---
title: "Elastic's Data Governance Revolution: The Role of Active Metadata"
url: "https://atlan.com/masterclass-elastic-recording/"
description: "Governance requires a delicate balance of technology and people. Learn how Elastic succeeds at governance with modern data catalogs utilizing active metadata."
keywords: "Dataops, data catalog, data catalog tools, atlan, data, metadata catalog, master data management, data discovery, data access, data catalog example, data catalog software, data catalog template, data catalog products, data catalog vendors, data lake data catalog, best data catalog tools, data catalog solutions, data catalog products, metadata management, data catalog creation tools"
---

> Atlan is hosting Context Conference, bringing together the leaders and builders at the frontier of giving AI the context it needs to understand their business. It runs online on October 28, 2026, from 11:00 AM to 2:00 PM ET. Atlan co-founder Prukalpa Sankar opens and closes the day. Leaders from AstraZeneca, BNY and Verizon share why they invest in context and what they get from it. Registrants get early access to The AI Context Gap, a new study from MIT Technology Review Insights. Register: https://atlan.com/context-conference/

# Elastic's data governance revolution: the role of active metadata

Events / On demand / Recorded May 19, 2023 / 43:30 / 8 chapters

Watch: https://atlan.com/masterclass-elastic-recording/

Takashi Ueki, Director of Enterprise Data & Analytics at Elastic, on governance for a remote, open-source company: meet people in the tools they use, design for personas, and make ownership real.

## What you'll see

### 01 / Governance for a distributed, open-source company

Watch 3:56-9:21: https://atlan.com/masterclass-elastic-recording/?t=236

Takashi sets Elastic's north star and three pillars, transparency, accountability and engagement, for a company that worked remotely before COVID. Governance has to fit people's day-to-day tools with as little disruption as possible.

### 02 / Choosing Atlan and putting it to work

Watch 9:21-19:54: https://atlan.com/masterclass-elastic-recording/?t=561

After proofs of concept with several catalogs, Atlan stood out for meeting people in the tools they already use. Three personas: a business user checking a new dashboard from inside it, an analyst alerted to pipeline breaks with a Jira ticket, and an architect finding duplicate assets before a migration.

### 03 / Ownership and making it personal

Watch 19:54-24:50: https://atlan.com/masterclass-elastic-recording/?t=1194

With Fivetran, Salesforce, dbt, Tableau, Sigma and BigQuery connected, the next step is ownership: owners with the authority to fix issues and accountability for quality, and a different pitch for Atlan to each group of users.

### 04 / Q&A: playbooks, domains and data quality

Watch 31:13-42:27: https://atlan.com/masterclass-elastic-recording/?t=1873

Jessie Buelteman on tagging assets in bulk with playbooks. Takashi on capturing knowledge from domain experts, keeping ownership with the business, and pairing quality checks with the downstream impact Atlan and EventBridge show.

## Speakers

- Austin Kronz, Director of Data Strategy, Atlan (host)
- Takashi Ueki, Director of Enterprise Data & Analytics, Elastic
- Jessie Buelteman, Enterprise Data Governance, Elastic

## Chapters

1. [0:00 Welcome and audience poll](https://atlan.com/masterclass-elastic-recording/?t=0)
2. [3:56 Elastic's governance north star and pillars](https://atlan.com/masterclass-elastic-recording/?t=236)
3. [9:21 Why Elastic chose Atlan](https://atlan.com/masterclass-elastic-recording/?t=561)
4. [12:58 Atlan through three personas](https://atlan.com/masterclass-elastic-recording/?t=778)
5. [19:54 What's next: ownership and accountability](https://atlan.com/masterclass-elastic-recording/?t=1194)
6. [24:50 Takeaways: people, process and adoption](https://atlan.com/masterclass-elastic-recording/?t=1490)
7. [31:13 Q&A: playbooks, domains and data quality](https://atlan.com/masterclass-elastic-recording/?t=1873)
8. [42:27 Wrap-up](https://atlan.com/masterclass-elastic-recording/?t=2547)

## Related videos

- [Implementing a modern data governance strategy](https://atlan.com/implementing-a-modern-data-governance-strategy/) (41 min)
- [Atlan Activate: 10x your data team's productivity with automations](https://atlan.com/activate-automation-recording/) (47 min)
- [Context Agents demo](https://atlan.com/context-agents-demo/) (13 min)
- [What is a context layer for AI systems?](https://atlan.com/know/enterprise-context-layer/videos/what-is-a-context-layer/) (10 min)

## Transcript

### 0:00 Welcome and audience poll

**Austin Kronz**

[0:00] All right. Welcome everyone, and thank you for joining our May master class. We have an awesome session planned for you today where we're gonna talk about the Elastic team data governance journey and their role with active metadata platforms and, and how they've gone about their process of selecting tools. We're really excited to be joined today by Takashi Ueki, the director of Enterprise Data and Analytics at Elastic. Takashi, welcome.

**Takashi Ueki**

[0:27] Thanks, Austin. Happy to be here, man.

**Austin Kronz**

[0:29] Yeah, of course. Of course. Before we jump right into the content that Takashi has for us today I think quick background on, you know, if you're not familiar with Atlan, just a quick overview of, you know, Atlan, we're pioneering this active metadata space and moving into this AI-led data catalog world. We're really fortunate to have some great industry recognition from Gartner Cool Vendor reports, market guides for Data Ops and active metadata management. You've probably seen the Forrester Wave. We're a top right leader in enterprise data catalogs for Data Ops.

