Drill to Detail - Drill to Detail Ep.69 'Looker, Tableau and Consolidation in the BI Industry' featuring Special Guests Tristan Handy and Stewart Bryson

Episode Date: June 24, 2019

Mark Rittman, Founder and CEO of Rittman Analytics is joined in this special episode of Drill to Detail by returning guests Tristan Handy, Founder and CEO of Fishtown Analytics and Stewart Bryson, Founder and CEO of Red Pill Analytics to talk about the recent acquisitions of Looker by Google Cloud Platform and Tableau by Salesforce, the wider story about consolidation in the BI industry and whether the trend in analytics tools is towards enterprise features … or open source?Five Thoughts on the Looker Acquisition by GCPCloud - Mark Rittman, Twitter ThreadA Wave of Acquisitions in Business Intelligence - Tristan Handy, Fishtown Analytics BlogGoogle to acquire Looker - Thomas Kurian Blog PostGoogle Cloud and Looker (PDF)Siebel Analytics - The Jewel in the Project Fusion Crown? - Mark Rittman’s Oracle Weblog via Internet ArchiveApplications in LookerSalesforce Signs Definitive Agreement to Acquire TableauSalesforce.com's Tableau Acquisition: Admitting Organic Innovation Failure?

Hi, I'm Mark Rittman and welcome to this special episode of Drill to Detail, where I'm joined by two returning guests who need a little introduction from me, Stuart Bryson from Red Pin Analytics and Tristan Handy from Fishtown Analytics joining me to talk about the news that everyone's been talking about in the BI and analytics industry, the takeover of Looker by Google Cloud Platform and Tableau by Salesforce and all the other kind of acquisitions and mergers that are happening in the analytics world at the moment. So Stuart, welcome back to the show and maybe introduce yourself to maybe one or two people who haven't heard of you. Thanks, Mark.

I really appreciate you having me back on. Always a good time. I'm Stuart Bryson. I'm a founder, CEO of Red Pill Analytics. We work in the, obviously the analytics space, data engineering space, mostly with cloud native technologies, but we still do a lot of work with on-premises technologies as well. And just really, really glad to get on and talk about what is a wide open BI and analytics space and with all the news we've had over the last few weeks.

And Tristan, it's great to have you back on the show. Yeah, sure. Thanks for having me back. Hey, my name's Tristan Handy. I run Fishtown Analytics and we maintain DBT, which is an open source data transformation tool. We also do a lot of consulting in the space. We work with VC-backed companies to implement the modern data stack. And the funny thing that I've noticed over the past three years of doing this is that a lot of the tools that we use in the modern data stack haven't changed that much. And so that's why it was so interesting to see all the shakeup in the past

couple of weeks and months, because my guess is that that's going to change moving forwards. Okay, Tristan. Well, it's great to have you back. And let's just summarize then the news that's happened over the last couple of weeks and pretty sort of like pretty amazing news really. So first of all, there was the acquisition of Looker by Google, a cash transaction, 2.6 billion, and they're now part of Google Cloud. So as well as that, about a week later, or a few days later, there was the acquisition of Tableau by Salesforce. Again, massive thing, and actually in kind of, I suppose, financial terms, much, much bigger than the Looker acquisition. And then before that, there was the acquisition of Periscope by Sisense. And actually, there were some acquisitions, not so much within analytics,

but to do with kind of data integration. There was Stitch by Talend and Illuma by Google. So a lot going on, but particularly the last week, the last couple of weeks when the Looker acquisition and the Tableau acquisition. So, I mean, let's start off really by talking about, you know, Looker's acquisition by Google. And I think actually the biggest, I think the person who had the biggest surprise or certainly was one of the first to comment on it was Tristan. You know, I think you commented on the DBT forums about that. So, you know, was it a surprise to you, Tristan, that Looker were acquired by Google?

Oh, was it a surprise? I thought it was maybe a little bit of a surprise. Yeah. I think it makes complete sense. And I can talk a little bit about that as well. But Looker was one of the real rocket ships in the BI space. And I've been to join a bunch of times. I've heard them share kind internal financial results before, and things are going great. So I am a little bit surprised that they did. I mean, obviously, it's a tremendous acquisition for them. I had always anticipated that they would end up IPO-ing, though. Okay, interesting. Stuart, what's your thoughts? Yeah, I think my reaction mirrors Tristan's as well. I mean, your initial sort of reaction, at least my reaction was shock. I mean, it really surprised me. And that lasted five minutes before sort of rationality kicked in. And it makes perfect sense. I mean, they have Data Studio and we've done Data Studio projects and Data Studio has some pretty decent capabilities, but it certainly doesn't have the full complement of what Looker has.

