The Data Stack Show - 173: Data Analytics Is a Team Sport, Featuring Jay Henderson of Alteryx

Episode Date: January 17, 2024

Highlights from this week’s conversation include:No Code Analytics (1:22)Analytics as a Team Sport (2:31)The workflow of someone without Alteryx (11:27)Alteryx's ability to handle diverse data sources (14:32)The balance between ease of use and complexity (23:06)Enabling casual end users with a no code interface (24:19)Taking analytics to the data (31:47)The boundaries between data engineers and end users (33:44)The importance of collaboration in analytics (34:12)The potential of every employee being a data worker (35:28)The human nature of the product and users in large enterprises (00:45:38)Final thoughts and takeaways (46:21)The Data Stack Show is a weekly podcast powered by RudderStack, the CDP for developers. Each week we’ll talk to data engineers, analysts, and data scientists about their experience around building and maintaining data infrastructure, delivering data and data products, and driving better outcomes across their businesses with data.RudderStack helps businesses make the most out of their customer data while ensuring data privacy and security. To learn more about RudderStack visit rudderstack.com.

Welcome to the Data Stack Show. Each week we explore the world of data by talking to the people shaping its future. You'll learn about new data technology and trends and how data teams and processes are run at top companies. The Data Stack Show is brought to you by Rudderstack, the CDP for developers. You can learn more at rudderstack.com. We're here with Jay Henderson from Alteryx. Jay, welcome to the Data Stack Show. We are thrilled to have you on. Hey, Eric. Great to be here.

All right. Well, give us a quick background and then tell us what you do today at Alteryx. Yeah. Hey, everybody. Really great to be on the show. I'm the SVP of products here at Alteryx. I lead product design and strategy across all of the different Alteryx products. I've got 20 years of product experience at mostly analytics companies. Before this, I was at IBM and I ran the Watson marketing division. I worked at SPSS, predictive analytics company, Clear Forest, a text mining company. So really spent my whole career in the analytics space being a product guy. And so pretty excited to talk you guys through some of the things we've got going on here at Alteryx and to talk about analytics and what's happening today and some of the things that are coming in the future.

That's awesome. By the way, Jay, one of the ways that I think about AlterX is that it's kind of one of the most successful examples of a tool or a product that's no code or low code, as we call it today. But AlterX has been doing this for a while. And I think it's a great opportunity for us to get deeper into that. Why we need that, how it plays well with, or doesn't play well maybe, we'll see with other tools out there. And also what the future looks like in the AI world, where it's kind of like the definition of like no code at the end. You don't even have to type, let's say. You can speak to your assistant and like do something magically. And I think we have the right person here today

like to talk about these things. So I'm really like excited about it. But how about you? Like what's in your mind and what are a few topics that you'd like to go deeper into today? Yeah. I mean, I think you said it. I think no-code analytics will be really fun to talk about, particularly as a way to democratize analytics to hundreds or thousands or tens of thousands of people in your organization. The scale can get

really interesting and really impressive. I'm also really interested in talking a little bit about how analytics can be a team sport and how the different people doing different types of analytics maybe can work together. And then I'd love to dig in a little bit around generative AI and some of the exciting things that are coming down the pike based on all of the crazy new algorithms that seem to be getting updated on a pretty regular basis here. That's awesome. What do you think, Eric? I think we should go and start talking. What do you think? Well, let's do it. Let's dig in. Let's do it. So much to talk about today, especially Alteryx being maybe one of the biggest companies that many people haven't heard of. So definitely want to dig into that. But give us your background.

How did you get into data initially and what was your journey to Alteryx? Yeah. I mean, I started off my career in one of the most exciting fields possible, which is accounting. If anybody on the show has ever been an accountant, maybe you'll agree, maybe you won't. But for me, that was really the doorway into doing things with data, getting involved with technology and technology systems and pretty quickly, you know, realized that I didn't really want to be an accountant, but all this data stuff was pretty cool and pretty fascinating. And then made my, my way to a bunch of different data companies over the years, whether it was sort of dealing with data from one particular type of system. I worked at a web analytics company dealing with particular types of algorithms. I worked at a place that made data mining and statistics software.

