Eric Kavanagh: Alright, ladies and gentlemen, Eric Kavanagh here with Hot Technologies. We’ve got Josh Howard and Wayne Eckerson on the line. We just had a fun little audio issue crash and burn right there, but we’re dialed back in and everything is rocking and rolling.
So, Wayne Eckerson I’ve known for many years now. He’s the principal consultant at the Eckerson Group. And Josh Howard I’ve also known for a long time. He’s the director of new products at Alteryx. These guys are both really, really excellent in their fields, and they’re going to be sharing with us a lot of ideas about how business and IT can foster better relationships and really collaborate and get some things done.
So, I’m going to push the next slide and hand it over to Wayne. So, tell me a bit about what’s going on.
Wayne Eckerson: Sure, Eric. It’s a pleasure to be here and talk about this issue. I’ve been in the States for a long time and have witnessed a gulf between business and IT, and a lot of that is due to their focus and their goals, what they’ve been hired to do. So it’s kind of a natural gulf, you could say, or gap between business and IT, but it does lead to some harmful outcomes. You know, IT has been hired to think long term, to build systems and applications, solutions that are permanent that offer economies with scale, high levels of reuse, and scalability, security, availability and reliability. Very conservative, slower-moving mindset. Business, on the other hand, is focused on meeting the needs of the customer, the point of interaction, much more short-term focused, incentives – and it could be dished out on a monthly or quarterly basis. Their focus is speed, agility and adaptability. So, there’s no surprise that there should be or could be friction between these two groups.
Next slide. So, this is kind of the dialog I sometimes hear at organizations where I go in to consult and where I feel like I’m playing the role of a marriage counselor, trying to get these two sides to one, acknowledge each other and their role in delivering business technology solutions. The business tends to think of IT as too slow, expensive and never deliver what they want, when they want it, how they want it. The IT tends to see the business as changing its mind all the time, adding new features. Then all these things move short-term, never seeing the big picture. The result oftentimes with this friction is that the casual use. There is the executive manager will say, “You know what? Just forget it. I know I’m not going to get the data I need, so I’ll just do without.” That’s pretty scary. The power user of data will say, "Just give me a dump of data and don’t bother me." And the BU leaders, if they really want information, they’ll just get their own budget, add their own people, and buy their own tools. IT says, “Alright, fine. But you know, good luck trying to maintain that on your own, because eventually it will break.” And it will. It will break either because no one is using it, because it wasn’t designed properly, or it will break because everyone’s using it, and you don’t have enough of technical experts on the ground, not enough resources to scale it. Or their expert leaves, and they’re out high and dry. Next slide.
Eric Kavanagh: This is a poll, so phone caller can actually push to poll. Hold on one second. So, I’m opening up this poll right now, hopefully you will see on your screen a pop-up. If you don’t, usually it will show up somewhere on the bottom. And go ahead. We’re curious to hear your answer on this.
OK, I got a few folks calling in now giving us some feedback. So, we’re asking: what degree is business aligned with IT in your organization? So, we’ve got a bunch of folks answering now. Thank you very much. So you’ve got very high, of course, high, moderate, low, very low. Be honest, we won’t share this with the other members of your team. We want you to give us your candid response. Alright, let me give us a few more seconds, and as we’re doing that, maybe Josh, we'll just bring you in real quick to help folks answering this question. Yeah, I love this process of collaboration. I mean, we’ve talked for years about a business/IT divide. I think that’s changing. I think it’s partially changing because of DevOps, the developers working more closely with the business. That kind of puts some heat off the IT side, but I think it’s also changing because of the cloud, quite frankly, because maybe people are just becoming more savvy about what they do in their workplace. But, what’s your thought on sort of the evolution of the IT/business divide?
Josh Howard: Yeah, you know, that’s an interesting topic, and it’s one that we’ll definitely get into here in a second, but, you know, I just think that the business is really forced to IT’s hand. That’s right, so, you know, for years everything was IT-led, and we’ve seen this come to pendulum swing back and forth from being IT-led to everything, you know, being bought through the business. And, I think we’re starting to see some centralization. I think, you know, you’re starting to see more organizations, stand-up centers of excellence, starting to see more and more businesses-intelligent companies, seeing centers being set up as well, and so it’s not, you know, IT or the business. We’re seeing a much better marriage of the two organizations and seeing these centers of excellence be set up that resides in both those organizations, and they’re having both IT and the business taking a seat at the table and order food. We need to pick other business objectives, and so I think that’s one of the trends that I think’s been very positive over the last few years or even longer. And I think that’s part of what we’re seeing.