And you could, you know, they've released BI Engine recently. The combination of those things you try to maybe put together and think, is this a real full analytics tool? And it's not. And I think they could continue their Google after all, and they could continue to, they have clever people and they could continue to build stuff. But I think what they really needed was to hit analytics hard and to hit it fast and to hit it now. And I think the only way to really do that is with an acquisition. And I think we've all used, everyone on this podcast, we've all used Looker and we, we all like

Looker quite a bit and it is a fully functioning, you know, end to end analytics tool. And I think that to, to also respond to something you said, Tristan, which is in your introduction, which is the tools haven't changed that much. Do you mean that the choices that we have or the actual functionality of those tools, They haven't been progressing. Is that what you meant? I meant the actual tools that we recommend to our clients who are kind of picking technology for the first time. Yeah. Yeah. And so I think that with Looker coming to Google, it's going to give them immediately a real analytics tool that can hopefully bring those data warehouses off-prem, bring those into the Google Cloud, and to have sort of end-to-end analytics, data movement, all of that happening inside of GCP. Okay.

Okay. So, I mean, why I it is it is in two ways so first of all was it a surprise that looker sold and then was it a surprise that google cloud bought you know and i think in terms of um looker selling i mean the i think the thing to bear in mind or certainly i was very aware of is there was a lot of vc money going into looker um and it had a very kind of high valuation um and so so somebody somewhere is going to want to make some money off of this. So the fact that they sold, I mean, I very much, you know, I suspect that people within Looker maybe didn't particularly want to sell at this point.

Maybe it's a little bit early. Maybe the IPO, as you said, is something there. But there's a lot of money, I guess, probably in the background there that wants to kind of like get a return, really. And so I think it's not a surprise that they sold as such. And I think, I mean, we'll get onto the product side later on, but I do think that probably it's good timing in terms of infusing Looker with some ideas and some sort of like maybe direction with the product,

because certainly, you know, going to join last year, I mean, great company, great people and so on, but I wasn't particularly kind of bowled over by, you know, innovation and ideas coming out, especially in terms of company great people and so on but i wasn't particularly kind of bowled over by you know innovation and ideas coming out in terms especially in terms of the packaged applications and so on so it wasn't wasn't a surprise to me that they sold uh and also you know in terms of was it a surprise that google cloud bought i mean as stewart said it's a kind of an obvious fit but i think the the major thing really the major kind of factor in this is Thomas Kurian being at Google Cloud. And I think, you know, the story here is quite a kind of parallel to the story that Stuart and I know from when Oracle bought Siebel Analytics. I mean, Stuart, do you want to kind

of maybe talk through what the sort of history of that was? And I suppose really, do you remember the days of Discover? I mean, Discover to me is a bit like kind of Google's data studio. Yeah, exactly. I mean, what happened then is a bit like kind of Google's data studio. Yeah, exactly. I mean, what happened then? Just kind of tell that story a little bit. Yeah, I agree. So, you know, at the time when Oracle acquired Siebel, you thought, you know, it's primarily for Siebel CRM.

And I think that was a big part of it. But I remember a blog post you wrote at the time, Mark, talking about how, you know, Siebel analytics was sort of the crown jewels, I think is the term you used. And it certainly made a lot of sense because Oracle had, in my mind, and I think yours as well, floundered in an analytics strategy. They had kind of cobbled together some tools and none of them really were end-to-end. A lot of them covered off some space here or covered off some functionality here and just a whole bunch of stuff. And they didn't have an end-to-end analytics tool with the capability to do curated, to do all the sort of stuff that Looker can do. And I think Thomas started to, or Thomas Curian started to

assemble pieces for an analytic strategy. And it definitely feels very similar. When you look at the Illuma acquisition, we need a way for on-premises data to get into the cloud easily. And now we need this end-to-end BI tool that's not sort of a hobby. I mean, that's going to come off wrong. Data Studio is a great tool, but it's not a serious analytics tool in my mind. And I think this is a serious analytics tool. And I don't think Thomas is done. I mean, he's going to put together everything he needs to go and bring on talks very specifically about industry-specific analytic solutions packaged for verticals and so on. I mean, that to me is where it's going. I mean, Tristan, what's your thoughts on this? So I wrote kind of a hot take blog post the day that the Tableau acquisition came out. thing that I was focused on is that, so I do a lot of solution selling to, you know, and I might end up doing this at higher velocity than, you know, most consultants because we work with smaller

companies. And so we kind of have more of those relationships. And so I get to see like tons of these come together and what always, you know, I'm an analyst. I'm maybe more focused on the back end than the front end. Like I don't have strong opinions about how a dashboard should look as much as I care about how the underlying technology works to produce that dashboard. But when I'm selling a solution to a client, it's the BI tool that actually drives their perception of that value. And so to the extent that Looker sales reps are in conversations every day, all day

about solution selling and analytics platform, they right now, or they prior to this acquisition were selling the benefit of Looker without actually expressing a strong opinion about what cloud partners best with Looker. And they weren't able to extract the value from the rest of the solution. And now Google's able to construct an end-to-end stack and extract the value of that entire solution, Looker is always going to be worth more in the context where it is driving value across the entire stack, as opposed to just grabbing the value from the top part. Stuart, I mean, your company and mine are both Looker partners. I mean, how do you think this will kind of play or benefit

or work in terms of being, you of being as a partner really for this as well? Yeah, that's one I've given a lot of thought to. We do professional services for Looker. I wonder if that relationship is one that will persist. We really truly hope so. I think that the attention that Looker shows our company is one that I truly hope continues. I don't know that that sort of care we get from the partner channel and the partner network at Looker, which has been so great for us, will continue. I don't have any reason to believe one way or the other, but obviously, when you've planted stakes in the ground, you do start to consider these things. I think one of the things that's interesting, I wouldn't mind hearing Tristan's take on this as well, is it was acquired, Looker was acquired by Google Cloud, right? And

that implies to me, when you look at all the assets that Google has, none of them are sold as licenses. They're all sold as sort of resources, right? So they're all sold by either uptime, compute, or pay-per-use. So this might actually increase the uptake of Looker by some companies who might go and give it a try because they spin up an instance and they don't have to pay whatever the annual subscription is to get started. They can try it out. They can spin up an instance, maybe similar to what you do with Cloud Data Fusion today, just an instance running and give it a try.