[1:03] And most recently been featured in three different categories as a leader one of which, data governance in the G2 reviews, which we're really proud of, because it's a hundred percent customer driven. So working with great customers like Elastic, like WeWork, Autodesk, Nasdaq, Unilever, News Corp, and, and many others. And we're really excited to host these master classes where we get a chance to show off what Takashi and his team are doing as well as what some of our other customers are doing and open it up to you as an audience to ask questions, because we know topics like governance everyone wants to know what, what everyone else is doing because it's a tricky space, right?

[1:45] We have an awesome agenda. With that being said quickly go over some logistics for the session. We're gonna start with Elastic's approach to data governance and how they are evaluating data catalogs and how they arrived at Atlan. We're gonna talk about how taking a persona-based approach to delivering governance use cases is key, and ultimately where Takashi and his team see the next steps being.

[2:11] Most importantly, we wanna we want to use about ten to 15 minutes at the end for open question and answer, which means please use the Zoom chat or the question and answer feature in Zoom, and we will be fielding questions throughout the, throughout the presentation, but also at the end as well. And we're lucky enough to be joined by Jessie Buelteman from Elastic as well, who will be in the chat and then hopping on to join us at the end for some que... Now hopefully, if you've joined a Atlan presentation before, and I'm doing my job correctly, you know we are gonna kick it off with audience participation.

[2:48] So before we hand it over to Takashi, I, I just wanna know, what is your biggest challenge when it comes to data governance today? There's a lot of different elements to governance, whether it's privacy, security, quality, discovery. We're really excited to kind of hear from you and be able to adapt maybe some of our content and answer some of these questions towards the end of the presentation as well. So we'll give this a couple minutes here or a couple seconds to hopefully get some good answers. I would expect some, some repeats. Yeah. Two right off the bat that I think are, are relevant. Stewardship. Yeah. Awesome. Discovery and metrics. Ownership.

[3:33] Yeah, I think ownership is definitely gonna be a recurring theme here. Seems to be wisdom of crowds here, ownership and, and automation taking a, a quick lead. Well, I think, Takashi I know you have some information on you know, how you are approaching ownership and how you have kind of rolled out a governance program. So this is probably - Yeah... The perfect time to, to turn it over to you.

### 3:56 Elastic's governance north star and pillars

**Takashi Ueki**

[3:56] Yeah. To, to your point, like, ownership's gonna be a big part of what we talk about, and I'm not gonna say that we've solved that. It's gonna be a big part of our next steps. But I, kind of like whoever wrote sins of the past on there definitely plays into a lot of the issues that, you know, we deal with and, you know, if I jump into it, right?

[4:14] I, I did wanna give a little bit of context first before we get into, you know, how we thought about Atlan and our governance program, and just kind of, you know, talk about why I joined Elastic or why this, you know, my role kind of existed to begin with, right? So I've been with Elastic for about two years now, and when I joined, my org, my role, you know, didn't exist. And it, it was, you know, specifically you know, the point of it was to address kind of the problem statements that you see on the screen.

[4:41] And a lot of these kind of what folks, you know, just have on the word cloud, but also I'm sure nothing new to the audience here, right? Disconnected reporting, multiple sources of truth lack of trust in the data, lack of ownership misaligned, you know, BI platforms. All those things kind of contributed to you know, what we wanted to address. So when I came in, I said, "All right, we need some type of kind of mission statement or north star to help address what that is," right? So as cliche as it sounds, you know, our, as a team across, you know enterprise data it's democratized trusted data across Elastic.

[5:16] And that's through enabling the availability and reliability of data and how we define it, how we consume it, deliver and govern it, right? So I wanted to set that, you know, context because when I look at my team, that's like what we're trying to accomplish. Now, a component of that is our data governance program itself, and that's, you know, the, what we're gonna focus on. But I wanted to give a sense of, when I think about it from a data architecture perspective, our BI stack, our enterprise reporting and analytics, anything that we're doing in, you know, underneath or within my organization is going to be pointing towards this ultimate mission.

[5:47] And then obviously, governance underpins a lot of what we're trying to do there and I'm gonna take even one more step back and say, okay, well, now let's think about it from the context of, you know, what environment that we're working in, right? So Elastic itself is a distributed company built on a heritage of open source, meaning even before COVID, we were a distributed company. Everybody, you know, was remote. And with that, and you, you start to think about it from a cultural perspective our kind of cultural statement or source code is, you know, what we like to call.

[6:20] There's a few components of our source code at the top of the screen, not all of, you know, what we have, but some of the ones that are more relevant to this conversation. So the first one that really speaks to, you know, kind of the concept of we are distributed by nature. We do come from a you know, varying backgrounds, whether it is you know, geographically, you know, background experience, right? And then the other two is, you know, around making sure that we're making progress, that we are keeping things simple as best as possible, and that we are, you know, looking to build things from a scale perspective, but also from a relevance perspective.

[6:55] And I think both, both of those are gonna be hugely important to a lot of what I discuss especially from a relevance perspective and keeping things, you know, as simple as possible from an experience. So what does that mean from a data governance perspective? That means, and kind of as Austin kind of alluded to earlier, right? Data governance is not always a, you know, sexy topic that people want to always focus on.

