 from our studios in the heart of Silicon Valley, Palo Alto, California, this is a CUBE Conversation. Hello, and welcome to the CUBE studios in Palo Alto, California for another CUBE Conversation where we go in-depth with thought leaders driving innovation across the tech industry. I'm your host, Peter Burris. The whole concept of self-service analytics has been with us for decades in the tech industry. Sometimes it's been successful, most times it hasn't been, but we're making great progress and have been over the last few years as the technology's mature, as the software becomes more potent, but very importantly as the users of analytics become that much more familiar with what's possible and that much more wanting of what they could be doing. But this notion of self-service analytics requires some new invention, some new innovation. What are they, how's that going to play out? Well, we're going to have a great conversation today with Stephanie McReynolds, who's the Senior Vice President of Marketing at Alation. Stephanie, thanks again for being on the CUBE. Thanks for inviting me, it's great to be back. So, tell us a little bit about, give us the update on Alation. So, as you know, Alation was one of the first companies to bring a data catalog to the market. And that market category has now been cemented and defined, depending on the industry analysts that we talk to, there could be 40 or 50 vendors now who are providing data catalogs to the market. So, this has become one of the hot technologies to include in a modern analytics stack. Particularly, we're seeing a lot of demand as companies move from on-premise deployments into the cloud. Not only are they thinking about how do we migrate our systems, our infrastructure into the cloud. But with data cataloging, more importantly, how do we migrate our users to the cloud? How do we get self-service users now to understand where to go to define data? How to understand it? How to trust it? What reuse can we do of existing assets? So, we're not just exploding the amount of processing that we're doing in the cloud. And so, that's been very exciting. It's helped us grow our business. We've now seen four straight years of triple digit revenue growth, which is amazing for a high growth company like us. Sure. We also have over 150 different organizations in production with a data catalog as part of their modern analytics stack. And many of those organizations are now moving into the thousands of users. So, eBay was probably our first customer to move into the over a thousand weekly logins. They're now up to about 4,000 weekly logins to Elation. But now we have customers like Boeing and General Electric and Pfizer and we just closed a deal with the US Air Force. So, we're starting to see all sorts of different industries and all sorts of different users from the analytics specialist in your organization like a data scientist or a data engineer all the way out to maybe a product manager or someone who doesn't really think of them as an analytics expert using Elation either directly or sometimes through one of our partnerships with folks like Tableau or MicroStrategy or Power BI. So, if we think about this notion of self-service analytics stuff and again, it's Elation has been a leader in defining this overall category. We think in terms of an individual who has some affinity for data but most importantly has questions that they think data could answer. And now they're out looking for data, take us through that process. They need to know where the data is, they need to know what it is, they need to know how to use it and they need to know what to do if they make a mistake. How is that, how are the data catalogs like Elation serving that and what's new? Yeah, so as consumers, this world of data cataloging is very similar if you go back to the introduction of the internet. How did you find a web page in the 90s? Pretty difficult. You had to know the exact URL to go to in most cases to find a web page and then Yahoo was introduced and Yahoo did a whole bunch of manual curation of those pages so that you could search for a page and find it. So Yahoo was like a big catalog? It was like a big catalog, an inventory of what was out there. So the original data catalogs you could argue were what we would call from a technical perspective a metadata repository. No business user wants to use a metadata repository but it created an inventory of what are all the data assets that we have in the organization and what's the description of those data assets, the metadata. So metadata repositories were kind of the original catalogs. The big breakthrough for data catalogs was how do we become the Google of finding data in the organization? So rather than manually curating everything that's out there and providing an end user with an answer, how can we use machine learning and AI to look at patterns of usage, what people are clicking on in terms of data assets, surface then those as data recommendations to any end user, whether they're an analytic specialist or they're just a self-service analytics user. And so that has been the real breakthrough of this new category called data cataloging. And so most folks are accessing a data catalog through a search interface or maybe they're writing a SQL query and there's SQL recommendations that are being provided by the catalog. Or using a tool that utilizes SQL. Or using a tool that utilizes SQL. And for most people and most employees in a large enterprise, when you get to those thousands of users, they're using some other tool like Tableau or MicroStrategy or a variety of different data visualization providers or data