So that might actually increase the uptake for some smaller companies. And it might increase the uptake for enterprises, which will go and instead of having to negotiate this contract, they'll spin up an instance and start building some things and see if it works. I honestly hadn't considered that yet. Does Google Cloud not sell licenses in the kind of traditional style that Looker licenses are sold under for any of their products? Not that I'm aware. And I actually went through and took a look at everything I could find in the console. And everything is either, you know, spin it up and pay for utilization, or it's pay per use, you know, BigQuery is pay per use.

That's really interesting. I mean, that would mean, if that ended up being the case, I would become the biggest fan of this acquisition of anyone out there. So I think that the primary value of Looker as an acquisition target is that they've done a really hard thing. They've actually taught a group of people, data analysts, how to do a brand new thing that previously wasn't really something that they spent a lot of time doing. Looker has taught them to write code as a primary way of spending their time. And it's taught them to think modularly about how they construct their data solutions. And I think that was a really hard thing to do. And I don't think that, you know, even given all of Google's resources, I think it would

be a really hard thing to create another product that took analysts along that same learning curve. So if they can adjust the pricing such that it's more accessible to more people, I think that's really a tremendous win. Okay. Okay. I mean, looking up, so looking at the, again, the blog that Thomas Curian wrote about the acquisition, you know, the first thing that he said that this would give Google Cloud in terms of capabilities is the ability to define business metrics once in a consistent way across data sources. And that was the thing that kind of interested me and Luca at the start, the fact that

it almost like reinvented this idea of the semantic model for, you know, an industry and and that was the thing that kind of interested me in look at the start the fact that it almost like reinvented this idea of the semantic model for you know an industry and people that were used to kind of i suppose just writing sql or or not having this thing at all i mean uh again any thoughts really i mean to you tristan actually first of all what's your what's your take on the value of the model within within looker because obviously there's kind of slightly mixed kind of i suppose uh mixed feelings about the DBT, but what's your feelings about the model in there? And what value do you think it will bring to Google Cloud going forward? So I really don't think there are, I don't have mixed feelings about the LookML model versus DBT. I think they do two different things. I think that there's the data transformation and preparation part, and then you layer on the

semantic model on top of that. And the job of the semantic model is to essentially teach the front end how to build itself such that non-technical, non-SQL users can drag and drop. And from my perspective, and honestly, I'm interested in hearing, I don't go back to the prior generation days as much. And so I don't know prior tools who have done semantic models before as well. But it seems surprising that in this generation of tools, Looker's the only one who's really gone hard at that angle. Yeah. Yeah, I agree. As someone that does come from that prior, consider yourself lucky in some ways, Tristan. But no, I mean, this is, when we go to talk to, Red Pill Analytics began primarily as an on-prem Oracle consultancy around

data warehousing. Mark and I have similar past histories, even more colleagues at a time. And when we go to talk to these customers that have these assets and are trying to get off of them, not having a semantic model or a metadata layer is a non-starter for a lot of enterprise customers. You could argue, Tristan, whether it should or shouldn't be, right? Because data curation and data movement is so much easier with tools such as dbt today than it ever was before and so it's arguable whether they should have that position but quite frankly they often do and i think that this this is the sort of the marriage of uh of thomas curian and google i'm trying to build an analytics platform it needs to have metadata if you're going to go after these sort of enterprise customers.

It needs to have a semantic model. Now, does it have the kind of semantic model that we're used to from the legacy tools? Not quite. And that's what I'm interested to see. It's still single source, whereas a lot of these enterprise legacy tools can generate queries for a single report across multiple data stores. And I think that it'll be interesting to see whether the product moves in that direction and what Mark and I are used to from a lot of those semantic tools or continues to move forward in sort of the more of what we're used to with a single data

warehouse in the modern stack. That's the interesting thing to me. Okay. Okay. So what about the topic of packaged applications? So again, I think in probably the last episode that we all recorded together, we were talking about packaged applications and would you potentially build packaged applications on top of Looker? Now, clearly Thomas Kurian and Google Cloud are going to want to do things around sales analysis apps or other things there. And I think that Looker's attempts at that so far

have been a little bit kind of half-hearted. I mean, again, maybe sort of like to Tristan, first of all, do you think they're going to invest in packaged apps and vertical apps? And where do you think it will go? So the past two looker joins have had very similar refrains um and the the platform has been the focus of both of them um and i think that there have been more compelling demos uh in as things have moved forwards um i still don't

get the sense that their packaged apps are a central thing that is a reality today. I think that they definitely care about moving in that direction. My take is that the most important thing will be, let's say that they develop a great packaged experience. How do you transition from that experience to the more freeform experience of exploration? Because I'm sure all of us are well aware how quickly packaged experiences break down because no two businesses are exactly identical. Yeah. Stuart, I mean, you and I have worked in the world of packaged applications before, and it's certainly a mixed blessing, isn't it, in terms of how far you can take it?