[7:19] And so when you are, you know, focusing on something like data governance where it's not something that, you know, people understand the importance of it, they want it, but they don't always want to necessarily participate in it or do the, you know, hard work of it, right? So when you approach it from that perspective, you know, what we wanted to make sure that we did was we need to meet people where they work. It, it needs to be a seamless, you know, embedded experience in their day-to-day job. If we keep on adding more and more responsibility and say, like, "Hey we are rolling out this data governance program.

[7:47] That means you have to do all these different things, and we're gonna, you know, hold you accountable for this, and you have to check all these boxes," it turns people off, right? And we lose the audience right away. So as we approach it, we wanna make sure that it's a part of their day-to-day experience, and it's as least disruptive as possible. And so with underneath that, you know, there's kind of three, you know, key pillars that we wanted to make sure that we focus on: transparency, accountability, and engagement.

[8:13] Transparency in, in terms of people knowing, you know, where the data is, who owns it, and and just being able to understand the data that exists within Elastic. Accountability, kind of reflective of the word cloud of knowing who the ownership is and who the stewards are, right? And holding those individuals accountable. And then engagement in terms of making sure that we have the right engagement across our business stakeholders as well as our, you know, technical IT counterparts as well. I, I think one of the misconceptions you know, the people have is like, "Hey, we have a data governance team. Go fix all of our problems, and you're now responsible for it."

[8:51] And, you know, what we wanna make sure that we, you know, helped out, you know kind of correct in, in that thinking is it, it's not a central dictate of, you know, myself and Jessie and the rest of the team saying, "This is what we're going to do." Our role is to make sure that we have the right framework, that we have the right guidance, that we are, you know, establishing the right ownership and stewardship, that we're engaging with them and collectively working towards a, a more governed environment, but more - and driving towards more quality data across the board.

### 9:21 Why Elastic chose Atlan

**Takashi Ueki**

[9:21] So with that in mind, you know, we said, "All right. We know we're going to need you know, a data catalog or, or something that's going to help us in this, you know, data governance journey." And when we started to look at it, we said, "All right. Well, we know that there are some legacy players out there that are tried and true. We know they have the capabilities. They've been in the market for a while. And then we also know that there's some new entrants, you know, coming in the kind of modern data catalogs or version three dot o, whatever, however you wanna describe it, right? Atlan being one of them."

[9:51] So what we did was we, we took kind of all these different, you know, vendors and said, "All right. We're gonna do a pretty thorough evaluation." And I think we did POCs with just about every single one of them. Honestly, like, they - every single one kind of hit the core capabilities that we're looking for. They all check the right boxes of, like, if I want a data catalog or, you know, a data governance, you know, tool. But there was kind of a another set of checklists that we also had, and this is kind of what I'm showing on the screen of does it promote usability or adoption? Meaning, we don't want just shelf wear.

[10:26] We want people to actually be engaging in it. When someone goes into the tool, are, are they gonna enjoy the experience? Are they gonna understand how to use it? So that, you know, goes into the clean UI, automated data metadata management, all these things. And then I kinda skipped over the first one. I want to kind of highlight that one last in terms of making sure it integrates with our primary enterprise applications. And this was hugely important for us, right?

[10:51] So if I think back to who Elastic is and the fact that we are distributed by nature, that we are all, you know, working remote, and I think a lot of companies are in this, in this space now too. If you think about us being in the office, you know, in person, if I have a question about, "Hey, what is the definition of, you know, this metric?" Or, "Hey, can you quickly approve you know, this workflow?" Whatever the case is, you just turn around and go to your colleague in the next cubicle or wherever they are, you can have that live conversation. That's your workspace.

[11:22] If you think about from a, you know, more distributed or remote perspective, our desktop is effectively our workspace or our cubicle, right? So if I'm introducing another application that people need to use and it's not integrated with the experience that they currently have, whether it's Slack or Jira or, you know, BigQuery or Tableau or whatever the case is, it just adds more clutter to your desk, right? And it gets lost in the shuffle. So I want - we wanted to make sure that whatever we chose was integrated with what the tools and the applications that, you know, are Elasticians are already using to make that a more seamless experience.

[12:02] And so when we kind of went through this checklist you know, Atlan was really the one that kind of sh-stood out and shined in a lot of ways. So that's, you know, how we end up going with Atlan for, you know, for this purpose. The other part that I think, you know, really stood out too was just the partnership that we had.

[12:18] From beginning to end, from the first conversation I had all the way through, you know, from a procurement and, you know, going forward, you know, over the last nine months, it truly has been a a partnership in terms of not just, "Hey, we're gonna sell you a product and, you know, en-enjoy your day," right? It was more so, "What are you trying to accomplish from an Elastic perspective? How can we help? What are the features that we could, you know, potentially enhance to drive our ultimate objectives?"

[12:43] And so because of that partnership, it's been, you know, we've been able to make a lot of progress, and we'll kind of walk through what that looks like in the next few slides. So you're probably wondering, okay, that's great, we, we selected something, now what does it look like?

### 12:58 Atlan through three personas

**Takashi Ueki**

[12:58] I wanted to kind of approach this through the lens of a few different personas, right? You know, the first one being i-if you're a business user and you say, "Hey, someone just sent me a new dashboard. I don't really know what it is. I don't know what the metrics are or where it's sourced from, or if I have any questions who I should talk to." O-one of the things that, you know, really has helped us a lot is the Chrome plugin, right? And again, this goes back to the concept of meeting people where they work already.

