Adam Wilson, Alteryx | AWS re:Invent 2022

This is texturong TV. Hey folks, we're at the AWS reinventure. We're talking to animals and who's a senior vice president with alterix, and we're talking about the convergence of alterix and trifacta that they acquired last February.

I think it was. Yeah, and we're also gonna have a chat about the democratization of Dana management and how all this is coming together with business intelligence Adam. Welcome the show.

Hey, it's great to be here. Thanks. Thanks for having me.

I don't think everybody knows who try fact it was and I'm not sure even everybody knows who will Turks is I want to walk us through exactly what two plus two equals five here. Yeah that all comes together. Yeah.

Now this has been it's been super fun and super exciting. So so back in February all tricks announced the acquisition of Trifecta, and now try fact has come in as the as the underlying Cloud platform for all the ultrix analytics cloud and what that's gonna do is to really take a lot of the ideas around self-service data preparation as well as and user analytics and make that broadly available to a much larger set of Constituents and so alteryx has you know over 8,000 customers and hundreds of thousands of users who've been using the product very successfully on the desktop, you know over the past 20 years or so and now we're taking the story in a big way to cloud and trifacta is excited to be part of that. So it's a lot of fun for us to to get to do this now on a much bigger stage and to be part of this this trend that you're seeing where everyone's focused on trying to help enable the people who know the data best to do more of the work so well exactly.

What do we mean by self-service because yes to be as an end user I'd call up the IT department and I'd be like, hey guys, I need this amount of data and you know, they'd be like looking back to you in about three months or so and that's the time they got that together. I forgot what I was caring about and what the question was and what the issue was. So what will be the role of it teams as we go forward if we have more self service and how does that self-service come about?

Yeah. No, I think it's really interesting because it's Shifting the dynamic right? So you have this impedance mismatch.

Are you often in most organizations have you know thousands of users who are being told to be more data-driven and you have a very small group of people who historically been in a position to provision the data that they need in order to actually fulfill that to fulfill that mandate. And so now what we're trying to do is to flip this around and to say why is it that that people who know the data best can't do this work on their own can we turn this into a user experience challenge powered by Machine learning that helps automate the most complicated stuff. So that end users who are not necessarily structured programmers can get in and can do more of this on their own.

So the nice part is now we can learn from the data and we can learn from how the user interacts with the data in order to speed up the process of going from raw to refined and that actually takes a lot of Burden off of it who often we're finding that they had backlogs of requests coming from the line of business that they couldn't fulfill. So now they can like hand this work to the end users. The nice part is the end users are going to thank them for giving them that work and that frees it up to crowdsource the best stuff to focus on the governance and controls around how data then gets moved into production and gets used across the Enterprise as well as to Source new and interesting data sets that could be brought in that could be used by the by the line of business and the analysts in order to to create new and interesting data products that might drive the business forward.

Do you think as a result that maybe we'll get better data because one of the dirty little secrets of it is that well most end users. Don't trust the data because they know who entered it it was them and they didn't really put a lot of effort into it because to them it was a chore. Yeah.

So are we gonna get to the point now where the quality of the data will improve because the people creating it or the same people using it and they'll be like well man this matters to me. Yeah. Yeah.

I think this is a really interesting topic because people have realized that if their data quality is bad then they're analytics and they're they're AI or probably worthless. And you know, can you imagine a state where we start automating more bad decisions faster based on bad data? So this is creating a burning platform for for companies to take data quality much more seriously.

So I do think that again when you give the data to the users who understand how it gets used to make business decisions. That naturally is going to create that shining bright light into dark Corners often that will allow us to start to standardize and clean up a lot of this data. The other thing though that I think is really powerful and one of the new capabilities that we introduced is something called adaptive data quality and what we mean by that is why as the data volumes get bigger as the data diversity gets bigger.

It's hard for a human being to upfront understand all of the validation and data quality logic that they would want to apply to do checks. So why can't the data sets suggest their own rules? Like why can't we let the machines do the work at scale and come back and say hey here things we see in the data that you might want to establish safeguards or rails around in order to ensure that you get sort of better quality analytics coming out of the data.

And so that sort of AI assisted approach we think is going to be more and more pervasive not just for data quality, but for data data integration as well as for just Analytics generally where the AI helps suggest the analytics that you should be paying attention to. We've been tilting at this business intelligence windmill for the better part of two decades now and I go into an office and I still see spreadsheets everywhere and I shake their head. So yeah, what will get people off of this spreadsheet addiction that they currently have and move to something that isn't that fundamentally a better application and yet everybody still runs around and uses a spreadsheet.