Yeah, it's a mixed blessing. I think our experience is coming from the Oracle packaged application world. And there was a problem immediately from those solutions. And that problem was that the source systems that they came from, which were ERPs were quite often customized. They shouldn't, they shouldn't have been customized often in the ways they were, but they were, I mean, each and every one were super customized. So that breaks down where the ETL and the, and the analytics that, that bring that stuff into an analytics application that has to be customized, or it has to be, you have to address those customizations. I think when you start looking at at least what Looker has built initially, which is marketing, digital marketing sources built in GCP, by the way, that was a nice choice

as it turns out to have built their marketing application on BigQuery. But I think those sources are less customized. I think there's more experience, there are more events, it's more event driven than it is really sort of ERP driven. So I think the way you model those solutions, there's not so much you can do from a customization perspective. I also think that the applications that you would focus on pulling data from today are written in a more modern way. And they're written with extensibility built in and APIs that are extensible. I think that you would probably have an easier time handling customizations that might occur in these more modern applications than you did in these monolithic, in some ways, awful ERP systems. So, I mean, looking at the press release or the blog from Thomas Currie, it talks about using Looker to deliver industry-specific analytics solutions in our key verticals, whether that's supply chain analytics, media analytics, and healthcare analytics.

But, you know, Stuart, I mean, the thing is there aren't any applications in Google Cloud. So I'm kind of wondering how this is going to work, really, and what are these applications they're going to provide analytics for? I mean, I guess you could – I agree with you. I hadn't thought of that. So suddenly Thomas is in an unusual situation where he doesn't have all the SaaS applications to source all the data. He doesn't have an immediate set of customers.

So he either has to A, go and build applications in the Google Cloud, and that's not something that's going to happen quickly. Or he's going to go and address all of the SaaS applications that are so common across enterprises today. So it's the sales forces of the world, Workday, ServiceNow. I'm going to assume, Mark, that these are the things he means. Those applications all have reasonably modern stacks so that you can get to their data with REST APIs and other sort of more modern ways to get data out of them. So he could go and do that. He could have shown his hand a bit in the idea that they're either going to build or buy some of these applications. And if they're going to start acquiring source applications as a way to serve a downstream analytics business,

that could truly, truly get, Google could truly get to be very, very big in the enterprise space. Imagine if, a bit like a Google G Suite, Google brought out a suite of ERP and CRM applications, and you paid for it by, you know, advertising your data or by, you know, maybe a low fee or something. I imagine if Google could actually index the world's transactions as well as the world's searches and email, you know, that would be an interesting sort of proposition really. And to run analytics on top of that would be sort of very compelling. Yeah. I just, I feel like there's this interesting line. You know, it almost goes back to like, the Ronald Coase theory of the firm, like, where do you draw

firm boundaries? Or in this, in this case, it's like, where do the economics of the hyperscale cloud providers end up encouraging them to go? I mean, to date, they have been very much infrastructure. And the big cloud providers are, by and large, not providing the applications as well. But there's no necessarily, there's no rule that says that has to continue to be the case. But if BI tends to want applications to be able to deliver these curated experiences on top of, we may see that start to happen, which I kind of like the way that things are right now where the cloud providers are kind of dumb pipes and everybody built on top of them. But maybe that will change.

So Mark, can I put just one comment on there? The one thing that he might be thinking about is not the applications that you and I have in mind, but it's applications that Google does have, such as YouTube and Search and AdWords. I don't think that's what he means. I don't think that's what he means, but he could mean that. And it would certainly be a place to start. It seems slightly unlikely, though, that Thomas Curran would have bought this all to provide reporting on top of youtube i mean that's i get what you're saying and we shouldn't and we shouldn't and we shouldn't dismiss that but you know this guy has a kind of brain the size of a planet really and and you know you think i kind of suspect there's something more going on really and and that could be interesting so but let's

look at the the um the tableau acquisition by salesforce i mean so obviously that was a massive deal in terms of kind of value i mean it's all shares as opposed to uh cash um and and you know you can't help thinking that maybe the timing of that has something to do with the look of one maybe they were second choice maybe they were whatever whatever but um i mean to uh again tristan first of all um what why do you think salesforce why do you think Tableau sold? And what does it say to you at the time? So honestly, this is the one where I want to hear what you guys have to say about this. So we don't actually, we don't use Tableau internally because it doesn't have that semantic layer.