[13:25] If I say you have to go to Atlan to figure out the definitions or better understand information about this dashboard every single time, it's a disjointed experience. But if I say, "Okay, our experience now is you have a dashboard. You automatically know that there's a Chrome plugin. You can pop that open if you have any questions about the content on the screen," that makes it a more seamless experience. It puts the information at the fingertips of the user.

[13:51] So it doesn't, you know, again, it doesn't make it a disjointed experience, and also drives the, you know, the point of, okay, this is our single source of, you know, what the definitions of the metrics or the data is and ensures that there's consistency across different dashboards and users. The, the next persona, you know, that we kind of focused on was more this kind of data analyst persona. And this one was, you know, pretty key for us in terms of, okay, we have these different pipelines there's an issue with it. Right now, the experience is something broke.

[14:30] You know, we have a Slack channel with all these different users in it, and then someone says, "Hey, I'm looking at, at the data on a dashboard or on BigQuery," or whatever the case is. "The trending looks off. Something's wrong here." And then we go scramble to go figure out what it is with, you know, our various technical partners, right? It-it's not the best experience, and what it also does is it breaks the trust in the qual - in, you know, the reliability of our data, right? If every single time our users are the ones flagging that something is wrong, then that tells them they can't always trust the data itself.

[15:02] So what we wanted to do was make sure that as, moving it as far upstream as possible, we want to automate you know, those pipeline breakages and be able to notify from a monitoring perspective, all the downstream impacted assets. So if something breaks in terms of our Fivetran, our Fivetran pipeline is not bringing in things properly or it's gone stale for, you know, a set amount of time, it triggers, hey, now we are going, you know, all downstream impacted assets, whether it's a BigQuery table or a Tableau dashboard, those owners are now alerted. We have a notification on those.

[15:39] That experience, you know, we haven't kind of put all those pieces together you know, kind of end-to-end in terms of you know, we'll be introducing dif-different dbt you know, monitoring and checks as well as within Fivetran and making sure that we can inte-integrate those properly, so those automatically trigger those warnings. And then also then can create a Jira ticket so that we can track progress on it. So that's kind of the end-to-end vision of what we're looking for.

[16:05] But at a minimum, it allows us to kind of kick off that notification directly within Atlan, and then ensure that all the downstream impacted assets are notified ahead of time, and not them coming to us, you know after the fact. The last one, and I know I put kind of data architect persona, but there's a few different ways that you can think about this in terms of popularity metrics and kind of, you know, we were using this a lot in terms of saying, okay, we have a pretty what is it? Someone said sins of the past, right? In the word cloud.

[16:43] Our current data warehouse obviously has been built up over time, and we need to add more structure to it and, you know, kind of clean it up. And you know, that's part of the journey that we're on. So when we look at it from that perspective, we say like, okay, well, what is actually being used? Do we actually need to m-migrate it over into a new cleaner environment? So this actually kind of helps us understand, you know, i-if there are kind of ambiguous or duplicative assets, which ones should, you know, are the ones actually being used? How often are they u-being used? In what context is it being used?

[17:14] And who's, and who's it being used by? So this has definitely helped us kind of rationalize, all right, where can we actually do this cleaning? The other side of it too is if we know what assets are actually being used within our warehouse, it also kind of helps to rationalize what we're bringing over.

[17:32] So if you think about it from a Fivetran perspective, we just say, you know, if we had set it up historically of like, all right, effectively select star, just bring in everything on this object from Salesforce into our warehouse, and then we realize nobody's actually u-using that object or, you know, twenty of those fields, we can reduce that, reduce the compute from a Fivetran perspective and save dollars there too. So it's been, it's helpful in a lot of different ways from that perspective.

[17:57] The other side of it too is if you think about it as a, you know, data analyst that's coming in new to the the business and you say, "All right, I need to go look at our opportunity data." You search in Atlan and you're like, "Wait We have three different ones. Which one should I be using? If you have the popularity metrics thing, I'd say like, "Oh, well, this is actually the one that people mostly use." You know, or, "Hey, this is the one that's being used by my team," right?

[18:20] Now, realistically, we'll get to the point of where you don't have that ambiguity, but in, you know, as we kind of move into that cleaner state, we at least have, you know, the ability for people to kind of self-serve themselves and understand which assets that they should be using. And so, you know, kind of wanted, you know, as I kind of close out, wanted to speak to what has that looked like for us, right? You know, we've been on this journey for about, I think, nine, ten, ten-ish months or so at this point.

[18:49] And we've made a lot of progress in terms of, you know, making sure that we are integrating with a lot of our, you know, different systems. So you see on the left-hand side, you know, Fivetran Salesforce, dbt, Tableau, Sigma as well as BigQuery. You know, we made sure that from an, a Okta SSO perspective, we-we're integrated. We have integration with Slack, and then on the right, you kind of see like, okay, some of these top-level metrics of you know, how many... Like, on the bottom one, right, as I mentioned about the Chrome plugin, we have five hundred and seventy-two, you know, chroming session views.

[19:20] We really wanna drive that because that's kind of, you know, how we're going to get more and more people engaged with Atlan itself. So again, the way, you know, the experience that we've had in terms of onboarding with Atlan, a lot of the native plugins and a lot of the, you know, kind of automated metadata scraping that, you know, is just inherent within Atlan itself has enabled us to kind of get a lot of this information in, right? So with all of that kind of now in place, you kind of get to the point of, okay, well, what's next for us?

