be, you know, if they’re below, if they’re below that size threshold?

[00:20:13] István: It’s a great question. Based on my experience, it is very, very hard to, or it’s easier, but it’s, it, traditionally it’s hard to get started with data warehousing. You know, you need to have an expertise, a team of data engineers, analysts in place. You need to hire them. It takes a lot of time, these traditional tools, you know, let’s go back to Amplitude.

[00:20:36] István: Actually, just to be fair, I really like Amplitude as a tool. It’s a great inspiration. I, I, nothing against them. It’s, we are not even pro, I, I wouldn’t say, I would say we are not even in the same, you know, market. We’re not competing for the same companies. but anyway, back to your question. It’s to get started, it’s very hard and companies fall back to this, you know, simple tool install.

[00:20:59] István: You put an SDK to your code, you collect your data, and everything’s just out of the blueworks. it’s much easier to get started with, with these tools in a, in the early phase of a startup. Then, let’s say with data warehousing, I would argue, this was actually even more true before, before 2020. In recent years, data warehousing and, you know, even CDPs have started to focus on moving the data to, to data warehouses and, making it easier, as easy as possible.

[00:21:33] István: So, I would argue if a, a, a good data engineer or good engineering team knows about, what are the tools that you can use for moving the data to the data warehouse is the same, you know, same, challenge. The same type of challenge to move to the data warehouse as it would be moving to the, to these external tools. It’s just not yet well understood. 

[00:21:56] Matt: And I suppose you’ve got that advice. We’re Google Cloud partners at Measurelab. And certainly that rings true to us that you see a lot of new sorts of accessibility and democratization tools like the growth of the data transfer service. Then, I mean we, me and Dara did a podcast on next 25, a couple of weeks ago that was all BigQuery, everything.

[00:22:21] Matt: BigQuery. It seems like they’re really trying to centralize so many services on our platform. So I suppose for yourself that’s quite. gratifying and, and, and confirmatory of like, well, that everyone’s putting our eggs in this data warehouse basket. And that’s, that fits nicely with what we’re, we’re trying to do.

[00:22:37] István: Yes. big fair, especially, I think they’re in a very good position. Their pricing model is amazing. It’s very appealing for startups. They have a free tier, you know, up to one terabyte. Then there is like $6 per terabyte of data scan. Super, super good pricing, and the tooling as well is getting better and better with Looker and, and all these things around GA four data.

[00:22:58] István: Now it’s like one click to move to BigQuery. So everybody is essentially set up to have a data warehouse. I would say if, if you’re in a Google Cloud. 

[00:23:07] Matt: Like you say, years ago when, when it was a little less accessible to get your data into the cloud data warehouse, potentially the, the, the clear thing is to go down some off the shelf ETL tool.

[00:23:19] Matt: But now with that pricing tier, if you’re a startup or a small company, and the accessibility to get the data into the warehouse and basically, I mean, for a lot of our clients in their first forays into the warehouse world, it’s free. you got free connections to GA four. It’s getting more and more appetizing to do a composable approach to centralizing warehouse data.

[00:23:43] Matt: That’s Because that’s going to be, and, and you can see a lot of these off the shelf tools scrambling to find their new reality, like at five tram buying centers and a few other things, trying to consolidate and get new features. 

[00:23:55] István: Also, the tooling, I, I think Amplitude is also moving to this warehouse native direction Slowly they are, they are still building on Snowflake.

[00:24:04] István: But I assume they’ll also try to do other data warehouses. AB testing tools like Apple that was recently acquired by Datadog. They also had a warehouse native approach to, you know, a lot AB test analytics. It is slowly getting to the, to the data world, this warehouse native approach. it’s still not exactly where I wish it would be, but it is, it is more and more appealing for everybody.

[00:24:30] Dara: Definitely. Why, why do you think that’s been slower? W why do you think it’s been slower on the, on the data side? 

[00:24:37] István: good question. To be honest, my full honest answer to this is, for businesses to be data warehouse, warehouse native, it is often candies in their existing business model.

[00:24:54] István: That’s why, that’s how I, I would say the, probably the biggest reason is how, imagine like yourself as amplitude, you are connecting to a, a data warehouse, some customer and. How are you going to monetize that? Currently their monetization strategy is volume based and suddenly they need to switch to something else, either a seed based, license based monetization, or like a fixed price, or they can’t keep up with the volume based pricing.