Later on, a text mining company, then focused on analytics in one function. So marketing analytics and customer analytics. And then found myself here at Alteryx, the natural landing spot where it all comes together, being able to do analytics against all kinds of data and different techniques. And it's pretty fun and pretty exciting stuff. Very cool. And I have to ask a question. So when we were catching up before the show, I discovered that you like curling as a hobby. And as someone who started out their career as an accountant and has worked so deeply in the world of data, I just need to know personally, you know, when you're throwing the stone, how much are you trying to be really calculated like an accountant? Or are you, is this an area where you're just doing this by feel, you know?

Yeah. I mean, like, like any good analyst, I sort of think like you're trying to balance the two things, right? Like you're trying to go with your gut and your intuition, but also, you know, you've got like, you know, a bunch of facts about how's the ice tonight. Is it bumpy? Is it fast? Is it slow? Is it pulling in one direction or another? So, you know, I think, you know, I try to bring that same perspective that I have an analytics to when I'm out on the ice trying to curl. Well said. Is that what kind of separates, do you think, someone who's really good at curling from someone who's, you know? I mean, look, I'm still relatively new. I've only been doing it for a couple of years. Something tells me that just like, look, practice is the key to actually getting really terrific. Love it.

Well, let's talk about Alteryx. So Alteryx is a huge company just sort of ubiquitously used for analytics use cases, especially, you know, sort of across the enterprise. But not sort of commonly mentioned, you know, as, you know, sort of modern data stack-ish tooling, analytics tooling. Can you tell us what Alteryx is and what it does? Just give us a 101 on the platform.

Yeah. So we call Alteryx an analytics automation platform. I think historically, if you talked to us 10 or 15 years ago, we'd be talking about prep and blend. And we've kind of grown beyond that and added some more interesting and sophisticated analytics. But, you know, look, the basics are you're trying to take data from multiple sources. So, you know, you got to have inputs and you want to bring them together. You want to do various transformations. You want to clean the data. You want to make new calculations. You want to do some advanced analytics. And then you want to do various transformations. You want to clean the data. You want to make new calculations. You want to do some advanced analytics. And then you want to put it somewhere. So you want to output it, whether that's a database, a flat file, it gets sent in an email. And then you want to be able to schedule it and have it happen repeatedly. Because that's really, while the ad hoc analytics is important, it adds value. Being able to schedule it and automate it is where you can really scale up the value.

Now, you might sort of say, well, that sounds like an ETL tool or an ELT tool. Like, isn't that exactly what those things do? I was just going to ask. Yeah. And I think like if you just look at the features and functions on paper, yeah, like they're kind of similar. I think the biggest difference in where you see kind of Alteryx has had terrific success is really with very casual end users. So these are people who a lot of times, you know, they certainly can't write code. So Alteryx has a no code interface where you get these

analytical building blocks that you drag and drop onto a canvas and connect together. So, you know, it's a no-code interface. The people who use it are often sort of doing the job of an analyst, but not always, you know, full-time analysts. Sometimes they're accountants or tax professionals or supply chain people. And so, you know, in that sense, sometimes we're not always traveling in the same circles as, you know, the other pieces of the modern data stack. We're not, you know, trying to have IT be buyers or users, you know, clearly they're going to be approvers and a bigger process for us, but the end users are really very casual. And so, you know, we're trying to help democratize analytics and, you know, put it in the hands of the people who have the most context about the business. So you'll see us with with the modern data stack, but just sometimes with

different end users and slightly different buyers. So interesting. I want to go one layer deeper to the foundation in terms of our definitions. And so we talk about analytics. You mentioned democratizing analytics, okay? Yeah. Analytics is one of those terms where, you know, it's funny because you can ask someone like, do you know what analytics are? And everyone would say, yeah.