And I honestly don't enjoy the process of creating a Tableau report. And, and, you know, we don't, we do plenty of analysis on top of Salesforce data, but we don't like get deep in the weeds with, with Salesforce. So my, this has to be a question about the, the selling process. I don't think that those two products are like in any way more joined at the hip than, than any other two products in this category. Yeah, I can't make heads or tails of this one. I mean, obviously, they have a huge customer base. And you can see this as a customer acquisition strategy, but if you're just going to lose them, because I can't see where the Tableau architecture, which is still very much desktop driven, even though you can now author

using a browser in the newer versions of Tableau, no one does. And that's a broad stroke, I know. But in general, most Tableau users are still developing stuff on the desktop and publishing it either to a server or to Tableau online. And so that architecture for a SaaS CRM system, that doesn't make a whole lot of sense to me. and maybe they feel like they can be the company that reboots the Tableau architecture to be one that can run alongside a SaaS application. But I think there's going to be a lot of work to be done. So if they see this as a strategy to try to move Tableau users to some of the other web-based analytics acquisitions they've made in the past, good luck with that. Because one of the things that a Tableau user does not want to give up is Tableau. And so I don't see that being a winning strategy. And I think that a retooling

of Tableau to be truly a modern web-based architecture is one that is going to take a while. Yeah, I mean, it's just really sad, isn't it, really? I think my first reaction was it was a bit sad. So obviously, on a financial scale, this is big news because, you know, Tableau are massive. And I think we sometimes, you know, it's a bit like a home for the bewildered old kind of like desktop technologies, you know, Tableau are massive. And I think we sometimes, you know, it's a bit like a home for the bewildered old kind of like desktop technologies, you know.

And I do get that. It's a big financial kind of deal, 15 billion, I think it was. Tableau is by far the biggest kind of vendor in this space. And Salesforce are massive. You know, Stuart and I know that when the Salesforce conference happens in San Francisco, it's even bigger than OpenWorld. It's two times bigger than OpenWorld. It's massive, yeah.

And so this is a big deal financially, but it just strikes me as a bit of a financial transaction, really. I think, as one of you said, it's a bit of a kind of a customer acquisition. And I mean, this makes no sense. I can't see anything that Tableau customers and the product will gain out of this. You know, it's not as if, you know, the whole point of Tableau is its desktop and all that kind of stuff. And I can't see what they'll get out of it, really.

I think that Stuart and I know the kind of fate of tools like this that get absorbed by these big enterprise companies. And then they spend the rest of their, doomed to spend the rest of their life just being integrated into into products that the customers from tableau don't want and and just makes it all run horribly and so on uh i'm not going for a job at tableau after this and um and then and then for sale and then for salesforce you know the whole the whole premise of their story is is to do away with kind of software and things you install and

now they've got tableau the kind of most obvious example of that. I just can't – I can see it as a financial thing. It's a bit like when one big company takes over another big company. And, you know, the customers are the kind of, in a way, the kind of the victims of this. But I can't see really any benefit of this at all, apart from it being maybe a good bill. And it's all shares, so there's been no cash changed around.

Maybe Tableau's investors want it out. Salesforce kind of want to do something. Maybe they're after Looker, first of all. And this is the kind of, they've got the second choice thing. I don't know. But, you know, no opinions there really. But it's, I mean, to Tristan, does any of that sound sensible? It doesn't.

And I'm so glad that I'm not the only one who feels a little cold from this. I read several articles about it, and I didn't feel like I ended reading those articles any smarter than when I went in. It feels like Tableau has a huge revenue stream. It's growing, and the retention is probably solid. And so, like you said, this feels like a financial transaction and there's probably a cross-selling opportunity with Salesforce and that's where the synergies lie.

Yeah, yeah. I mean, and also, it does make you wonder. Sorry, Stuart. Oh, sorry, Mark. I was going to say a couple of things. First off, your Tableau partnership paperwork should be in the mail or in the post, as you guys say. But secondly, what's interesting about, and I'll draw a parallel to Looker here. I started, and you as well, Mark, we started working with these legacy tools.

And from the time they started trying to modernize, Looker was born and sold, right? So in a lot of ways, it's difficult to go and retool a platform. It's just not easier. In some ways, it's easier to start over. And I think that if they truly are going to try to make Tableau something that runs alongside or more of a web play, more of a SaaS play, they got their hands full. And I think it could be difficult. Yeah. Yeah. And it also makes you wonder what is happening with the kind of Einstein products now.

So there's a fantastic product called BeyondCore, which I was raving about for ages. And funny enough, the guy who was the CEO was going to come on the show back at the start. And then the acquisition happened by Salesforce and never heard of again. And he thinks, and I don't know what's happened to him or the product or whatever, but have you seen it at all? Yeah. I mean, I obviously never used it. It's all academic. I haven't been hands-on with it, but certainly I've looked at it. I remember when the acquisition occurred. You know, Salesforce has acquired several things.

Well, they built Wave Analytics, and then they bought BeyondCore as a way to try to, I assume, it was going to be to report on that. But there seems to be kind of a black hole of what's going on with their analytics acquisitions. I'm not seeing fruit from any of this work they're doing. And maybe this could be the Hyperion acquisition in some respects. You know, Stuart and I again remember there was a product line from Oracle for OLAP called Express, and then it was Oracle OLAP after that, and technically very good. But it didn't particularly get traction with kind of, I suppose, a large audience.

And then Oracle went out and bought probably the equivalent product in the market, which was Hyperion. And, you know, for a long time people were going, well, this is just a duplicate of what you've got. And in a way Salesforce have all these products that they would – you know, they have the equivalent products to Tableau already. They've got BeyondCore. They've got those things there.