[00:25:21] István: But it is very hard to justify for the customer, I need to pay x hundred thousands of dollars for the data that I already pay for in Snowflake. So it’s that. I would say in short, for the existing incumbent businesses, it is a feature that they use to keep their all clients and it’s not a product itself that they think of.

[00:25:47] István: So this is my, my, short and honest answer for this. 

[00:25:50] Matt: Yeah, it makes sense. Out, out of interest, how do you see, so say, just to, to step back slightly to the idea of it getting more and more accessible for people to centralize data in, in data warehouses, there’s also this. very popular term at the minute.

[00:26:06] Matt: It’s not a new one, but this whole shift left the idea that these sort of federated data teams within a company own everything. So that’s like you have your, to, to speak in Measurelab’s, cadence, your marketing analytics team who own their own models and almost their own mini marketing data warehouse.

[00:26:25] Matt: Do, do you see that that trend happening more and more within, within your industry where there’s, what there’s less one big Goliath with the keys in the center and it’s, and everyone’s just trying to chip away a bit off the edge where they’re kind of getting ownership. 

[00:26:41] István: I would say we see that, but it is really dependent on the, of the, of the company itself. How is it set up, the organization, the capacity in like the data team or teams? Currently, we mostly see companies that have a few people in data. Any business that you work with. And there, I think it’s still like a one centralized team dealing with all the data, which has some benefits and also some disadvantages.

[00:27:15] István: Definitely, you know, it’s always a debate, like if you should centralize and have somebody that is taking care of everything, you know, or distribute the, the, the responsibilities. But you might end up in like, you know, multiple times implementing the same, same thing essentially, or even worse, some contradicting information in multiple these, these, let’s say these, these sub data warehouses.

[00:27:39] István: Yeah. So in my experience, every current customer we have, it is mostly small native teams that we work with. 

[00:27:45] Matt: It almost feels like there’s a combination of the two models, doesn’t it? Where you, you’ve got some centralized ownership and to, to stop the repeated data and the multiple answers to the same question, but also some sense of autonomy for the smaller teams.

[00:27:58] Matt: Because it, it does feel as well that the pace with which things are changing and new things are coming online and, and, and the AI features that say Google are chucking out Bitcoin and stuff like that. For people to take advantage of them and really grab hold of them, they’re going to need data in which to do that with.

[00:28:13] Matt: And if it is all slowed down by bureaucracy and red tape, that’s going to be tricky for companies to keep pace. Yes, yes. 

[00:28:21] Dara: Definitely. Yeah. So can, can I ask what, mostly out of my own curiosity, but I think there’ll be listeners probably who will, who will be in a similar position where they. They get it, they get the benefits.

[00:28:33] Dara: And I appreciate what you said earlier, but there may be a point of scale where this suddenly becomes almost like a no-brainer or at least something you should really consider. What, what’s then involved? So let’s say you are, you know, you work in a data team. You, you, you pitch to the business to say, right, we’re at that size, now we’ve got enough data, let’s move to a, a, a warehouse native solution.

[00:28:53] Dara: What, like, even on a high level, what’s involved? What, what does that process look like? What, what skills are needed internally? How do you go about implementing that solution? How do you then go about educating the wider business around how to make use of that? You know, what’s, what, what does that kind of look like typically?

[00:29:10] István: It’s a great question. I would separate these two types of companies again. one that already has an existing solution. Let’s keep talking about amplitude. If they decide to, you know, you, we need to get rid of Amplitude because the scale we are operating at is impossible to maintain anymore with Amplitude.

[00:29:31] István: The biggest challenge is that most of the users of that product’s platform are really, really embedded into the platform. So they have their charts, their dashboards and everything. It’s tremendous work to get away from that platform. So there is some kind of vendor locking, which is obviously coming, coming with every tool like that.

[00:29:53] István: That’s one part. For this, this group of companies that already have the, the, the, the tooling really, and also they have a data warehouse. From the technical point of view, to have a warehouse data approach to it, there is not really much to do in the data warehouse. There is no need for extra data modeling or, or, you know, sometimes transformations or preparation of the data.