Eric Kavanagh: Can’t blame me that I’ll throw over to you, and I’ll read out the results. Depending on your browser, you may see the results already, but just to give it to you: The question of course, “To what degree is business aligned with IT?” Very high got 7 percent, high got 8 percent, moderate got the vast majority, it’s 29 percent, low is 10 percent, and very low is 0 percent. That’s basically the total, so really what you’re looking at is most people said moderate, 21 out of 73. Six out of 73 said high, five said very high, and then of course we have a whole bunch of people who just didn’t answer, but most, actually 43 out of 73, people didn’t respond, but I appreciate your time. And with that I want to push this next slide. And I believe, Josh, you were going to talk a bit.
Josh Howard: Yeah, and so, you know, kind of where I was going was we’ve seen a lot of change in the last five years, or even going back ten years. And it really used to be the wild west, and then I’m guessing there’s probably some folks here on the line that still think it’s the wild west in their organization, but it used to be where everything was completely locked down and rigid, and everything was forced through a centralized IT team, and that was just how BI was delivered. But the problem was that the business users weren’t using it. They never got the results they needed. They couldn’t, you know, munge data together like they needed, and so you just saw, you know, organizations abandoning their BI practice in a lot of cases. They just weren’t getting the usage that they expected, and, you know, that’s understandable because the users, they wanted easy-to-use tools where they could take, you know, data sources and do some of their own integration work.
But they didn’t want to wait around for IT to do this for them. And so what we saw was, you got all these business teams go off and purchase their own license, their own visualization tools, and had their shadow IT buddies set up a data mart, and they were off. But that led to a whole new set of problems. Yes, the business was able to get the flexibility and agility and some of the results they needed much faster, but still left IT, you know, trying to figure out, “How do we govern this? How do we scale this?”
Because also what was happening, they were building up these data marts. They were starting to operationalize a lot of the reporting and visualizations, then they’d just go back over to IT to get the fix, and so it’s just not scalable. It wasn’t the cure, and so those were some of the issues. But it doesn’t have to be a tug-of-war between the business, who wants ease of use, and IT, who wants to govern it. It’s really about getting everyone on the same page and pulling in the same direction. I think there really is a, you know, best-of-breed approach that can satisfy the needs of both users. Slide.
Eric Kavanagh: Alrighty. There you go.
Josh Howard: Yeah, thanks. And so the way that we’re approaching at Alteryx is we’re really looking at it from an analytic governance standpoint. And so, you know, I’m not using the word “data governance” here because I think data governance is much more of a framework that encompasses a lot of different things, but really just focused in on these three key areas of how the data’s being managed, how it’s being accessed, and how we’re securing it.
First off, on the data management side, when you’re looking to enable self-service tools, you want to make sure that, you know, those users have access to all the different data sources they may need. And so, again, this is part of the problem that we saw with traditional BI tools like MicroStrategy and Cognos and OB was, you know, it was just tapping into a centralized data warehouse, but those business users really wanted to take that data and blend it in with other data sources to get additional results.
I mean, so you want to make sure that directly to all those different data sources, regardless if they’re relational or non-relational, and do it in a way that’s not going to make the data redundant. And so, you want to make sure that you’re using in-memory technologies so that you’re tapping into those federated data sources and not duplicating that data elsewhere in the organization, because that just causes a whole set of issues.
And then you want to make sure that you’re looking at things like data accessibility and data security, making sure that the data’s being encrypted, making sure that you’ve got the right permissions and authorizations in place. And what we recommend is use the systems that your IT teams have already set up, so things like Active Directory and Windows authentication. Tapping into those systems that can pass through that authentication all the way down to the application, and that way you can ensure the right users are getting access to the right data.
It’s really about moving from a state of control to a state of enablement, and doing so with guardrails. So, you know, analytics of guardrails, where IT is giving all the tools to be successful, but they’re also monitoring it, making sure that it’s consistent, it’s reliable, and that they’re doing it with the right permissions in place, and making sure that those users only have access to the right data. Next slide.
Eric Kavanagh: Alright, Dr. Wayne.
Wayne Eckerson: Yeah, so this is my slide. This just shows the dimensions of self-service, which Josh was talking about. That is the business mean of demand these days, but they don’t want to wait, as Josh said, for IT to deliver stuff, and IT used to do it all. They used to build the architecture and manage the infrastructure and pick the tools and build the applications, the reports, the dashboard, and that just does not work for a vast majority of users out there. And now we’re near self-service. We’ve got self-service reporting, self-service dashboards, which I call [inaudible], self-service visual discovery. We’ve got self-service data integration, or data preparation. We’ve got self-service advanced analytics, where there are some data scientists. So we’re thinking all these capabilities available to people, to business people, who are inclined to do things on their own.