Yeah. Well, it's interesting. We did a study recently with IDC that said that you know over 50% of organization people and organizations have access to no more sophisticated tools than a spreadsheet when it comes to analytics so not surprising right and I think that what you're finding is that more and more organizations are realizing that if if you can take more of a platform Centric approach and for us that's the ultrix analytics Cloud platform where you can start to now bring together different constituents where you've got data Engineers, you got business analysts, you've got data scientists who are all collaboratively curating data together.

It creates a force multiplying effect. Of the organization because now everybody's not sort of on their own Island manipulating data in their own way, but you're actually leveraging the power of the cloud in order to to do this in a way that's going to allow more technical users to work with less technical users who have more business context in order to get to kind of better data products faster and we think that's a winning formula and you're starting to really see as more and more of the data gravity shifts the cloud you're starting to see more and more of that happening in Enterprises at scale. How smart will we these apps get you mentioned AI do we get to the point where maybe I'll walk in the office one day and it'll tell me here's the three things.

You're likely get fired for and you should go do something about this or you know, how much do I need to know to ask the question? Yeah versus how much can the thing tell me what question to ask. Yeah.

Well one of the offerings that's in the ultrix analytics Cloud that we're particularly excited about is that an offering called Auto insights and rather than you know, building static, you know dashboards or reports that get passed around the idea is exactly this which is we're going to introspect the data and we're going to come back and say hey given what we see going on in the data here are some things we think you might want to pay attention to and so, you know for us this is really exciting because now it's starting to really put the power in the hands of the end users to then go in and not sort of start from scratch and not consume things in the static and the static way especially as a date is constantly changing but you're able to keep up with the analytics as they're changing and the system itself again reveals. Might be the most interesting Trends patterns causal relationships correlations and things that that when you walk in and sit down allows you to focus your time and attention on those those areas now we have all this data that sits in these databases and we used to hire these people. Well, I called them sequel jockeys.

They used to sit there and come come up with all these great queries and yeah, give your result. Yeah what happens to all those people if we're gonna self-service our own needs in the end users are doing all the querying what's their role? What becomes them?

Yeah. Yeah. No and I think that it's it's let me give you an example of how this played out at one of our large Financial Services customers.

So this is one of the largest banks in the planet. They have a group of quants that split roughly equally between fixed income and equities. They had about three to 400 of these individuals that are constantly looking for for algorithms that they can trade on.

There was a group with nit that was supporting them. That was roughly 10 people so you can imagine In the backlog of work that they had because these quants are constantly trying to stitch together data new and interesting ways in order to gain some advantage in in the market. And so so the backlog measured in years can't hire your way out of that because they don't have the budget to like just hire tons and tons of additional people also really hard to find those skills.

So what they said was what we're going to do is we're going to create open up raw zones in our data Lake and we're gonna allow a lot of interactive experimentation. We're going to put the data in there. Let the quants go after the data do that in a self-service way and then the minute they actually create training data that births and algorithm now, we're gonna get back involved and now we're going to say hey, we'll help you operationalize this at scale because if we're gonna make multi-million dollar trading decisions, we want to know where did that data set come from the train that algorithm.

How is it transformed along the way who touched it? Because now we have to have full audit Trail on and chain of custody on what happened with that data, and so now a lot of The governance the control the ReUse all being done, you know by that smaller it organization, but allowing these guys to run much faster because they can do it in a self-service way. And in that game if you're a little faster, you don't just win a little more you win all of it for some period of time until the market catches up with whatever insights or algorithms you've created.

And so that becomes a really powerful not just operational efficiency story, but a really powerful like Top Line business growth story. The other thing that the it organization said to us is we implemented this as they said then we're also in a better position to go out and really find other interesting data sets that we could bring in that these guys could start to spend time with so that's government data. That's dark data that's you know, data data from data Brokers.

It's data that's littered all over the inside of the organization that they can start to bring together. So just elevates the Strategic nature of their role rather than forcing them to spend all their time, you know building Pipelines. Wants to you know requirements that are coming over the wall.

Hello, I'm validate the data sets because it's one wag one said it's one thing to be wrong. It's another thing to be wrong at scale. Yeah.

Yeah, how do I kind of know that the data is reliable and especially if I'm bringing in this third party stuff and I'm comparing contrasting and what's my level of confidence in the data? Yeah. Well, that's where you're seeing this massive Trend around observability where people are starting to say Okay, first step was can I sort of find where I have data assets or where I can where I can get them together then can I start to like integrate them and can I do that at scale now across a much broader set of users and use cases, but very quickly this again this Burning platform around understanding the quality understanding how the data May evolve or change because just because you find a data source and get your pipelines created and start to do your analytics doesn't mean that that data isn't going to shift or drift in some fashion whether that's the actual schema or whether that's the date itself.