[00:30:19] István: The reason why, because they already have a BI tool such as, let’s say, power BI or like, you know, Looker or Tableau. They already the, the data to be clean and performant enough. you know, even with Mali, they are, they already had to do the effort to make it, you, well, well optimized and the data to be, to be essentially clean and usable.

[00:30:39] István: So there is not much here that the team must do to switch to a warehouse data approach. It’s more likely to, one by one, shift the thinking of the people that now from now on, you will use this different tool. Although it looks very similar, your charts are going to be gone, your dashboards are gone, the buttons are slightly mixed up, but you have to use that from now on because it essentially makes sense.

[00:31:04] István: Everything else was what was there before. It doesn’t make sense anymore. So that’s that, the shift in the organization is the hardest part. for companies that already have the solution, for companies that don’t have a solution, that is surprising. A lot of them actually, they are, their problem is a bit different, but it’s, it’s also very similar.

[00:31:26] István: The product managers or we also serve marketing people. They don’t know these tools. They don’t, haven’t seen most likely amplitude mix all these programmatic tools. So they needed to be educated so that, so from now on, instead of you asking the data team for some basic query to, you know, one of your questions must be answered instead of the data team answering you, we need, in two days, you have this tool that you can use and answer for yourself.

[00:31:54] István: That’s a lot of education. So it’s still a problem. It’s not easier than the other one. Actually, again, technology wise, like data platform, data model, and transformations. If you have a data warehouse and you’re not starting out, you most likely have already made the necessary steps to make it performant and clean and everything and usable. So at least for us, we don’t require the company to do extra, you know, extra steps. 

[00:32:19] Matt: What it, it might be a good segue, I suppose, there, just to talk a little about, a little bit about the technology of, of what it is you are doing. So, the reason I think that is obviously you, you, you’ve got basic model data and you’re confident that you can, your system can talk that data and pull out meaningful insights.

[00:32:38] Matt: So, so how, how exactly is that layer working? I know you can’t give us your secret source, but how is that layer working ? 

[00:32:45] István: There is no secret actually. I’m happy to talk about how it works. Maybe let’s start with the integration process. How we do integration to the warehousing, to the warehouses.

[00:32:55] István: the goal of this integration is to, the data team connects their tables for the warehouse from a big query. Multiple tables or, or a single table, doesn’t matter. It can be even sometimes we have customers that have like a thousand tables connected to the, to the workspace in, in mid. So the integration’s goal is to kind of collect metadata from those tables, first of all.

[00:33:21] István: And by metadata we mean, in product analytics, the whole thing is based on events and properties and that you have user properties and in the data warehouse tradition that you have tables and schemas. So you’ll need to have a transformation in, in the tool that transforms those tables into meaningful events and properties and all that stuff that is already familiar for the users.

[00:33:46] István: for us, this is called the indexing process. You can imagine it as like you connect your thousand tables in, in it. So you press a button and what we do is kind of go to the table, read out the, the schema, the column names, the, everything that is like kind of metadata from the table. We also get a small sample of it, like a very tiny sample, just to see a couple of extra information.

[00:34:12] István: What could be the potential filter values for properties, et cetera. things like that. And this indexing usually takes, depending on the company and the, the, the data models they have from like one minute to half an hour. And once that is done, you, you got the events and properties and, and user properties type of setup like you are from there within any traditional product analytics.

[00:34:37] István: and if you have that, you can start using the application, build your funnels, build your retention queries, and it’ll translate that into SQL queries. And the secret sauce is that if we are doing it in an automatic way that is generating highly efficient SQL queries, I can give you an example.

[00:34:59] István: Let’s say a funnel query. You know, you want to measure. Page visit to check out in an e-commerce company, what is the conversion rate during like the last month? How many people check, do, checkout, from the visitors? How a data analyst would approach this with SQL is doing kind of left join and some filtering.

[00:35:19] István: And this is okay, this is easy to write for data lines, but it’s very, very suboptimal in performance. So for example, what we do is, a lot of magic with window functions and kind of unioning tables instead of joining them, which will result in a much better CPU complexity or like, time complexity of the queries.

[00:35:41] István: And essentially your experience with the tool is very similar, that of Amplitude and Mixpanel. 

[00:35:49] Matt: And is, is that a combination of knowledge of the, of common data sets that, most of your customers within particular sectors are going to be coming at you with like, you know, GA four, you know, the schema, you know, Google ads, et cetera.