### 19:54 What's next: ownership and accountability

**Takashi Ueki**

[19:55] And this is where I'm gonna take a step back and say, you know, think about it from an overall data governance perspective. So when I talk about this, you know, this slide, it's not just specific to Atlan itself. You know, as thinking back to that word cloud, you know, ownership was the largest word out there, right? So ownership and accountability, what does that mean? You know, as we started on this kind of journey with Atlan and kind of our governance program itself about a year, year and a half ago, we identified who the owners of different, you know, asset, or data assets should be.

[20:27] And we got a lot of nodding of the heads of, "Yes, yes, I agree, I'm the owner of this asset. Yes, yes, I agree, I should be the steward of it." Nothing actually improves, right? So it's like, okay, well, how do we move the needle on that? And, you know, a lot of the things that we're thinking about as we move into our next fiscal year that started, you know, this month, is driving the ownership and accountability through formalized, you know, data contracts, saying, "Okay, well, you are you are the data or the business data owner of these data assets because you fit into this criteria.

[20:57] Great, you have, you know, that authority, but that also comes with, you know, these responsibilities," right? So making sure that we're able to hold folks accountable for the quality of the data, that they're actually truly owners of it, that they're assigning the right stewardship of it. So then we have that set kind of contract in place. We've established kind of the expectations that we believe from a central kind of governance perspective, what that looks like, and then also then now setting up ways to hold them accountable. When we think about a, let's just, you know, say a given field, right?

[21:30] It's like, okay, well, right now it's seventy percent complete and accurate on as of today. It should be a hundred. Okay, well, how are we actually progressing in terms of are we improving in that space? So making sure that we have the data to kind of point to, to say it's improved over time or it's kind of, you know, stayed flat or, you know, gone down, and using that to then hold those folks accountable, right? So being very much more targeted in that way. The other one kind of, you know, it says make it personal or, you know, these different personas of how to think about Atlan.

[22:05] When you think about Atlan, and even I struggle with this sometimes as a, you know, Atlan evangelist, right? What is Atlan, right? It's like, well, there's so many different things. There's so many different angles that you can think about from an Atlan perspective and, you know, you think about the event next month for Atlan AI. There's just, like, so much more, and it's gonna be you know, such a powerful platform. And that's important to us and our team and, you know, a lot of our users. But trying to then position it for our business users, it doesn't always resonate, right?

[22:34] They don't care that, like, oh, we have a metadata, you know you know, catalog or whatever, you know, however you wanna describe it. It's like, well, what does that mean, right? It-it's kind of like if I think about my mom and I said, "Okay, I have Excel. I use it all the time. You... I can create pivot tables. I can do this, I can do this, formulas, and all this great stuff." She's like, "Yeah, I just need something to add up a few numbers." Like, "Oh, you just want a calculator. Oh, okay, well, E-Excel is y-your calculator," right?

[23:01] So it's kind of, you know, figuring out different ways that we can market Atlan to different users to kind of strip down what's relevant to them. So from a persona perspective, I say, "All right, we have different, you know, business users say Atlan is effectively our Tableau catalog. As you come in and you wanna know what all reports and dashboards are there, here's how you would experience it," right? "Oh, that makes sense to me. Okay, now I'll go there." Or if it's a new exec coming in and say like, "Hey, how do we measure the business? What are our key metrics across the business?" "Oh, Atlan is our metrics catalog. Here, go here.

[23:39] This is how you would experience it. You can click through and see all the different metrics that we have." So it's just a one-stop shop just for that. Then they say, "Oh, great, now I know this. I know the definition, and I can click into it and then see all the associated reports where I can actually, you know, analyze that metric." So it, it changes the way that we kind of are positioning Atlan for different user groups and making more relevant to them specific to their, you know, actual needs. And then the last one is, you know, kind of putting it all together.

[24:09] If I, you know, think about it from our team perspective, what we did a lot over the last year was build, you know, bringing a lot of the component parts that we need for our team to be successful. Now it's how do we kind of make it all, you know, fit all together in terms of dbt, Elementary, Slack, Sigma, Tableau, Jira. How do all these, you know, individual component parts work together? How do we make a, a seamless experience to to maximize the ROI for what we're trying to accomplish from a enterprise data perspective? So a, a lot, you know, that we're gonna be doing for this you know, upcoming year.

[24:43] And, you know, would love to, you know, connect with more of you all just to kind of get a sense of like how you guys are approaching some of these same problems too.

### 24:50 Takeaways: people, process and adoption

**Austin Kronz**

[24:50] Yeah. Awesome. I, I really appreciate it. I mean, I think you made a lot of really good points about really personalization of users, user experience, right?

**Takashi Ueki**

[25:00] Yeah...

**Austin Kronz**

[25:01] I think you also make the connection too that the lack of that flow in user experience is not the inherent problem, it's the trust that gets broken when you have Yeah... Constantly ask someone to, you know, go out of their day-to-day workflow. So I think it's, it's really important. The ownership and accountability, definitely a recurring theme. We come back to it all at the end as we had in the word cloud. So maybe just one more question for the audience here. You know, how, how are you establishing accountability or, you know, what does data ownership mean? This, this won't quite be a word cloud, more so a running free text field.