science tools to actually access that data. So a big part of our strategy and elation has been how do we surface this data recommendation engine in those third-party products? And then if you think about it, once you're surfacing that information and providing some value to those end users, the next thing you want to do is make sure that they're using that data accurately. And that's a non-trivial problem to solve because the analytics and data is complicated. And metadata is extremely complicated. Because often it's written in a language, it's arcane and intended to be precise from a data standpoint, but that's not easily consumable or easily accessible by your average human being. Right, so a label, for example, on a table in a database might be cust underscore seg underscore 257. What does that mean? It means we can process it really quickly in the system. But it's used as to a human birth. As a marketing manager, right? I'm like, hey, I want to do some customer segmentation analysis and I want to find out if people who live in California might behave differently if I provide them an offer than people who live in Massachusetts. I, it's not intuitive to say, oh, that's in customer underscore seg underscore. So what data catalogs are doing is they're thinking about that marketing manager. They're thinking about that pure business user and helping make that translation between the business terminology. Hey, I want to run some customer segmentation analysis for the West with the technical physical model that underlies the data in that database, which is customer underscore seg underscore 257 is the table you need to access to get the answer to that question. So as organizations start to adopt more self-servicing analytics, it's important that we're managing not just the data itself and this translation from technical metadata to business metadata, but there's another layer that's becoming even more important as organizations embrace self-servicing analytics. And that's how is this data actually being processed? What is the logic that's being used to traverse different data sets that end users now have access to? So if I take gender information in one table and I have information on income in another table and I have some private information that identifies those two customers as the same in those two tables, in some use cases, I can join that data. If I'm doing marketing campaigns, I likely can join that data. If I'm running a loan approval process here in the United States, I cannot join that data. That's a legal limitation. It's not a technical issue. It's a legal government issue. And so here's where there's this discussion in folks that are knowledgeable about data and data management. There's a discussion of how do we govern this data? But I think by saying how do we govern this data, we're kind of covering up what's actually going on because you don't have to govern the data so much as govern the analysis. How is this joined? How are we combining these two data sets? If I just govern the data for accuracy, I might not know the usage scenario, which is someone wants to combine these two things, which makes it illegal. Separately, it's fine, combined, it's illegal. So now we need to think about how do we govern the analytics themselves, the logic that is being used? And that gets kind of complicated, right? For a marketing manager to understand the difference between those things on the surface, it doesn't really make sense. It only makes sense when the context of that government regulation is shared and explained. And in the course of your workflow and dragging and dropping in a Tableau report, you might not remember that. That's right. And the derivative output that you create that other people might then be able to use because it's back in the data catalog doesn't explicitly note often that this data was generated as a combination of a join that might not be in compliance with any number of different rules. Right. So about a year and a half ago, we introduced a new feature in our data catalog called Trust Check. Yeah, I really like this. This is a really interesting thing. And that was meant to be a way where we could alert end users to these issues. Hey, you're trying to run this analytic and that's not allowed. We're gonna give you a warning. We're not gonna let you run that query. We're gonna stop you in your place. So that was a way in the workflow of someone while they're typing a SQL statement or while they're dragging and dropping in Tableau to surface that up. Now, some of the vendors we work with, like Tableau, have doubled down on this concept of how do they integrate with an enterprise data catalog to make this even easier? So at Tableau conference last week, they introduced a new metadata API, they introduced a Tableau catalog and the opportunity for these types of alerts to be pushed into the Tableau catalog, as well as directly into reports and worksheets and dashboards that end users are using. So let me make sure I got that. So it means that you can put a lot of the compliance rules inside Alation and have a metadata API so that Alation effectively is governing the utilization of data inside the Tableau catalog. That's right. So think about the integration with Tableau as this communication mechanism to surface up these policies that are stored centrally in your data catalog. And so this is important, this notion of a central place of reference. We used to talk about data catalogs just as a central place of reference for where all your data assets lie in the organization. And we have some automated ways to crawl those sources and create a centralized inventory. What