[00:36:02] Matt: I, is that coupled with ai? Like is LLM helping generate any of that sql or is it all the knowledge of, of those schemas and being able to put together these good efficient window functions and knowing how to join them together into queries? 

[00:36:17] István: Yes, it’s a great question. Yeah. We don’t use LM, it’s, this can be systematic, you know, this, this approach is systematic.

[00:36:24] István: So you, if you do the same clicks in the application, it’ll produce the same SQL all the time. And I have a strong belief that this part itself, it’s not ideal for alums Because it’s, it’s kind of obvious what you need to do. There are a few things you have to consider when you look at the tables. You get the metadata from tables, you get information sites such as, what is the date partition column, you know, for GA four for example, you get the part, you get the information.

[00:36:51] István: What is the even type or even name column? You get it from the metadata, and based on that metadata collection and the information we have, we can kind of systematically generate very, very efficient ways without the use of lms. With that said, I’m really a big fan of LMS and AI and I think there is a lot of, a lot of role in this for them for product analytics, but not exactly the SQL queries generation.

[00:37:17] István: I, I think it’s, I think there’s many companies trying with text to SQL approaches, which makes a lot of sense when it’s a small amount of data in, in, at our scales that we operate, these high volume data sets that we operate with. I think there will be, quite so many problems with, lms, just text to SQL approach.

[00:37:39] Matt: We have alluded to that a little bit over the past sort of few weeks where, off the back, I suppose of Google’s big announcements of a lot of LLM stuff going into BigQuery and some text to SQL type type approaches that if you didn’t have your ducks in a row from a data perspective in your warehouse, it could lead to some expensive problems potentially.

[00:37:58] Matt: Because it can be quite autonomous in what it does. And, just sort of go off and start doing things. which is why I asked the AI question Because I wasn’t completely sure how you were doing it, but that, that’s interesting that it’s just, just deep knowledge and, and smart putting things together and, and, and functions, et cetera.

[00:38:18] István: To be honest, there are not that many, there’s not much secret sauce in this. You can, you know, identify like 10 cases and how companies, you know, store their data. And, if they don’t do it that way, as the, as our covered, you know, covered models, it is either going to be extremely slow for them anyway.

[00:38:42] István: Even a data analyst query would be very, very slow for them. So they, they, it is, it’s very beneficial for them to model it in a way that it’s fast in, in our tool as well. So yeah, that’s, there are not that many things that, you know, we have to be smart with, smart about. 

[00:39:00] Dara: Would, would the idea with, with the kind of optimized SQL queries and being able to do these things in a kind of more automated and streamlined way, the, the, you know, the kind of like on the surface level, you think, well, that’s got a real advantage to the company because then it can make these tools more accessible to people who don’t have SQL skills.

[00:39:21] Dara: Do, do you, maybe you’ll have to speculate a bit on this, but from the customers that you work with or the people that you know who are using, even if it’s not necessarily Mitzu, but just some, warehouse native solution being skeptical. I kind of imagine the reality would be that it actually makes the life easier for the analysts, but not because people start using these tools themselves, but just because they can save themselves having to do, you know, SQL coding every single time they get a request.

[00:39:48] Dara: But it’s that classic thing, isn’t it, of like, sometimes people just want to go to the person who’s got analyst in their job title Because they feel they’ll trust the answer more than, you know, don’t give me the power to do it, because then I have to trust myself. So I, I could imagine maybe the benefits are still there, but they might not always be.

[00:40:06] Dara: You know, there’s this kind of ideal view, isn’t it? That it becomes this like, you know, everybody’s using the, the, you know, creating their own reports. Everybody’s making their own queries, but I suspect that probably doesn’t happen that often. 

[00:40:19] István: You are right. Exactly right. It’s again, coming back to the two types of companies. One is familiar with product analytics tooling, and the other one is not. If the other, you know, the one that doesn’t know the, the, the product managers don’t know product analytics tools and they used to be talking to the data team asking for, you know, answers is going to be the same thing. They are going to still go to the data team and ask him for answers, but it’s going to be much faster.

[00:40:46] István: The answer will be much faster given. That’s a big difference. I, I, we had some success in educating people. It takes a lot of time. 

[00:40:54] Dara: Yeah. It’s a big undertaking, isn’t it? 