[25:36] We'll leave this open too, so even after we move on to the, the Q&A portion, it'd be great to just see, you know, your take on what ownership means and how we drive accountability. 'Cause I think it's something that the, the space in general is hungry for. I know in Atlan, right, you can have multiple types of owners, right? But the, the what does that mean for them is also a, a very important piece too. So, yeah, producers of content. Yeah. Yeah, I think there's a, you know, probably a balance between stewardship, data producers, the business users. Yeah. Voluntarily. Yeah, being voluntold. Yeah. Well, we can leave this up. Takashi, go ahead.

**Takashi Ueki**

[26:23] And also, yeah, to that point, you know, I think Jessie and I and the team have probably gone back and forth I don't know how many different times about what the right way to identify ownership specifically even in Atlan and what other, like, custom fields we can add in to, like, provide that color and context. And then you get to the point of like, wait, we have too many roles assigned to this. It's actually now more ambiguous about who actually is the owner and what the responsibilities are, right? Like, let's actually just strip it down. Put one person there, that's it.

[26:57] Now we know, okay, well if there are any questions just go to this person or, you know, their designate, right? I, I think we were trying to thread the needle a little bit too much, and now, you know, when you as Anand said, you know, so you kind of have to like just long told of like, all right, well, you are the producer, you are the owner, now this is the responsibilities that come with it.

**Austin Kronz**

[27:18] Yeah. I think there, there's probably a whole webinar we could do on this, the concept of like a data product manager almost being that single point of contact that determines is this a question for the technical owner or is this a question for the business owner, right? And they almost become the liaison between all the different stakeholder groups, much like any product manager would be. So I think that push towards data products is something that we might see change, change this paradigm a little bit.

**Takashi Ueki**

[27:46] Yeah.

**Austin Kronz**

[27:48] Awesome.

**Takashi Ueki**

[27:48] Because you also want to make sure that whoever the owner is that's assigned has the right sway and impact if something is wrong. 'Cause usually, you know, like when people say like, "Oh, there's a data issue," I always kind of tell them, it's like, well yeah, the data, data itself can't just be wrong and bad. It's like data is a byproduct of a, a system, a process or a policy, right? So who is at that, you know, who has that authority, who is in that role that can kind of shape the right, you know, policy or process change or work with, you know, has the right influencer relationships to change from a system perspective.

[28:21] That's really kind of what you want to make sure you have in that ownership perspective.

**Austin Kronz**

[28:26] Yeah. So they need to it's, it's like identify the authority, but also the authority to remediate the issue, right? To, to remedy the issue. So, Yeah... Awesome. We'll, we'll keep this up in the background, so if you have the link, keep, keep giving your responses there. Just to wrap up so we have time for the, the question and answer, 'cause we have a couple that came in. Takashi, tell me if this is a fair summarization too, right?

[28:50] Selecting data governance tools is both functional requirement, but also you mentioned, you know, finding a partner, finding someone that blended, you know, the vendor principles with your own company's principles and kind of moving at the same pace. User experience, we just talked about that a lot, right? The user experience for a variety of data personas. It's kind of hard to say even who is on a data team these days, right? I mean, everyone kind of interacts in different ways. So catering to all these personas you know, the variety that we have is, is extremely important.

[29:22] And then the last piece, which I think probably spend a lot of time on as you mentioned, you're getting to this phase now too, where, you know, there's the technical setup of, you know, the metadata management, active metadata, data catalogs, all the tooling, right? But there's this whole people process change management journey to ensure the actual adoption of the policies. So focus equally on, on both. Probably plan ahead there too.

**Takashi Ueki**

[29:47] Yeah. And on that first point, like when I, when I talk about like kind of, you know, in the context of like meeting people where they are, it's like that's also like the company itself. Our jour - We were at the beginning of our journey, which means I didn't necessarily need a, you know, a tool that had all of these full suite of, you know, features that have been built up over decades. It's just too heavy and clunky for what we need, right? There's like very basic things we say like, "We absolutely need to have this."

[30:14] And the great thing about, you know, working with, you know, Atlan or like these, the, you know, companies that are on the more cutting edge of it We get to partner together to, like, shape, okay, well, it doesn't have to be the same experience as what all the other, you know, legacy players have had. It's like, this is kind of what we want to accomplish. What is that then now going to look like? And being able to provide that feedback and be on the journey with Atlan as well in that context.

[30:38] So it's really just understanding what appetite, where we are in our journey, but also what appetite for, like, kind of disruptive change do you have? You know, there may be other orgs out there that say like, "Hey, no, I just want this. I just want a data catalog. That's all I want. Just give it to me because that's what I'm used to." And, you know, it's going to be, you know, a, a more rigid kind of mentality, right? I'm sure Atlan has a playbook to kind of just, you know, displace those and kind of disrupting that, you know, in those areas as well.

[31:03] But for us, it was like, I want someone that is going to be able to be a little bit more nimble and kind of work with us to shape the experience that we think we need.

### 31:13 Q&A: playbooks, domains and data quality

**Austin Kronz**

[31:13] Awesome. I think transitions to some of the, the questions, we can maybe get into the, the Q&A portion now too. So feel free, just a reminder to the audience, submit them in the chat or in the question and answer feature. But one of the, the questions I had, and I think we had a question too about setup and, you know, connections and things like this. And Jessie, I know you're, you're here as well, so feel free to pop on screen with us. But there he is. Talk about what it was... I guess, what was the setup like?