Next slide. We’re getting some feedback here, Eric, just to let you know. So, you know, self-service on the surface looks like a win-win for both the business and the IT department. Users get what they want when they want it, how they want it. The IT department gets type of users, they get to offload the work, and they get to deliver things indirectly, but either way... In a lot of situations self-service has some significant downsides that you have to be careful about. And Josh was giving you some remedies for some of these downsides.
Go to the next slide, Eric, and we’ll just see that organizations’ self-service as kind of a tidal wave of force, which are duplicate, conflicting. And it gets to the point where no one trusts anyone else’s report except their own, which is not a good state of affairs. You could even say it’s worse than when they started. You basically have an architecture that’s comprised of shadow reporting systems, data extracts, which ultimately increase cost and overhead and redundancy and duplication and, consequently, increases risk in the organization. So, self-service is about standards where governance is really just the Tower of Babel. Everybody’s communicating, but no one’s listening. Next slide.
Eric Kavanagh: That’s a great quote, I like that. “Everybody is communicating, but nobody is listening.” I think that about sums it up in some places. Alight, here you go.
Wayne Eckerson: So, you know, I’ll get to the remedies as well, but a lot of businesses think that the purpose of self-service is to get rid of IT. Well, there’s a lot of counterintuitive things in business, and this is one of them. The purpose of self-service was not to limit IT from the equation but foster greater collaboration with it. Another irony of self-service that I didn’t put here is that it requires a lot of standardization to support self-service. It’s kind of like, think of driving on a road, right? There are a lot of rules that we have to adhere to. Everyone—
Automated Voice: Conference recording has stopped.
Eric Kavanagh: Don’t worry about that. That’s just the backup. Keep going.
Wayne Eckerson: OK. So, and IT really is the group that needs to put together those standards. And once those standards are in place and accepted and adopted, hey, then we can do self-service ‘til the moon comes out. Next slide.
Eric Kavanagh: I think we’re back over to Josh.
Josh Howard: Right, yeah, and I agree with a lot of that, Wayne, that you were saying is. But the thing is, if you want to get more value out of data, again, we’ve got to get out of the business of having IT control everything and getting into the business of enabling. So that means empowering users with their own analysis tools and not just IT. This doesn’t mean that you have to give them the keys to the kingdom. You can do so with those guardrails else existing. Leverage the existing systems in place, leverage your authorization tools, Active Directory, your permissions, and this is going to ensure that, you know, someone isn’t giving data to someone they shouldn’t. And so, by doing all these things is, you are empowering those analysts to deliver greater value and doing it in a way that’s governed.
Next slide. But the reality is that IT is never going to be able to keep up with the variety of different ways an analyst is going to want to view the data, manipulate it. And so, not only that, but you don’t have the time to keep up with those requests as well. The legacy systems, the waterfall processes. If you just look at an ETL process for adding a table, it can take, you know, weeks if not months in some cases. And so, you want to be able to keep pace with that change of business.
If you want to, in fact, create a culture of analytics, you’ve got to enable those users to do that. And then once you do that, the benefits can be truly amazing. You know, when we first started talking about five/ten years ago, business intelligence projects, I mean it was often quoted 70–80 percent of all BI projects would fail. And that’s just not the case anymore. When you arm business users with the right tools, we’re seeing some tremendous results and tremendous value, and that’s the reason why self-service tools are spreading like wildfire through an organization. That’s because of the success that we’re seeing.
And I’ve got a use case that I’ll talk about here in a minute as well, but, you know, we literally have tens of thousands of users doing self-service analytics and scale. And these users are delivering insights faster, they’re creating new products, and they’re reacting to changing business conditions a lot faster in order to stay ahead of the competition.
You know, the second thing is that, you know, they’re also spending less time prepping data and more time doing the analysis. It’s just another component to it, and I’ve got an example here from CNA where they had a number of analysts that were taking time-consuming approaches, that were taking weeks or months and now getting those down to minutes. That’s without exaggeration. We literally have lots of these examples of customers doing this, and this is truly a win-win scenario. Analysts are happy that they’re not having to, you know, they’re getting to their data faster. IT’s happy because, you know, they can focus on their strategic initiatives without fretting about governance, and then finally the executive teams are happy because finally they’ve got business and IT teams working together to create that analytic culture. Back to you.