[00:40:59] István: Oh, and the other thing is like if you are coming off the world of amplitude and, and mixed or like GA and you know what these tool’s capable of and you are just given a slightly different tool to you’ll, you’ll figure it out as a product manager, you’ll always start using it.

[00:41:15] István: So it’s more like a culture question or like an organizational question. but definitely we had much bigger success in onboarding new users when they already knew what tool like that. 

[00:41:24] Matt: I had a question, which is a little bit more, I suppose you still have similar que similar problems to say, a company that’s doing text to SQL in that your output is only ever going to be as good as that data that is sitting in the warehouse for the, for the customer if they haven’t.

[00:41:45] Matt: I know you said that, you know, most of your customers are coming to you either with an established warehouse, so they’ve got some modeling, they’ve got some reporting tables. but if there’s other things further to the left of that from the co point of collection and getting all of that data right, so that what you are, what’s being produced in your platform isn’t, doesn’t look right, but isn’t Right.

[00:42:06] Matt: Do you do much in terms of helping clients understand that side of things to make sure that the data is solid before it gets ingested by your platform? 

[00:42:17] István: Yeah, I mean, in short, yes. We help a lot, our clients, we help, like, you know, they, they understand if there is an anomaly. They, they are not, you know, they can’t explain with some, you know, some release or something.

[00:42:30] István: We help them sometimes even debug on, on like on the roll level data, that layer, we help them debug this. I think it’s, it’s, it, this problem is not related to Mitzu. In general, it’s related to data collection and, and how you manage data. The one positive part of this, why I think we are, we shine at, in, in this, in this data quality is that there is, it’s not a black box that we built.

[00:42:59] István: Every s QL upgrade that we generate is visible to you. You can review it, you can, you can read, you can copy, paste it to your BI tool if you want. And so we are taking away that uncertainty or like, like, let’s say, source of risk where you are moving your data one more step to the right, let’s say to an external solution like third party analytics tools.

[00:43:22] István: so your data can fail in collection time. You know, like storage time. Some are, some things can go wrong. Data modeling time, you can have some problems and you know, and you, it can also have issues when you are moving it away to a third party tool. So with this, you are getting, at least you’re getting one less, place of, of a potential risk of your data being corrupted.

[00:43:47] Matt: Yeah, I, I guess, I guess that that, I guess my point, not that it was am Mitzu responsibility more that it, you know, you are going to get, anybody’s going to get more outta your tool if they’ve done their due diligence on, like you say, from point of collection to, to transport to modeling. Once they’ve done all that and they’ve, they’ve done it properly and maybe they, like you say, they, they grow their team internally to do that in some way.

[00:44:12] Matt: They engage with an external party to do the fundamentals in that way. They’re really going to get a lot of benefit from your tool once that fundamental foundation’s in place. 

[00:44:22] István: Yeah. I again, like the worst things get, things that can happen is like in one of your tools you see some numbers and the other tool shows.

[00:44:31] István: Completely different numbers for the same, supposedly the same source, and at least that is, you know, taken out of the equation, that problem. 

[00:44:39] Dara: What, what about, so a lot of the talk, so far, and I know this is kind of how you, how you position, you know, Mitzu, as a, as a product analytics platform.

[00:44:50] Dara: How does marketing analytics fit into this? Because I know mix, but I’ll probably get this the wrong way around, but I, I, I think I’m right in saying one of them recently added marketing analytics and maybe the other one had it before, and, but it, it, it’s like, you know, the, I guess the question is like, do, should they both sit under one, you know, should, should they be, using one tool or should they be separate or is it going to be a case by case basis?

[00:45:13] István: Oh yeah, it’s a, it’s a good one. I think the use cases are somewhat overlapping for product and market analytics. At least what we see. We have, we have both, from the customer base, we have both. Like some only use it for products, some use it for marketing only. some use for post, Also, it depends on the industry.

[00:45:33] István: I would say traditionally, e-commerce, travel industry, companies that don’t require your user to be signed in to be using the application. I think its product and marketing is roughly the same thing in, in the, in the data, at least in the, in this, like in this use cases, from the use case point of view in let’s, let’s stick to e-commerce.

[00:45:57] István: What do you want to maximize the number of visitors on your site and to maximize your conversion rate on the website? And conversion rate is like funnel analytics. Problem. And the same thing you can see, have it in the product if you want. You want to see how many people convert to a new feature in the mobile app or something.