[31:45] You know, how quickly were you able to get up and running in terms of, like, technical setup at least? And then, you know, the question specifically was about APIs, you know, Atlan being an API-first platform, but we have dozens of out-of-the-box connectors. I mean, what, what did that first part of the journey after selecting a tool look like, Yeah.

**Takashi Ueki**

[32:05] Y-yeah, I mean, it's... One, it was pretty seamless. It was pretty quick in terms of getting, you know, connected to our different data sources, you know, right, kicking off the kind of the scraping of our metadata and getting all that into Atlan, right? Or even if it's like, "Hey, we have these," a quote unquote data catalog that was maintained in a Google Sheet file, and then just kind of importing all that in. Like, all of that, like of just getting...

[32:29] And we did a lot of that ahead of even before signing with Atlan to say like, "All right, let's just start gathering all this information so that once we, you know, hit go, we're able to now bring it in, you know, very quickly." And Jessie, you can add onto this, but I think it was like a four to six-week kind of timeframe of like, all right, let's just work to get all the right technical, you know connections set up, bring in what we have, bring in the right, you know, users, all that kind of information before we started, you know, opening it up to a lot of other users.

**Austin Kronz**

[32:58] Awesome. Question about just automation, right? Now that you're, like you said, nine, ten months in or so into the journey, I mean, when a new use case is arise, you mentioned playbooks. I think, Jessie, we had a good conversation about the Atlan playbooks. I mean, what role does automation play in the governance program now overall? And I guess, how does that look like whenever you're, you know, implementing new policy or something like that within the platform?

**Takashi Ueki**

[33:24] Yeah. I'll let Jessie talk about more specifically on the, you know, the playbook side of it. But, you know, one area that has continued to come up that we weren't able to readily address before we had Atlan with the lineage that we have is a lot of situations where our CRM team is like, "Hey, we're going to be making these changes in Salesforce. We're going to be deprecating these, you know, different fields, or we're going to be updating, you know, the logic behind this." And they're like, "All right, we're going to go ahead with that. We're going live." And it's like, hold up.

[33:53] With kind of the impact analysis that we can do directly in Atlan, it's like a five, ten-minute exercise of, all right, here's everything that you're now going to be impacting downstream within BigQuery, within Tableau, within Sigma. Have you talked to the owners of those people of, are they okay with these changes? And, you know, it actually put a pause on something they were going to roll out you know, for a couple of weeks, and then now we're in the middle of another one that we're kind of going through. Like, do you understand what the impact is to your to our user base if you go through with this?

[34:23] So I think that's, you know, hugely important for us. And then, you know, Jessie, I'll, I'll let you chime in on the playbook side.

**Jessie Buelteman**

[34:28] Yeah. Thanks. Yeah, for the playbooks when we first integrated with Atlan, there wasn't a way to do mass curation. So actually, what we're doing, we're partnering with Atlan, and they were extracting all our assets essentially through an API and providing that through a, through a G Sheet or an Excel spreadsheet. And then we could go and add owners, descriptions, provide that back, they would upload it. So with the implementation of playbooks, it, it allows you to do that. It's just real simple command center way to do mass curation.

[34:59] So we use it a lot for owner assignment classification tagging, whether it's for pers - like PII information or even we created some custom classifications to highlight our enterprise-certified data sets or our highly used visible dashboards from Tableau. So we, we use that functionality quite a bit, and it takes a lot of pressure off of doing that manually through the API export or even going in and finding the assets individually and going through and tagging those. So pretty big time saver there.

**Austin Kronz**

[35:35] That's great. That's great. Couple questions in the chat coming in. We talked about technical, I guess, rollout, right? What about on the people side? There's a question here from Alan about do you already have data domains, right? Or, you know, how do you formalize this data stewardship? Mentioning data contracts here as well probably related. What, what is the rollout from a people? Do you pick one or two teams, one or two domains? What - how did you approach it?

**Takashi Ueki**

[36:05] We kind of have approached it... We, we've approached it in a number of different ways. Part of it was just given some turnover in some of our some functions where a lot of the folks that actually had information or knew about you know, what a, you know, description should No longer were with the company, so I was like, "Okay, well who who can we just chase to just go get that information," right? But I, but what I would say is like, you know, how we started to target is more so what are the...

[36:37] Thinking about it from a business perspective, what are the metrics that we all kind of, you know care about from a sales perspective, a finance perspective, marketing and product perspective, right? A lot of that is around our, our bookings, our telemetry and usage data. Okay, if that's the case what information is relevant to those, that are key to those, and then kind of working backwards to, okay, well who are the owners for, like, who are the domain owners in those areas?

[37:06] And then having a little bit, you know, more targeted conversations with them of like, "Okay, well, your opportunity object, which has, you know, 600 fields in, in, you know, Salesforce," when you just clean this all up, it's like, well, that's not a very productive conversation, right? It's like, "Okay, well you're the owner of this. You're responsible for all of this, but here's the, you know, let's just say 20 to 25 fields that we actually care about that are critical to all these different users. We need to focus on that," right?

[37:31] So it's just making sure that when we have those conversation with those owners, that we make it a little bit more targeted and a little bit more palatable of what we expect from them.