So putting in place a lot of monitoring a lot. Of controls a lot of checks that ensure that those things are detected and that and that people can respond to them efficiently has become kind of the next, you know Frontier for competition. I think in this market because that's what the end users are saying is they start to really think about about you know, really the data products driving their business and allowing them to cater along till segments in their Market or to manage and model risk better.

That's the thing. That's really going to I think separate a lot of the the successful companies as we go forward, you know from from those that you know, really haven't figured out how to harness a lot of the data they have under management. So you mentioned competition what differentiates alteryx from everybody else.

There's no shortage of bi and analytics applications out there. So what should people be looking for? Yeah.

Yeah. Well and I think that, you know from from our perspective this really is about tapping into what's going on, you know in the line of business, you know with the people who are making the business. Legends, right.

So I think you know the Hallmark of the company from the very beginning was, you know, getting this technology into the hands of the end users and and truly democratizing what was historically a very technical very expensive problem. So we talk a lot about analytics Automation and we talk a lot about doing that for the end users who you know who are not again necessarily structured programmers, but live in data every day are incredibly data driven. They're more likely to look like someone that's an accountant or someone that's a financial analyst and then they are someone who is a computer science graduate.

And so I think that's allowed the company to you know, now approaching a billion dollars in Revenue, you know group 75% last quarter in Revenue, you know over, you know, several hundred thousand users across 8,000 accounts, you know, this is this is a been a very successful organization that ultimately is starting to now think about Cloud as an amplifying effect on Everything that all the success that we've had historically on the desktop and so I think as we start to bring that together across democratizing machine learning democratizing what I called the sort of Auto insights or some of those analytic insights as well as democratizing. What is historically been an ETL process bringing all that together in ultrix analytics Cloud gives everyone a unified Approach at a time when the market is really saying listen, we're trying to figure out how to do more with less. We're trying to figure out how to automate more things.

We're trying to figure out whether what's going on right now in our business is the economy or is it us? Right? And and so they're looking for a unified platform in order to do that in a platform that they can cost effectively deploy and scale across their entire organization.

And that's where we're trying to get people to sort of lift their heads up from just being inside of spreadsheets all the time and starting to sort of think about this in a new way. And that is really again been a lot of what is resulted in the success of the company. Yeah.

What's your best advice to folks then or conversely? What's that thing that you see people doing? That just makes you shake your head?

Yeah, you know, it's it's really interesting. I think that I'll give you one other kind of example just to illustrate ways things that used to make me shake my head that I think are finally starting to change a little bit. So we do work with one of the largest pharmaceutical companies and they looked at the the clinical trial process and they said, you know, it takes somewhere between 10 to 12 years to take a drug from inception to FDA approval and their belief was that a lot of this was due to the fact that they were spending too much time trying to get data stitched together across all these different activities that were involved in that process.

And so you're thinking about things like experiment data assay data clinical trial data real world evidence, right? So they they're starting to put sensors into devices that they're deploying that are going through clinical trials so they can see Behavior data. So the interesting part of this was that they used to in some cases wait until dat.

Would get published to the government as part of the approval process in order to pull that data back in in a standardized way so they could even use it internally. Can you imagine having to go through the government in order to get clean access to your own data? So they step back and they said we need to rethink this radically and we need to also think about ways in which we can get the scientists not data scientists, but literally the chemists right that are working in this area to be able to do again more of this on their own because they know what the codes mean they understand how the data that they're getting from from, you know, the real world evidence and from the clinical trials and the experiment data all should kind of harmonize together so they can do the analysis that they're doing and their belief is and what they're proving out right now is that they can cut that time from inception to FDA approval more than in half.

So you're talking about going from 10 to 12 years down to something that looks more like five years and that's not only powerful for their business, but that's powerful for the health of the population, right? and and so I think for us seeing that cultural shift and seeing that that change, you know, that's the kind of stuff that I would go in and you know sort of shake my head at previously thinking like this is such a data Rich environment and the impact of getting this right is so massive, you know, why why can't we sort of rally organizations more fundamentally to see that they at their at their at their heart are really, you know, data fundamentally data companies and that this can be a massive competitive Advantage if they can increase the agility and what they're doing and I think now you're actually starting to see that play out, you know in you know in places like pharmaceutical and financial services and these data intensive industries that often are very highly regulated and in the past, you know, those regulations and their inability to get the data harmonized, you know resulted in frankly just slowness and resulted in a lot of frustration from you know from the users that you know, We're trying to get productive with data. All right, guys, you heard it here if you're running your business on the back of spreadsheets.

You're probably following behind. Hey, thanks. Okay, great to see you.

Thanks for the time. All right guys back to you.