[00:46:16] István: It’s the same for us. It’s the same. There’s no difference. There can be some difference in the underlying data for marketing and product, or like, for example, in marketing data, you most likely don’t have people that are, you know, authenticated. You might, you most likely have a visitor ID on or like an authenticated id.

[00:46:34] István: While in product, you are capable of seeing people that are recurring coming back to your product. You are capable of measuring retention or at least easier to measure retention to your product because everybody must sign in, you know, and it’s, it’s a, that’s, that’s different. But fundamentally the tool itself, I don’t think it should look different.

[00:46:54] István: It’s, the features in the tool should look pretty similar. 

[00:46:57] Dara: And, and then I, I, I kind of agree with you on that and then, but then I, I, I wonder, do you, you know, do you see this pattern where, in some cases somebody might still use a tool like GA four because of the integrations that offers with the other, you know, with the Google ads and, and you know, the, the, the DMP and all the rest of it.

[00:47:15] Dara: So I, is that you, you’re nodding, assuming it is something you see where people would still use GA four maybe for those native connections? Oh, yes, definitely. 

[00:47:24] István: Definitely. Yeah. Yeah. It’s, it’s, I think GA four, at least because of the ecosystem, is very useful around it. You know, it’s, yeah, as you mentioned, Google Ads is, is great, you know, tech manager, everything is, is kind of well fitting together.

[00:47:36] István: On the other hand, we also see people exporting their G four data to pick very, doing a bit of, bit of transformation, removing the repeated types, which is like an annoying thing in the G four data and, transferring, being that to Jason. and basically after that it’s, it’s, you are good to go to do a product, like marketing analytics on it.

[00:47:57] István: Yeah. So that’s definitely, I mean, these warehouse native tools, they don’t replace everything. Obviously, you know, you’ll have your dedicated tools. I think GF four is, is, is, you know, very good to, you know, it very well fits into this Google ecosystem. While our, like our tool warehouses, other warehousing tools, they are most likely a, a, a readonly type of, tool.

[00:48:20] István: Like you pull information out of the data warehouse, you don’t get like, you know, extra features beyond that. 

[00:48:26] Matt: So we, as we mentioned earlier in the podcast, about 2020 BigQuery and other data warehouse platforms have started to level up their games a little bit and, and make some of this data warehouse and analytics more plausible.

[00:48:42] Matt: What. What have they done? What have, what have they changed since 2020 that’s really leveled them up and made them in this way? I, 

[00:48:49] István: I, I think the answer is kind of like, there are so many open source solutions coming out like, you know, iceberg and Delta tables that enable this. And all the other players, like Big Query and Snowflake must act on them because they were lagging behind.

[00:49:05] István: They were slower than Databricks, significantly slower than Databricks. And now it’s such a big competition in this warehousing space to be performant, you know, and, and cost effective that essentially it’s, it’s, it has, it, it, it has been a revolution in it in the last couple of years. 

[00:49:20] Dara: I, I, I’m, I’m possibly being naive here as the least technical of the three of us.

[00:49:25] Dara: but what, what would stop, and, and, and maybe this is something that keeps you up at night, but what would stop Google, taking Google with BigQuery, but obviously it could be another, data warehouse, vendor, whatever. What would stop them from doing, like do you think they have an appetite to actually get into this space given that they are the ones actually storing the data?

[00:49:47] Dara: Do you think that they would be interested in actually adding this layer on top, or is that just not acting, that’s just not their, you know, that’s not their area of interest. 

[00:49:57] István: There are a lot of potential candidates that could implement this. Like any BI tool looker, you know, or like, you know, power bi, data warehouse solutions.

[00:50:08] István: Like it’s a big, very. Also they could do that. I, I think the answer comes down to how well it fits to their current vision of the product and, how I see it, let’s say BigQuery and let’s say Looker, they want to stay as a generic solution for a vital audience. You know, and this is, this product analytics is for like product managers, and if you screen it’s for marketing managers as well.

[00:50:37] István: This is like a small subset of everything they could do. So the question would be like, I would, I would rephrase the question like, why don’t they do all types of verticals? You know, and it suddenly becomes a different question, but it is the answer for it because, you know, why would you just pick this one and not do all the other verticals?