And it's like, okay, could you define it for me, right? And then they probably stop for a second and they're thinking, well, that's actually kind of hard to put a short definition on because you could argue that the transformations that you're running on the data in an ETL process qualify as analytics. You could argue that building a chart and a report is analytics and sort of everything in between. So

from AlterX view, can you define analytics? Can you give us sort of a working definition? Yeah, I mean, I think I do think of it as very broad, right? I think I do count transformations as analytics. I think a lot of times in those transformations, you're doing calculations. You're creating a new metric by combining the different data sources together. Sometimes you might be doing something really sophisticated from a statistics perspective, looking for one or two standard needs above or below the mean. It could be generating a report

or a graph that's going to highlight an insight. And I love this idea that analytics is very broad, right? It's taking that data, refining it, finding insights into the business. And I think it's just, you know, it's incredibly important for an entire organization to be aligned around driving value out of their data. And while you need the help from specialists, you need the help of the data engineer, you need the help of the data scientist. At the end of the day, you need to empower more people with these kinds of capabilities in the organization. It can't be trapped inside of a small team, especially the bigger the company, the more you need to train and enable them and give them the right kinds of tools in order to get value out of that data. Jay, I want you to paint a picture for us. And I think a use case might be an interesting way to tackle this. So could we talk about someone who does not have Alteryx and they need to perform some sort of analytics task, right? What's their workflow going to be both from a tooling standpoint

and from a team collaboration standpoint, right? Maybe that's a little generous, right? What do they need to go get from other teams? And then what does that look like to use, you know, all the tricks? You know, a lot of what we see is we run into a lot of people using Excel. You know, again, if you think about us touching very casual end users, that shouldn't be too surprising that, you know, organizations are relying on Excel to do some very complicated things. And so, you know, in your organization,

if you're looking at the guy who is awesome with pivot tables and VLOOKUP and other things, you know, those are people who are great candidates to graduate into Alteryx. And, you know, you'd be surprised at these, you know, those are people who are great candidates to graduate into Alteryx. And, you know, you'd be surprised at these, you know, sort of very sophisticated things

people are building out in Excel, but in the end wind up being very fragile, you know, wind up sort of not being part of the enterprise governance that you have around, you know, where it's kept and how you keep things up to date. And they're just very fragile and sort of don't automate very well. And that's a lot of what we can do in Alteryx is, you know, take those formulas out of an Excel file, give you a very simple representation of how data is coming in, all of the different transformations happening to it,

not have to, you know, remember some complicated formula for VLOOKUP and put it in something that, you know, can be automated and repeated and won't be so fragile and will be easier to just sort of understand and visualize exactly what's happening to the data. And so I think that's like a pretty basic example of where, you know, a lot of times you'll see great adoption of Alteryx is, you know, sort of those really advanced Excel users. Yeah. Yeah. That makes total sense.

You know, it's, you know, anyone who's ever, you know, had to, I put this in air quotes, you know, for the listeners who aren't on the video call, but to elaborate, you know, on a 450 megabyte Excel file with some pretty gnarly VLOOKUPs and some macros on it, and you're emailing or just using an internal network to iterate on that file, version control is underscore one, underscore two, makes total sense. And I agree with you. It's actually astounding the level of complexity. I mean, people end up almost building software inside of... So one other question on that use case, so that was sort of graduating from Excel. A lot of times those Excel files that you're talking about have inputs from other systems. Is that part of it, the challenge that Alteryx solves as well,

right? Because someone's getting a CSV export from where they're dumping it into a spreadsheet and running all their VLOOKUPs. Yeah. The really kind of interesting thing about the inputs that Alteryx can take in is, look, we're great at connecting to all the different pieces of your modern data stack. So we've got great support for Snowflake and Databricks and, you know, Trino or, you know, Dremio or, you know, whatever cool thing you're using. Yeah, we can pull data out of there. In our experience, though, sort of, you know, while data centralization projects are amazing, you know, what companies done with their data centralization project?