Eric Kavanagh: Alright. We did have another poll, so you should be able to see those results out there in the audience. We should see that already in your polling panel, but the question was, “Has your organization received the promise of self-service?” I can tell you that the respondents have a resounding, “No.”
I think that speaks for where we are on the industry, but I think you’ve made a couple of really, really good points there, Josh, namely that enabling self-service, albeit with some standards like Wayne was discussing, does in fact allow you to build in governance. That’s the guardrails we’ve talked about, right? The governance policy can be phased into the delivery system, and that’s when you actually achieve governance while empowering the analysts to be self-serviced. Is that right, Josh?
Josh Howard: Yeah, that’s exactly right.
Eric Kavanagh: Yes, so the respondents—
Wayne Eckerson: So, Eric, those results are interesting, you know. I would say that the cause of that is either IT’s still in control, users aren’t getting self-service and getting what they want when they need it, or, you know, they have under-governed self-service. And both are bad. So, it’s hard to actually hit the needle with self-service, to have a governed environment that gives users all the information they need and functionality they need to get the insights they need and take the action they have to. It’s hard, hard, but, you know—
[multiple people speaking at once]
Wayne Eckerson: —you’re faced now with the tools like, you know, Alteryx, very powerful tools, very powerful. So, we have the ability now that we can—
Eric Kavanagh: And you have several reasons your raw deal with Sonic went under a little bit, so just watch out for basic audio [inaudible]. I’m a little bit surprised, and I think that this is actually probably good news for Alteryx because they have a solution to enable self-service. Because in the old way of doing things with lots of different tools, for example, with lots of integration points, people are kind of running around, just trying to keep up with the status quo, and I think that’s one of the real challenges.
One of our clients had a comment a few weeks ago that’s been ringing in my ears ever since he referred to the “tyranny of urgency” and how that tends to dominate several organizations and prevent change. You’re always urgent state, you’re always running around just trying to get stuff done that already needs to be done. And that basically prevents you from doing new things.
At a certain point you have to stop the music, recognize one chair is going to go away, but the rest of the chairs need to sit down at the table and start tossing some collaboration until we work together. But that’s kind of how I view this whole picture. So yeah, the answers typically were 23 of 43 said, “No,” 6 of 43 people said, “Yes,” and 6 of 43 people said, “Not sure,” but 38 people or so didn’t answer. But that’s a pretty resounding, “No.” With that, I want to get into a case study.
I’ll hand it back to you, Josh. Take it away.
Josh Howard: Yeah, and so earlier I talked about, you know, this collaboration between business and IT. I really do feel like we’ve seen some pretty big changes, and more and more organizations are moving in this direction, enabling self-service and seeing those results that I was talking about. And Ford is a great example of that. Ford’s been, of course, using data and analytics for decades, but like a lot of organizations, it was really just done in pockets of the organization. There was little oversight to consistency and coordination, and, you know, they also had data governance practices that were inconsistent.
And so they had a huge issue; they had over 4,600 data sources, and so, you can imagine the challenge of doing this at a size of a company like Ford. And so what they did was, going back just two years ago, they formed the Global Data Insights and Analytics Unit, which is a centralized center of excellence, consisting of teams made up of, you know, data workers, so data analysts, data scientists of the sort.
You can think of this COE a lot like an HR department or a finance department that serves the entire organization. That’s exactly what this new team was set up to do, and so they were able to identify and go after their own high-priority challenges and work with different business units that tackle, you know, different problems. But the whole idea was that they wanted to aim and change that conversation to focus on the business challenge itself, right, and fulfilling those business needs. And, you know, they’ve started with one data analyst to begin with a couple years ago, and one Alteryx license, and a combination of Tableau and QlikView.
Now, they’ve now rolled Alteryx out to over 1,200 data scientists in the last two years, and they’re hiring more. And so, it’s been really amazing to see that take place within their organization and use cases that they’re solving are unbelievable. They’re using Alteryx in order to solve manufacturing line issues all the way down to their NASCAR races, so it’s really fascinating to see some of the results that they’re driving. And, you know, what’s interesting is, you know, some of these use cases, single use cases are saving tens of millions of dollars, and so it’s very easy to justify for them. And that’s just one use case, and it’s now being literally used across hundreds of different business cases and across those 1,200 data analysts and data scientists. So, phenomenal results and we’re really pleased with the partnership that we have with Ford.