[00:50:57] István: Let’s say, why don’t you cover, I don’t know, the energy sector, the FI finance sector with verticalized tools? It is very, very hard to build a team to do this, you know, to build this. I, me personally, I started to work on this years ago. You know, there is a lot of, a lot of expertise coming with it and a lot of experience and we are a small team, but we are laser focused on this problem.

[00:51:20] István: This one or two people that use the application. I would say ide. Google would be much better off buying us out. Yeah, of course. Then finding their own thing, you know? Yeah, yeah. I’m, I’m making. 

[00:51:35] Dara: Which, which of course is often what they do, you know, so, so yeah. Yeah, it is. Yeah. Yeah. 1, 1, 1 other area, and then I, I, I, I think Matthew probably had a, you know, couple of, couple extra questions as well, but one, one final area for me that we haven’t really kind of gone into is around, kind of data privacy.

[00:51:54] Dara: And I guess that’s one of the, I’m going to state the obvious, but that’s a big advantage in somebody if they did go down the, warehouse native route is, again, you’re keeping all that, you know, the customer’s data is staying in one place rather than pumping it off into, you know, God knows where. so that must be a growing, you know, a growing trend that, that would obviously benefit, companies like Mitzu.

[00:52:15] Dara: But from a company perspective, you know, you’re, you’re reducing the number of places where something could go wrong in terms of data privacy. 

[00:52:24] István: Yes, definitely. We heard that from customers. A couple of industries are typical for this FinTech, a tech, B2B SaaS as well. When you have, like, your customers are mostly, you know, dealing business, dealing with businesses and their customers essentially, you know, they require them to, to keep the data safe.

[00:52:46] István: it is very often actually the motivation to move to warehouse native solutions in, actually, surprisingly, I was expecting it less, but in the US it is, as much of a concern as in, let’s say Europe. So, we have this in, from a couple of customers in the US. Basically they got low lawsuit suits because their data was, you know, corrupted.

[00:53:10] István: It’s easy to get burned, you know, by this, you know, if you, if you are, for example, in the ed tech industry. So, yeah, it’s a, it’s a, it’s a growing concern. I, I would even probably. Risk to say that with the current political situation, happening in the world, it’s, it’s going, going to grow even further. Yeah, but definitely it’s, it’s already there.

[00:53:31] István: It was already there for us before, you know, before the current climate, you know, started, you know. 

[00:53:40] Matt: I’ve got two remaining questions. I’m going to ask them both at the same time and let you decide which order you want to, you want to tackle them. The first question is about the branding of Mitzu in this sort of Hello Kitty-esque sort of little cat.

[00:53:54] Matt: I’m really interested in what, what led to that, and maybe even the name of, of, of the company as well. and the second was, was more of a bit of a crystal ball gaze. You, you mentioned that you were, you were very interested in AI and LLMs and, and all that stuff. It doesn’t have to be about AI and LLMs, but just where do you see I.

[00:54:13] Matt: Things are progressing. What do you see the next sort of five years being within product analytics, marketing analytics, and the world of tech generally? So take them however you wish. 

[00:54:23] István: I’ll start with the easier one with, as I mentioned, it started out as a hobby project I’m working on alone.

[00:54:32] István: And it became like an open source project. It was an open source project for a while actually. And initially the open source nature of it was an idea for marketing. I wanted to acquire leads, you know, potential customers through the developer community or data engineering community that are traditionally going for open source tools, with the hope that they will turn into paying customers.

[00:54:59] István: And naturally, if you’re going for like the, the, you know, like the individual contributors, let’s say data engineers, data analysts, you must appeal to them as a friendly tool, you know. Essentially like it, well, it, it welcomes you, you know, you can play with it. It’s like, it’s like a cat. You can play with it, you know?

[00:55:16] István: It’s like that. I, that was the idea behind it, behind the logo. We switched from open source to like, you know, SaaS business, but we kept the logo, which was an advice from one of the designers. We have, basically even if you are doing B2B sales, you know, we do enterprise B2B sales. So like sales cycles of half a year, you know the brand.

[00:55:40] István: It is, it’s still required. You need to, you are talking to humans, you’re talking to people and they’ll remember the cat. They won’t remember like an abstract whatever logo, you know, and they often, like, you know, they talk to each other. This is what we hear, hear back that, oh, you remember this CAT tool? What was the name of the tool? What was this cat? Insight. 