How long have we been talking about, you know, big centralized data warehouses and breaking silos? Yeah. Like the really interesting thing about Alteryx is it can let you accommodate the realities of your data infrastructure. And the fact that there's always, you know, some file that's not getting updated in the data warehouse, something that sure it's getting updated, but you know, the data updates, you know, hourly and the load into the warehouse is weekly or, you know, from your partner community, you're getting, you know, extracts out of their systems that are getting FTP or emailed to somewhere. And, and all a really great system to let, you know, the poor analyst who has to, you know, get that data and do something with it,

accomplish it, right? Because it puts the power and the tools in their hands directly. So they're not sort of beholden to IT. They're not, you know, they can actually self-serve around bringing this to date, the data together, you know, cleansing it and pulling it together into whatever output they're trying to do. In some cases that can just be loading it back into the data warehouse or the database. In other cases, it might be producing a report, but it gives you tremendous flexibility to deal with the realities of your very messy data that probably isn't all actually where you need it to be. Yeah. Makes total sense. Now, I want to ask the question here, because you mentioned when you were chatting before the call that even just in the couple of years that you've been there, the company's revenue

is almost double. Did I hear that correctly? Yeah, yeah. So ARR in particular was a little north of $500 million when I joined about two and a half years ago, and tremendous growth. It's been a pretty fun ride. Yeah. I mean, every company in the modern data stack that's a startup wants their valuation

to be a billion dollars, which is probably going to be a lot harder in this environment. You kind of describe Alteryx as sort of the biggest company that a lot of people have never heard of from a data perspective. Why do you think that is? You've been in and around the space for decades. Yeah. And, you know, before I came here, wasn't super familiar with Alteryx, right? Look, I think that's part of what we're looking to change. I think we are sort of, you know, talking a lot more across an enterprise. You know, I think a big reason of it is our end users are sometimes a little bit different than the rest of the modern data stack, right? While we have data engineers that use our product and we have people in IT who use our product, those aren't necessarily the people we're talking to every day.

And when we say we have 500,000 end users, a lot of them really are accountants or tax professionals or supply chain people. And frankly, their exposure to the rest of the modern data stack is relatively small. You know, all they know is, you know, IT put all their data in Snowflake or their data science team, you know, has a bunch of stuff they want to access to in Databricks. And so, you know, it's a little bit, you know, trying to help the people who have problems that are analytic problems, you know, who are being asked to play the role of analysts, but have some day job empowering them and sort of bringing them closer into, you know, the rest of the modern data stack that I think a bunch of your listeners are probably way more familiar with and helping build some of those bridges. I think that's a huge part of

our mission. And in, you know, frankly, it's what's been driving our success is you can scale up analytics. You can put, you know, these capabilities in tens of thousands of people in your company, or if it's a smaller company, dozens of people, right? It doesn't have to be, we don't need the folks to be gatekeepers. We can empower people. We can do it in a way that they can still govern and comply with the policies, but we can actually empower casual end users with sophisticated analytics. Yeah, that's incredible. Well, I know Costas has a bunch of questions. I actually have a question for you leading products. This is just a personal curiosity of mine, because when you think about a company that's clearly an enterprise, having multiple thousands of users within a single account, so a lot of companies are going to consider that a massive enterprise account. Yet you describe solving problems for a casual end user. How do you approach that from a product standpoint? Because it's clear that you need to have a hardened product that can withstand the needs of the Fortune 500. Yet there's a very human,

I just get a really strong sense of a human element to how you're describing your end user. And a lot of times people think, well, a big enterprise company, they don't care about the end user. But you have a very human tone. So how do you reconcile those things? I mean, it's fascinating. I would say we're very fortunate here at Alteryx

in that we really grew up selling a seated designer to one person in an organization, having them fall in love with it, that person telling five people in their organization and them telling five friends and them telling five friends. And it really is what fueled what I think of as the first wave of growth for Alteryx. And it's created, you know, raving fans for us. I've never met a more passionate and fervent, you know, user base and like, gosh, there's 500,000 of them. So like, there's a lot of them, like, you know, people have Alteryx tattoos, like it's, I mean, it's nuts. Oh, yeah. Oh, yeah. Now, what I would say is we definitely, you know, hit a wall where sort of the growth started to plateau a little bit. And in part is because, you know, look, once you get to a certain number of seats in an organization, you know, IT kind of sticks their head up and goes like, hey, what are you guys doing over there?