But what Tableau brings is a solid product with a solid user base, and it works, and it delivers the goods. And maybe this is their Hyperion acquisition. I mean, Stuart, is that something else? Yeah, it makes sense if you can retain those customers, and Oracle did. So I would say the Hyperion acquisition was a better one than the Express one as far as the customers that they were able to maintain. You know, maybe it's not as difficult as I'm thinking. Maybe they're going to continue to just run Tableau as a business and try to cross-sell where they can and just continue to do what Tableau has done well, which is empower analytics users.

But Tristan, as you've said, it's not my way of building things, right? I prefer sort of a looker approach, the no software approach, but maybe they can make a go of it. Anything where you have to repeatedly hit save as and create like a series of assets, none of which are connected to each other. It feels like Excel to me.

Do you, I mean, to both of you, do you think this is, I think actually you've said this, Tristan, does this mean that in fact, all these tools that started off as desktop tools and lightweight tools are being inexorably kind of drawn into enterprise? So let me put on my DBT hat for a second and take off my consulting hat. We have been building DBT for three years. The early adopters were VC-backed startups and their data teams, and it's grown through organic word of mouth through that.

And then it reached some point, and that point for us was like December, January, where we just started getting interest from the enterprise. And it hit some maturity inflection or organic adoption inflection. I don't really know, but we started getting these big inbound requests. And we have experienced the same thing that the rest of the ecosystem has experienced where you've got the existing users who are using an open source product, and maybe they pay a fairly small contract size for the, the proprietary dbt cloud. Um, but, but then there's enterprises who right off the bat say like, we want to run this on prem and we're willing to pay. It's not like 10 X it's like a hundred X. Um, and, and, And it's hard not to have that at least influence your product roadmap. And we're by no means stepping away from caring about the rest of that ecosystem as well. But

just feeling it from the other side of the coin, I can really empathize with these companies where somebody shows up at your doorstep and they say, I've already used your product and here's a bag of money. It's hard to say no to that. So, Stuart, I mean, you're a red pill and so is my company as well. But you particularly straddle that kind of enterprise and new analytics sort of world. So do you think there's a need for tools like Looker to become more enterprise? And do you think there's a kind of demand for it in the enterprise? Yeah, it's interesting. First off, Tristan, if you ever have too many of those bags of money, you got mine.

Still very few, very few bags so far. Yeah, I mean, it's an interesting one. Mark, as you know, we have pivoted from a pure sort of enterprise consulting company to one that does a little bit of both. And in some ways, we sort of try to take enterprises in a more modern direction is one of the things we do. even mid-sized companies, or even some smaller departments within big companies that are all about cloud services and giving them a try and sort of pushing some buttons and seeing what they can do, trying open source software and experimenting with open source software and seeing the value of open source software. But in general, enterprises are afraid of open source software, but in general, enterprises are afraid of open source software. And in some ways they're afraid of cloud services. I'm painting with a broad stroke there and things are changing, but a lot of enterprises would say that open source project looks really good. It's

a shame I can't pay for it. And I've often struggled with that mentality. I get that they want support and I get that they want to be able to have a neck to choke if something goes wrong. But if you look at just the quality of something like Oracle support over the years, I'm going to get some calls on this one, but the quality of Oracle support over the years, I'd rather have open source in a lot of cases. So that whole idea of it having to be pay, it can't be open source, it has to be paid. And Tristan, it has to be on-prem. I can't believe we still see that. But I know in the DBT discussions that you and I've had about how do I approach enterprise customers, they want to be thinking about

on-prem implementations of dbt cloud. And it just is so frustrating. It seems like we should be past that. But everyone has to grow. These companies have to evolve. In some cases, it's just simply there's a fear of the unknown, a fear of not controlling your compute resources in the way you've always controlled them, etc. But I think that perhaps as companies like Google buy products like Looker and companies like Salesforce buy products like Tableau, I think that just gives everyone a little bit more comfort. If Google continues to go after the enterprise market and cater to them

with cloud services that feel like these enterprises need them to feel, then I think that we'll start to see some success. And I think that it is a clear path that we are going to continue to watch products, cloud services, and things that we love just the way they are get gobbled up. And we have to learn to live with that a bit too. Do you think that, I mean, actually, fair enough, I recorded an episode of the show yesterday with Bruno Aziza from Oracle about what Oracle are doing in this space. And Oracle's proposition is very much, Stuart, as you know, is about integration. And it's about, it's not about features and it's not about, you know, that kind of thing.

It's about integration with all the kind of tools and processes you use and so on. But do you think that's where a lot of the emphasis will go with kind of, with Tableau and Looker in the future, integration with the rest of the ecosystem in that world. I just feel like we've watched this movie before. I mean, if you look at what Oracle meant by integration in those days, it wasn't really integration for the customer. It was integration for Oracle so they could run it all in a combined stack. And so I think if you continue to see, if you don't see this as customer-driven, what is integration that's valuable for the customer?