[00:56:00] Matt: yeah, I, like, I, it, everything else in the space feels very utilitarian. Do you know what I mean? It does differentiate it as a brand for sure. Yes. So, 

[00:56:08] István: by the way, there is a, a security company called Vanta, again, super B2B Enterprise salesy, you know, like it kind be came, cannot be diverse than that, you know, and they are, they have a logo for Lama A Lama, you know it, and it’s the same story.

[00:56:24] István: They, that’s what actually was the motivation as well. Like, it’s the same story. You are still selling to humans and you want to be on, in their mind, you know, by the name, Mitzu, it’s the name of a cat in, my name Native Tang, which is Hungarian. Yeah. So it’s, just like that. And, originally, I had a partner in the beginning.

[00:56:47] István: We were sitting in a cafe called Mitzi. And we were like brainstorming the name of the company and everything, and they’re like, oh, after one hour we said like, okay, just let’s just call it mid and have this catalog go. And it’s, it stuck and I think it’s too late to change it. So we might upgrade it to a tiger if we go up, up market.

[00:57:09] István: I like it. I like it. Yeah. and about the vision question and the looking forward and AI and alarms? I, I think it’s, I, I’m a mixed, I’m a mixed feeling about this in like, the AI in general. I, I, I really like AI tools like, for example marketing blog post generation or like, you know, like helping you write emails and everything.

[00:57:32] István: I also really like them in software engineering. we use it a lot actually in the company, with data. I think the, how to say, it’s a bit riskier to use, especially the text. Text to sql. I would, I would argue that. it is coming, but it’s not there yet. So my, my, my usual, like my, my internal touring test is if I ask a question from the text to SQL code, tool today, and I’ll ask the same question next week, will it give me the same answer?

[00:58:06] István: And the answer is often not. And this, the text to SQL part, I’m a bit skeptical at least. we are visioning in the product a text to mid two query, let’s put it this way, type of, type of feature that basically you can chat with the application, but it won’t generate SQL queries, but it’ll generate the blocks that are currently generated systematically in the, in the code.

[00:58:31] István: That’s the, that’s the way we look at like, like first MVP will be like this, and down the line, maybe a few years from now, I am visioning like, in this space that there will be like product analysts, agents. So basically you give it a task or something like that and it’ll work like a product list or like autonomously.

[00:58:55] Matt: Yeah. I do wonder if, I mean, we’ve got the same experience with our internal SQL SQL bots that they are, they don’t get consistency and they, they can, they can spot where things aren’t right. I do wonder if it’s just the absolute sea of web development data that’s, and information that’s out on the web, that’s the core of their training set where they are so much more adept at that.

[00:59:22] Matt: And where, you know, similar to what you’re discussing there, you’ve got Google and, and OpenAI creating these agents that can go to GitHub and work through issues and do those kinds of things. Because it’s got that corpus of underlying training data. Maybe, maybe the SQL and, and that kind of training data isn’t there yet.

[00:59:40] István: Yeah, I, I often run into, I, I, I often run into like, you know, GGPT using GG PT for like C code. Even syntax errors are appearing quite often. So it’s, it’s there. It’s, but I wouldn’t trust, let’s say, you know, imagine a company going for a board meeting and showing some data to the board, and I would, you wouldn’t generate a, a, a query with the, with the AI tool for that, you know, that, that influential meeting.

[01:00:07] Matt: Yeah. Yeah, it’s very conspicuous that Google, when in next 25 recently we had a clipboard of very curated prompts. Prompt. Yeah. Yeah, yeah. Yeah. 

[01:00:20] Dara: It’s been an absolute pleasure having you on the podcast. Thank you again for your, your, your time. It’s been a really interesting conversation. So yes, thank you.

[01:00:28] Dara: Thank you so much. It was a pleasure to be here. That’s it for this week’s episode of The Measure Pod. We hope you enjoyed it and picked up something useful along the way. If you haven’t already, make sure to subscribe on whatever platform you’re listening on so you don’t miss future episodes. 

[01:00:42] Matt: And if you’re enjoying the show, we’d really appreciate it if you left us a quick review. It really helps more people discover the pod and keeps us motivated to bring back more. So thanks for listening, and we’ll catch you next time.

• Spotify
• YouTube
• RSS Feed