Like, how come you got all these seats? And really that inflection point was around when I started and having been at a larger company like IBM, part of why I came here was to help bring more enterprise readiness features to the Altrix platform and help put in the governance, the SDLC, all of the things that large organizations were looking to have in a platform that they were going to scale to hundreds or thousands or tens of thousands of people in their org. And so I think a lot of that has been the work we've that passionate fan base in the companies because, you know, there was no way, you know, anybody's ripping that out of their hands because it was so critical to getting their job done. I think that gave us the room to mature, you know, the backend infrastructure pieces that would meet those needs. Thank you, Eric. Thank you for giving me the microphone. I have some very

pressing questions I have to ask. So Jay, let's talk a little bit about the product and about this need for low-code, no-code kind of interface to these systems, right? And my question is the following. The problem that I always had with these systems, not so specifically in the data space, because we can see platforms trying to provide this kind of experience across the whole industry, from building websites to like i don't know doing ai in the mail and i felt and that's like i really want like your like thoughts here as a product person right it always felt like there is like a very delicate balance between making like a product that it's easy to use or turn it into something that looks easy to use but actually is like annoying to use because it's very i mean there

is a reason that okay machines are complex stuff right and there is a reason that we use something like a language to program them. The semantics are very rich there, and that's why we need programming languages. But transforming that into a user interface, for example, is a very hard task, actually. It's not easy. You can very easily end up in a system that is, at the end,

harder to use, in a way way or has a very bad experience. Can you help me understand how, as a product person, you navigate that? Considering also a user who is not technical, right? We're talking about an accountant here. So how do we turn these people into a data engineer in a way without even them knowing that they are turning themselves into data engineer. Yeah, it's fascinating. I mean, I think there's, you know, a few different building blocks to having a great no-code

interface to enabling somebody who's such a casual end user. You know, obviously, we have a particular sort of, you know, framework and user interface paradigm. It is a, you know, a flowchart interface. There are, you know, nodes or building blocks. What I would tell you is, yeah, that, you know, that core metaphor is important, but I've seen other, you know, low-code products that have, you know, flowcharts that don't work for a really casual end user. So some of it is also just sort of, you know, spending a lot of time with the users, watching them use it, doing usability studies, having a strong design team, and understanding deeply the use cases that the customers are trying to solve, because, you know, frankly, a lot of the ease comes from, you know, kind of the interface itself and it facilitating getting to the outcomes the users try to drive very quickly. And so, you know, we ease of use that we've got and helping drive the thrill of solving and being able to get to an answer really quickly. The other thing I like to think is that we also have some really nice escape hatches within the no-code interface that will let you do more sophisticated things. So yeah, you can cut and paste some SQL that you found on Google

or, you know, a little bit of Python code that if you hit a wall and there's some, you know, more sophisticated thing you're trying to do that, that there is a way to accomplish what you need. And that flexibility often can be really powerful, you know, because you're sort of giving people the opportunity to get past whatever roadblock they have. And so, you know, because you're sort of giving people the opportunity to get past whatever roadblock they have. And so, you know, I think there's a number of different things that you can, that you can bring together into one product to help, you know, keep the ease of use where it needs to be to make sure customers are, are getting the value out of the thing that you're using.

Okay. That's great. And one more like product related question. of the thing that you're using. Okay. That's great. And one more like product related question. So, okay. It's one thing like to try and it's built, you know, like a pretty complex product, but for one persona, right? Like one very specific, like type of user, which might be like a data engineer or it might be an ML engineer, but pretty much like all these people are like the same, like they speak the same language first of all, which might be like a data engineer or it might be an ML engineer. But pretty much,

all these people are the same. They speak the same language, first of all. Their vocabulary is not that different from one to the other. So even when you're just talking with them, you're going to hear the same things. But when we're talking about democratizing access to data through a no-code or low-code system. We're literally talking about like professionals inside an organization that may be coming like from completely different like contexts, right? Like someone who is coming from logistics and someone who's an accountant are like completely different people. Like they understand the world in very different terms, right?

And they are exposed to different tools. They use different language. But both of them might be using alterings. So how do you do that? Because that's, I think, even harder to do. And I'm saying that as a person who has tried to build products and I've seen how complex it gets by just introducing one more persona in there. I can't imagine how it is when the space of personas

is actually open, right? It can't be literally like anyone. Yeah. I mean, look, I think the thing that ties all those people together is their lackability to code, right? They don't know SQL. They don't know Python. But you know what they do know?