**Austin Kronz**

[37:41] Makes sense. So, so there's already some... Kind of take the knowledge of the organization, right, and codify it a little bit Yeah... Inside of there. Yeah, yeah. I guess related to data contracts, I don't know if it's something you're exploring or even just, you know, this whole shifting governance to the left kind of principle, but is that something Elastic is thinking about or that's, I was gonna say, playing around with? Yeah.

**Takashi Ueki**

[38:05] Yeah, definitely. You know, we, we definitely want to make sure that the ownership and the governance sits with the business, you know, and who is actually responsible and accountable for the data itself, right? 'cause a, a lot of times what you end up with having is like, well, like I mentioned before, it's like, "Okay, well you're the owner," but they don't want to actually do the hard work of, you know, stewarding the data properly, right? They want the power of saying like, "Oh, well, I get to dictate what, you know, goes in and what the definitions are, this, this, and this, and who has access," all that kind of stuff.

[38:38] But okay, well what about the quality? It's like, "Well, aren't you guys doing that?" It's like, "N- no, you're actually accountable and responsible for that," right? So it's definitely making sure that we move, you know, to your point, Austin, as much to the left as possible and as you know, as directly into the business and really making sure that we are identifying the people that truly are accountable, that are willing to be accountable for it and that we are very clear on how we're going to hold them accountable for the quality of it.

**Austin Kronz**

[39:07] Just jumping off of the, the quality, there are a couple questions about quality, observability and I remember some of our conversations about this. But you mentioned dbt tests and some other things. I'm sure there's other checkpoints along the pipeline, but how, how are you approaching that? Or, or you know, what role will something like dbt tests versus looking at an separate data quality or observability platform how does that factor into the governance picture?

**Takashi Ueki**

[39:33] Yeah. I, I think on that end yeah I think in the chat I was like, "Why are we building it from scratch," right, versus buying something new. And this kind of goes back to the, my comment earlier about where we are in our journey. There are, we are at a point where it's like, okay, there are just basic things that we need to be, you know, monitoring and observing for, which we can accomplish with dbt and Elementary packages and then surfacing in different ways, right? These are very basic things. We can get to the point of like, okay, do we wanna make it more sophisticated? Do we wanna observe in, you know, more different ways?

[40:05] It's like, well, we'll get there, but right now we need to make sure that at least, you know, our data is not stale, that it's actually coming over properly, that, you know, there's certain completeness or accuracy or validity on certain fields. Very basic stuff, right? It's like, but we need to make sure we get into that rhythm and that practice. It's not just on us of like, "Okay, well could we do more?" It's like, yes. Well, going back to the conversations with the business data owners, do they have the appetite to consume all of this, right?

[40:33] It's like we say like, "Hey, we just signed a contract with, you know, Monte Carlo or Bigeye, and here's your dashboard. We're observing all this stuff. What are you gonna do about it?" And they're like, "I, I don't even know what any of this means," right? So it's like, okay, well, we have to have, go on this journey with our business data owners to say, "Here's very basic stuff. This should be 100% complete. It's not. Go fix this." It's like, "Oh, okay, I understand that," right?

[40:56] So I think that's kind of where, you know, where we are in, in our journey in this space of observability of our data and kind of monitoring for, you know, the quality is like, you know, start simple, make it digestible, and then as we move forward and make it more sophisticated, we're always gonna have that conversation of do we need something else that's, you know, going to play better in this space?

**Austin Kronz**

[41:16] That's really helpful. Yeah. I, I think the, the sort of marriage of the visibility, so something like Atlan may provide, plus whatever quality observability checks you may run, is also really important - Yeah... Kind Yeah... Report on Yeah... And really automate it. Yeah. Good.

**Jessie Buelteman**

[41:33] Yeah, I think the question also, it, because I think they saw the slide where Atlan we were using Atlan and EventBridge to show all the downstream assets. So that's one thing like that dbt and Elementary doesn't necessarily provide. It can say, "Oh, your pipeline broke," or, "This table didn't refresh in time." But where Atlan comes in and that, that connection there is, okay, maybe a raw table or something broke, but what does that mean to all the downstream assets, right? So what does that mean for my Tableau dashboard or the report that I'm using? And Atlan can use the lineage, lineage that it scraped and say, "Look, this upstream table broke.

[42:08] Now this impacts the end users," and can put that notification or that announcement on the asset within Atlan or send that Slack message to to the, you know, subscribed users, which is important to let them know, right, of the, the failure upstream and that they might not know about otherwise.

### 42:27 Wrap-up

**Austin Kronz**

[42:27] Yeah. That's awesome. Well, I know we're, we're coming up on, on time here. I wanted to just again, say, say thank you to the Elastic team, and it's been great working with you through all this. I know we had a chance to present at the Gartner Summit in the US now doing this. So it's really great to see the progress, and I'm excited to see what the next 10 months with Atlan and the governance journey has in store for you. So, Yeah... If you'd like to get in touch with us, feel free to follow us on, on LinkedIn Atlan, Takashi, Jessie, myself. We have our Atlan Activate coming up on June 15th.

[43:03] I hope you've seen the videos everywhere around LinkedIn and social media of Atlan AI, the trailer, if you will. Well, the full movie will be coming out June 15th so please tune in to the Atlan Activate to see the, the Atlan AI preview. Join the wait list to get access to an early beta, and we really appreciate all your time. So thanks again, and we'll see you next month for another masterclass. Thanks.