Most of these products have REST APIs, and there are already other services that tie things together from multiple clouds. There are services that try to bridge the gap between the REST API from this product and the REST API from that product. I've been fine with best of breed when these tools all are designed

to not just work with one another, but work with anything. That's the main way they've been designed. If Google and Salesforce acquire these tools and make them such that they work with a few less clicks or a few less API calls with their own products, that's great. As long as they continue to be extensible through other people's products just as easily. I mean, if you look at all the things that can run with Looker, that work with Looker, if you look at all the things that can work with BigQuery, you don't have to silo yourself on a GCP stack. If GCP gives me something that's inherently easier

to plug together and still best of breed, I'm all for it. But those two scenarios rarely occur at the same time. So I think that the way I'm thinking about innovation at Looker moving forward is that there's maybe there's two categories. And Stuart, this is like largely along the lines of what you were just saying. I think that there is a core experience that at this point probably will not fundamentally change. The first category is things that I don't think will actually happen because they're not actually aligned with integration with GCP. my struggles doing dynamic joins in LookML based on a date attribute. And how this really, like, LookML's join structure, like, doesn't facilitate this super, super well. You have to do a lot of work to make it work well. And I was wondering, hey, is there going to be any, like, different query pattern that gets released in LookML. And I think the answer to

that is probably not. Probably we've got the, in that category of things, we've probably got what we've got. But then there's this whole other category of, you know, the BigQuery team in particular seems interested in extending the SQL dialect in, or their proprietary SQL dialect in ways that expose more and more functionality that isn't like what you would traditionally call database processing type functionality. So they have the ability to train, I think it's both linear and logistic regression now in BigQuery using essentially like SQL type syntax. And I'm sure that you can, you know, if it's SQL, you can kind of hack it together in Looker,

but that's not a first class experience. So what if, you know, the power of Google's machine learning were going to be exposed through a LookML or, you know, first class Looker experience? Like that's the type of stuff that if we're going to see innovation here, I think it's in that direction. I mean, Stuart, you and I have lived through this with the Oracle world.

And Oracle bought Sibling Analytics. They therefore had the opportunity to, I suppose, leverage more of the Oracle database with that. And things like machine learning and data mining have been in there for many, many years. It never really happened, did it though? And there was always this kind of, I suppose, tension between focusing on the Oracle database and losing the customers of all the other databases. I mean, what was your memory of that, Stuart? No, absolutely. My memory was that every time they doubled down on their joint investment and making something work a little bit better with the Oracle database, it sort of took away from the core capability of an analytics

product. Now, I'm not sure what we're going to see. If we're going to see Looker become primarily an extension of BigQuery, I'm not sure that I would be for or against that, right? I'm not positive yet. Exactly. If, if it gave us tons and tons of functionality around exposing the models that exist in, in BigQuery, I would love that. But, you know, we also still have a lot of, of Snowflake that Snowflake looker combination was a strong one. I think that in some ways even stronger than BigQuery. Google is serious about building an analytics capability that will work with other things besides just their core components. If you look at the whole way they're going, though, I think that there's a good chance that they will continue to be interoperable. enabling customers to build Google services, first-class Google services with different technologies. You saw this with Redis.

You saw this with Kafka or Confluence. And now Snowflake runs in GCP. That's one of the announcements. So I think it would be great if Looker, if they continue to innovate Tristan in some directions that are just for GCP products, while also just innovating, you know, look at look ML extensions, you know, translate a whole bunch of different SQL languages. If they continue to build that, that core functionality as well. I think that would be great. I will say one last thing is that, you know, I have, I did feel like some of the,

some of the areas that Looker could have been addressing, could have been addressing or not being addressed. And I think that that's been happening, you know, a complete Git integration for everything in, in, in Looker. And I kept thinking that that was just around the corner and it's been two years. So I don't know if maybe the injection of Google engineers will start to address some of those things. But you never know. I've seen this go swimmingly and I've seen this go really badly, these acquisitions. And I just have to, I love both GCP and Looker, so I have to hope that it goes in the right direction. Okay, interesting.

And I suppose back to, I mean, the last thing, I think in your post you wrote recently, Tristan, about the acquisitions, you had open source the next wave of BI, and you were kind of putting the point across, I suppose, that this could actually be an opportunity for open source. What do you mean by that, really? And what was your thinking around that? Yeah, so, and I did have a question mark at the end of that heading in my post.

So there are several BI front-end tools that have some real traction that are open source. Metabase, Redash, and Superset. And Metabase just recently raised an A round. I think it was $7 million. I didn't look that up recently, but I believe that's what it was. And through the grapevine, I think that Max, who is the founder of Superset, has raised some money and Redash has been a commercial entity for a long time. And these products have thousands of, like tens of thousands of commits, hundreds of committers, at least, and I think tens of thousands of stars on GitHub. So these are like really significant projects. And none of them are perfect.

We've used all of them in real client settings before, and really none of them are perfect. But that kind of traction in an open source context, like products have staying power. And we should really anticipate that they are going to be around and they're only going to get better. Yeah, it is interesting, Tristan, in that I think that I'm 100% on board with, I think this being an opportunity. The question for us is like, you know, if you look at like Red Pill and Fishtown, we typically, you know, have a different set of customers. I think there's an overlap in the middle, but we tend to work with a lot of enterprise customers and you tend to work or exclusively work, I think, with venture funded startups. So I think what you see from your customers and what they're willing to try out, I think, and good for you because that's probably a lot more interesting about them willing to kick the tires on some open source software because at the end of the day, saving money means a lot to them. They tend to be just so afraid of venturing outside of something that they can't provision with money.