They know Excel. And, you know, sort of if you can tell them like, well, you know, Hey, imagine you take those 10 Excel files and combine it into one, and then you can sort it and you can write formulas to make new calculations. There are, I think some very, you know, familiar metaphors that can span those roles, you know, from things like spreadsheets. And frankly, you know, familiar metaphors that can span those roles, you know, from things like spreadsheets. And frankly, you know, that is very much where a lot of those end users are coming from. And so, you know, hey, our formula function, you know, it draws on that experience that people have across Excel. So I do think there's some familiar metaphors, but, you know, look, I think it also helps to have great training and great onboarding experiences

for the customer. And, you know, we provide a lot of example workflows for how to accomplish different things. You know, that we have a community where, you know, people can ask for help and advice. We have a program called the Alteryx Aces, which are super users and ambassadors of our product that will, you know, answer questions in that community board. And so there's a lot of things that kind of surround the products themselves, I think, that help facilitate getting people familiar with it.

Maybe the other thing I think that's really interesting is people are starving for insights and, you know, they're kind of drowning in data. And so when you can give them a tool that feels intuitive, it's amazing to just watch their enthusiasm and sort of their willingness to learn how to do things inside the application. And sort of, I think that there is this idea we call the thrill of solving that I think is fun to watch unfold, but also very motivating for the people and makes them willing to engage and learn how to use some software and learn how to do things like, you know, join a data set together when they don't know anything about what a join means, and to look at inner and outer joins and things like that, that are sort of very data engineering concepts, but exposed in a familiar and easy to use way. Yeah, 100%, that's super interesting. All right.

And okay, we talked a lot about the non-technical people, but usually in larger, special organizations, there are also technical people in there. And part of their job is to govern the data and manage the data and expose the data to the rest of the organization. And I think a big part of what was happening these past couple of years, especially with the rise of the cloud warehouses, was like, okay, let's get all the data, put them in one place and have like the data engineers or like the email engineers or like data platform or people or whatever we want to call them to govern in a way the access to this data.

So, and okay, the bread and butter of someone like a data engineer is like a pipeline, right? That's what they're doing for a living at the end. How an environment, in a real environment, other than a company that has real, I would say, use case around their data, how systems like AlterX play together with, let's say, something like Spark or like Databricks, like the more, let's say, data engineering tooling out there, right? And then let's talk about the technology a little bit, and then we will talk about the people because that's even more interesting, to be honest. Yeah.

So, I mean, you know, one of the things to note is, you know, I think we're in the middle of a really interesting inflection point for analytics. If you think about the old way to do analytics was you would get an extract out of the data warehouse and then put it in some analytics tool to twist and turn and slice and dice it. That model is really getting flipped on its head, in my opinion, where now instead of sort of bringing the data to the analytics, we're instead of sort of bringing the data to the analytics, we're switching that to be taking the analytics to the data. And so if you look at the investments we're making around push down, around in-database processing, around cloud native

compute, we're able to take the analytics to where the data is and actually leave it inside of the warehouse as we're creating new metrics and doing different selections and sorting and filtering and all the different analytic things that we need that we want to do, you're able to sort of not have to egress the data. You can leave it where it is. And I think that's sort of one of the more interesting and exciting trends that we've seen, you know, in analytics. And so, yeah, we'll connect into Databricks. Yeah,

we'll push down into Spark and, you know, we'll leverage your Databricks Unity catalog if you've got that all connected up. So I think there is really interesting opportunity to sort of leverage the existing investments that your data teams are making and sort of, you know, gosh, you've spent all this money on Snowflake. You've spent all this money on Databricks. You need all these people to access it. You want to derive the value from it. You want to get the data in the hands of the people who have the context for the data and how it applies to the business and let them use it and activate it. So, you know, in some ways I think it's, you know, Alteryx has been pretty great partners, particularly to the database vendors, where we're helping them realize the value of the big investments they've made in some of those systems.