we've added in our new release, which is coming out here shortly, is the ability to centralize all your policies in that catalog, as well as the pointers to your data in that catalog. So you have a single source of reference for how this data needs to be governed, as well as a single source of reference for how this data is used in the organization. So does that mean ultimately that someone can try to do something, trust check and say, no, you can't. But this new capability will say, and here's why. Or here's what you do. That's right. The scriptive step that says, let me explain why you can't do it. That's right. Let me not just stop your query until you know. Let me give you the details as to why this query isn't a good query and what you might be able to do to modify that query should you still want to run it. And so all of that context is available for any end user to become more aware of what is the system doing and why is it recommending. And on the flip side, in the world before we had something like trust check, the only opportunity for an IT team to stop those queries was just to stop them without explanation or to try to publish manuals and ask people to run tests like the DMV so that they memorized all those rules of governance. Yeah, self-service, but if there's a problem, you got to call us. That's right, that's right. So what we're trying to do is trying to make the work of those governance teams, those IT teams, much easier by scaling them because we all know that the volume of data that's being created, the volume of analysis that's being created is far greater than any individual can keep up with. So we're trying to scale those precious data expert resources. Digitizing them. Yeah, exactly. It's a digital transformation of how we acquire data necessary. And then make it super transparent for the end user as to why they're being told yes or no so that we remove this friction that's existed between business and IT when trying to perform analytics. But I want to build a little bit on one of the things I thought I heard you say, and that is that the idea that this new feature, this new capability will actually prescribe an alternative logical way for you to get your information that might be in compliance. Have I got that right? Yeah, that's right because what we also have in the catalog is a workflow that allows individuals called stewards, analytic stewards to be able to make recommendations and certifications. So if there's a policy that says thou shalt not use the data in this way, the stewards can then say, but here's an alternative mechanism. Here's an alternative method. And by the way, not only are we making this as a recommendation, but this is certified for success. We know that our best analysts have already tried this out or we know that this complies with government regulation. And so this is a more active way then for the two parties to collaborate together in a distributed way. That's asynchronous. And so it's easy for everyone, no matter what hour of the day they're working or where they're globally located. And it helps progress analytics throughout the organization. Well, very importantly, it increases the likelihood that someone who is told you now have a self-service capability doesn't find themselves abandoning it the first time that somebody says no, because we've seen that over and over with a lot of these query tools, right? That somebody says, oh wow, look at this new capability until the screen metaphorically goes dark. Right, until it becomes too complicated. And then you're like, oh, I guess I wasn't really trained on this. And it doesn't get adopted. But this is a way to, it's a very human centered way to bring that self-service analyst into the system and be a full participant and how you generate value out of it. And help them along. So the ultimate goal that we have as an organization is to help organizations become, our customers become data literate populations. And you can only become data literate if you get comfortable working with the data. And it's not a black box to you. So the more transparency that we can create through our policy center, through documenting the data for the end users and making it more easy for them to access, the better. And so in the next version of the Elation product, not only have we implemented features for analytics stewards to use to certify these different assets, to log their policies, to ensure that they can document those policies fully with examples and use cases. But we're also bringing to market a professional services offering from our own team that says, look, given that we've now worked with about 20% of our installed base and observed how they roll out stewardship initiatives and how they assign stewards and how they manage this process and how they provide incentives. We've done a lot of thinking about what are some of the best practices for having a strong analytic stewardship practice if you're a self-servicing analytics oriented organization. And so our professional services team is now available to help organizations roll out this type of initiative, make it successful and have that be supported with product. So the psychological incentives of how you get one of these programs really healthy is important. Look, this is, you guys have always been very focused on ensuring that your customers were able to adopt value proposition, not just by the value proposition. Stafford McReynolds, Senior Vice President of Marketing, Elation, once again, thanks for being on theCUBE. Thanks for having me. And thank you for joining us for another CUBE Conversation. I'm Peter Burris, see you next time.