And that is frustrating. And I think if there's one thing that I would like to see is if the movement of big companies, enterprise companies to the cloud means that they change their way of thinking. At least they start to think about not owning everything inside of their data center and not owning everything with with sysadmins and installations and all the sort of ways they're used to doing things. And if that movement to the cloud means they start to open up in other ways, such as open source software, especially if that open source software is a little easier to use. Right. So as companies like Google make things like Confluent available as a service, I think you'll see a lot more. Nothing makes open source easier than a managed service, right? So I think if we start to see managed services as a way for the enterprise to start getting comfortable with some of these things, I think that it could be really great for everyone. So if we start to see some of these products you've mentioned become available as managed services, as things that are easy to spin up and try out, especially in one of the big three clouds.

I think that it could be great for everyone. But I think as long as they're being installed on-prem, the deck might be stacked against it. I mean, I think the question mark at the end of your statement there, Tristan, you know, open source, NetSwave with BI question mark, I think that's very at the end of your statement there, Tristan, open source, NetSuite, BI question mark, I think that's very telling, really. And I think, you know, I mean, really, I mean, I've had Max Bushman on the show before talking about Superset and the data engineering movement.

There's some clever stuff in there, but it's not kind of changing the paradigm about things. And, I mean, we used to use, we called it meat base in my old company. That is a SQL tool over. I mean, I think the other point you made in that blog post, Tristan, we talked about competition will shift to providing new experiences and talking about,

well, maybe the next BI tool that takes off will actually not just be another spin on kind of dashboarding and so on, but will be something different really. And I think that in a way Looker was that, wasn't it on but will be something different really and i think that in a way looker was that wasn't it look is quite different really to the tools that came before it and you know in a way tableau was quite different to things came before that and i think that to my mind you know what we should be looking for what is the new paradigm and the new experience here i

mean just you know maybe just to finish off you know what was you what were you thinking on that and we'll cut to you at the end um yeah so just on the um on the open source side to close that that thread in my brain um the i i agree that um the existing open source tools uh are are kind of replays of the same experience um and and really we're seeing them uh they're at the stage where i think looker and mode and periscope were in 2013 which is to say like you know very early um but they're going to climb the same the same learning curve and they're going to go through the same process that that the tools that we've just seen uh go through um on the on the new experiences side the one thing that i uh uh that it's like really stands out of my brain is that i'm in in several Slack communities, and there's a lot of people who, even though they use Looker or whatever other BI tool for almost all of their analytics, they swear by a company called Amplitude for their event-based analytics. And Amplitude is just a product that collects the same event stream as any other event tool.

And it includes a storage component. So you query it within some Amplitude-hosted environment. But then, and this is why people actually love it, they have a user experience that is built specifically to query event data. It does funnel analyses really well. It does cohort analyses really well. It's just like built for purpose. And that's the kind of experience that there's, from a technology perspective,

there's no reason why that can't be delivered also on top of BigQuery and in a kind of modular way. But no BI front end has kind of thought of building a best-in-class tool to facilitate a particular experience. Okay, Stuart. So just to found this off, what would you say is maybe the next wave of this stuff coming along? Where do you see the next innovation, the next look are kind of coming from really? Yeah, I think to a certain degree dashboards are going to become less important. I think it's like a, dashboards are going to, are going to become less important. I think

it's, it's like a cup holder in a car. I think you're going to have to be able to, to provide dashboards, but I think they're going to become less relevant. And it's the contextual information, which I think is the new experience that you might've been talking about, Tristan, is different ways that we're going to consume data. A lot of it's going to be mobile in some way. And does that mean a trickle feed through an app that gives me information just in time? Is it a BI experience or an analytics experience that responds to the context my phone's giving me? Is it different ways of consuming data without even realizing you're consuming it? In most cases, it's not even visual. I think that that was one of the great stories about Looker was the built by Looker embedded

Looker use case, which was the full 130% REST API coverage and all the things you could do with it. We never did much with it with our customers because we were always focused on building dashboards. So I think if you continue to build, I think a product with the fundamental architecture of Looker could be extended in some really interesting ways. But we weren't really seeing it done. It's not because Looker couldn't facilitate it. In a lot of cases, it's because the customers weren't interested.

So I think as customers become more interested in these ways of of thinking about tooling um as being more than just a dashboard i think we'll start to see some of that some of that uptake and if whatever tool is able to satisfy these new use cases these new experiences i think that that might be of course the next one to get acquired, right? Okay, excellent. Well, Stuart, Tristan, it's been fantastic speaking to you. It's been a very entertaining kind of conversation. So what we'll do is I'll put the link to your blog post, Tristan, in the show notes and to your company's Red Pill, Stuart, and to mine as well. But thank you very much. It's been a brilliant conversation and hopefully speak to you soon. Thank you.

Great. Thanks, everyone. Thank you.