Yeah, 100%. And how about the people involved? How data engineers, for example, feel about giving the power to anyone to go and build pipelines at the end, right? That is going like to run on like their systems. So how is the, what are like the boundaries there between what the data engineer cares about or should care about and what, let's say, the last mile pipeline that like a user needs? Yeah. It's, you know, look, first of all, what I would tell you is, you know, analytics is a team sport, right? These people need to collaborate. And frankly, I think as a vendor, the vendors need to do a better job helping these people collaborate because it's not, it's, you know, it's harder than it needs to be. And, you know, there's all the people in the organizational dynamics too. And so, you know, I don't want to make it sound easier or convince people that all you got to do is buy some software from Alteryx and it'll all just work. But I think you actually kind of touched on it, like what the sort of best practices around how a lot of companies are set up, which is,

you know, you need those data engineers building those pipelines, sort of creating the data sets that are the source for all the things people are doing in Alteryx. And they're the base things that people are selecting from, that they're combining with their spreadsheets and doing all the downstream calculations. And I think it is tough to know sometimes which things do you want to push upstream into the warehouse versus downstream, let the end users take care of? And some organizations are afraid to give hundreds or thousands of people access to that data because we believe in its importance to running an effective business. And so what I would say is I can't really convince myself that a successful business in the future won't have every single employee being a data worker. You know, just like the way we used to talk about knowledge workers. Every single person in the company is going to need access to data to get their job done. And the exciting thing about it now is you can still derive competitive advantage from it.

And so, you know, you can be at the forefront of this trend and create real competitive advantage for your organization at scale if you can provide the right kinds of data and the right kinds of analytics to every single person in the company. Yeah. And why are companies afraid of that? Because what you're describing, and by the way, I totally agree with you, it's very important, right? How people can create an organization without having access like to the data.

So why this has to be- I think sometimes they're scared for good reasons. You know, there's regulatory concerns. There's, you know, privacy concerns. You know, data can be sensitive stuff. And, you know, sort of having the right governance structure, I think is important to being successful with these sorts of efforts. And I think, you know, sort of empowering end users, you know, can feel like a scary thing. I think that the governance angle is one. I also

think there's big skills gaps, you know, from all these end users. I think there's concern that sort of, will the users know what to do with the data once they have it? Will they be able to, you know, sort of create productive insights out of it? And I guess my argument there would be, first of all, I think, you know, you can create competitive advantage if you give them access to the data. I think oftentimes the people closest to the business also understand the implications of the data much better than centralized teams. And I think the software has come a long way just in being able to automatically surface insights that are interesting or important or have changed. And so, you know, I think there's, we've come a long way to sort of lowering the barriers of entry

into surfacing insights. And I think those people who have that context are going to be great at doing it. So, yeah, I think those are probably the two biggest things I hear from our customers, you know, worried about governance

and worried about the skills gaps. But I think there are things you can do programmatically to put people in a position to be successful with it. Yeah, and talking about the gapping skills, I think it's a very good opportunity to get into the Gen AI trends. So let's talk a little bit about that. Because in a way a way we're talking

about like generative AI, like it sounds like the, you know, the ultimate no code, low code, whatever solution, right? You don't even have to type like practically technically if you don't want

to. So what's, how's AlterX like looking into that? Like, and let's try like to remove the hype from it and keep the substance because there's also a lot of hype out there. And I think it's important to do that. One of the things I'd suggest for every single person listening to this podcast

is if you haven't signed up for ChatGPT Pro, go do it and play around with the advanced analytics. And it is, you know, pretty mind blowing to, to use that thing. You can upload a file of data and you can just say things like, tell me something interesting about my data or perform some advanced analytics on my data. And the things that it can do, I think, are a really clear indicator for how generative AI is going to impact the analytic space. Now, there's a whole bunch of problems with sort of that as a model for analytics. First of all, I think, you know, first of all, no one wants to load up their data to open AI and have the data leave their four walls.

The things that ChatGPT does with the data aren't repeatable. So I can't schedule it. I can't make it go on an ongoing basis. I think also it doesn't have the context of your business and the things that are happening in your company or your industry or your function. And so there will be fine tuning of the models to give it better context. And so, you know, I also think that asset that it produces, what ChatGPT does is they wind up writing Python code to analyze the data. Well, that's great. But if I don't know anything about Python, cool, I can get an answer. But I can't look at that Python code and know whether it's actually doing models, fine tune on top of a base model, and sort of give your company's data so the model can be deployed in your four walls. So companies should be comfortable feeding their own data into it to fine tune it. And then we can do chat to SQL, chat to Python, but also chat to workflows. And now as a really casual

end user, I get an asset that it creates where I can look and see, oh, look, it really is pulling the right data sources. It really is filtering and sorting and that formula for calculating the tax implication looks right. And so it's sort of creating an asset that a very casual end user could look at and feel good about. And so, you know, to me, as I think about generative AI being able to do, you know, chat to Alteryx workflow or chat to SQL and chat to Python. Now, all of a sudden, like itilot with us that's going to help, you know, help us navigate across all these different lenses of, of light or languages of analytics that we want to talk. So to me, I just, I think we're at a really, you know, transformative moment. I think this is the most disruptive technology in generations.

And I think we're going to see some really exciting things in the analytics space. Yeah, okay, that's amazing. Eric, I have to give the mic back to you because I know when we are talking about AI, you have so many questions. Yeah, no, I mean, we're pretty close to the buzzer here. You know, I'm interested to know, I agree with you, Jay.

I really think the transformative potential is immense. But I'm also interested to know where you think the initial failures are going to be. You know, especially as it relates to relates to how LLMs impact analytics. There's a lot of promise out there. I think one of the things we can say about AI is that it's created this ocean of promises that it's really hard to distinguish between what's overblown and what's real because it's very clear that there's immense potential there. So help us understand which promises are maybe not going to come true when it comes to the world of data and analytics. Yeah. I mean, it's a really great question. The analogy for me, at least,

this feels like the moment when the internet first got big. It does not feel like the moment when blockchain was at its peak height. Yeah. Yeah. And I would say, look, it's not that sort of the internet didn't have, you know, we had diapers.com still. So, you know, it's not like there wasn't failures during that phase, but, you know, there were enough successes where, you know, we've now created enduring business value. And so, you know, as I kind of look at generative AI, you know, I think your list of generative experience experiments right now probably needs to be pretty long. Right. And I think you're right. Not all of them are going to be successful, but I think for me, the thing that feels very obvious to me is that there will be lots of successes and that, that sort of maybe not everything's going to stick, but you're going

to find some things that are going to just, you know, be incredibly valuable. And so I guess my advice to companies is like, you better be running some of those experiments and you better be figuring out how it applies to your industry, your function, your, you know, you as a, as an individual, because,

you know, I think it's, it will be the most transformative thing that happens to all of us. And I realize I'm feeding into the hype, but guys, there's a reason there's hype, right? Like there's,

there is, there's real promise there of delivering business value. And frankly, just, you know, taking analytics to the analytics to an entirely different level and getting not just data into the hands of people and not just insights, but giving them the ability to interact to create the insights that will impact their day-to-day jobs and decisions. So, you know, try it out, see what works, you know, and you know, you got to balance moving quickly with the need for enterprise governance and to, you know, proceed in ways that aren't, you know, that are going to fit within your company's policies and things like that. But man, there better be, you know, a couple dozen experiments you guys are all running. That's my advice. Yeah. Wise words. Jay, so great to learn about Alteryx. I just love hearing, again,

I'm just going to reiterate sort of the human nature of the way that you think about the product and your users, you know, even within large enterprises. I think there's just a lot for all of us to learn from the way that you're leading that. And thanks for giving us some of your time on the Data Stack Show. Yeah. Thanks, guys. This was really fun.

We hope you enjoyed this episode of the Data Stack Show. Be sure to subscribe on your favorite podcast app to get notified about new episodes every week. We'd also love your feedback. You can email me, ericdodds, at eric at datastackshow.com. That's E-R-I-C at datastackshow.com. The show is brought to you by Rudderstack, the CDP for developers. Learn how to build a CDP on your data warehouse at rudderstack.com. Thank you.