Wayne Eckerson: Alright, this is my slide. So, you know, I teach a [inaudible] class on self-service analytics, and this is kind of a summary, a very high-level summary, of the solutions that I bring to a table for the audience. And I’ll try to explain this pretty quickly. You know, I see self-service, well one, there is no one self-service. Everyone has a different definition of self-service inside an organization, so what’s self-service to a CEO is certainly not self-service to a data scientist. But in general, there are two classes of users. The first class, you know, more casual users, executive managers, frontline workers are in the top-down world in blue.
And, you know, I call them “data consumers” or “data explorers,” and they’re pretty much thinking output, you know, reports and dashboards, hopefully interactive that people built for them, either IT or their colleagues, and consuming that as is. Explorers tend to open those things up and edit them in place, but they don’t necessarily want to start with a blank sheet of paper. No way are they being paid to do that. Not paid necessarily the analysts. That’s what the people in the bottom-up world do, the data scientists and the data analysts, who have additionally, data analysts work with spreadsheets, access to databases. And data scientists have more pull with, you know, the data mine workbenches. A lot of the self-service tools that have come out have really empowered this bottom-up crew. It’d be much more productive than they ever could do before. They cannot only, you know, do their own reports and dashboards, they can also go get their own data, blend it, match it together, and so on. I’ve actually seen this triumvirate of tools come out and import the bottom-up world. The data catalogs so they can go find the data either prep tools so they can match it together, and data visualization tools so they can analyze, visualize, and share that. I think we’ll see that tool set become one, and I think actually Alteryx is just on the way towards doing that.
So I call this bottom-up world “real self-service,” whereas the top-down world I call it more “silver-service” because we’re kind of giving information given on a silver platter. It’s been pre-packaged to some extent. Still interactive, still editable, but someone had to think about who the people were that were going to consume this and tailor it to meet their specific needs. You can see in the top-down world you’ve got, you know, the more heavy duty centralized groups, the data governance committee, which, you know, puts its [inaudible] on data sites and reports. And the data warehousing team that tries to integrate data for decision-making. That’s a more traditional IT-oriented centralized top-down governance process. Whereas in the bottom-up world, which is more like 10 percent, 20 percent of the organization, they’re getting governance from the grassroots level by actually opening data sets up, looking at them, commenting on them, tagging those data sets – basically building shared mean of the data from the ground up. You’re getting catalogs and data marketplaces, and an organization needs both these worlds. In fact, they feed each other, very synergistic, they’re two sides of the same coin. If you don’t have analysts out there in every department, operations fail, the marketing, finance. You’re missing all kinds of insights that you need to drive the business because they’re generating answers to questions that people couldn’t have figured out what they were the day before. And certainly IT couldn’t or developers couldn’t build those reports or dashboards. So they’re kind of substantiating the next wave of the requirements and the next wave of insights that should be packaged up and put in the top-down world.
Now the problem is when the bottom-up world publishes reports to the top-down world that haven’t been certified or governed, and you get conflicting reports, duplicates, and things like that. So, in my world it helps to have a data governance gateway between these two worlds, and that is alright, if a data analyst started [inaudible] creates and comes up with a new insight and builds a report. People like it, and then, you know, they want to continue to publish that report and share it, perhaps more broadly to the entire enterprise, it needs to be reviewed by the data governance, and hopefully very quickly, to ensure it conforms to standards. It may need to be written into a standard platform, new data may need to be added to the standard enterprise repository. And what we’re seeing now is the tools like Alteryx are actually embedding the workflows needed to support this promotion process where we’re promoting in a report that’s become popular to get a watermark or a scale as enterprise-caliber certified report or data set. So, that’s some of the data governance state weighed in a nutshell as a review process. There could be a production handoff with development teams, and there might be permissions and governance built inside the BI tools, the analytic tools, or those workflows. Next slide.
Eric Kavanagh: Alright, I think we’re back to Josh on this one.
Josh Howard: Yeah, and so, you know, when you talked about the moving from a number of these different tools, and what I’ve found in my own, you know, research is that most analysts are using 10 to 12 different tools in order to get their analysis job done. And, you know, they may be using a data cataloging solution to find the data, they may be using a data prep solution, they may be using a data visualization tool, something for advanced analytics, predictive analytics, and data science tools for deploying and managing that. And we really think that this should be served through a single platform, and we think that’s where the industry’s going. And so, most people know of all tricks towards data prep and blend capabilities and its tight integration with tools like Tableau and Power BI.
But, you know, we’re way more than just data prep tool. We are really an end-to-end platform for those data analysts and citizen data scientists, providing the ability to discover that data, prepare it, blend it, analyze it, and do it in a repeatable way and a repeatable workflow. And then deploy and share those assets to the scale, and so it’s really what Alteryx is all about. And we’ve got an amazing community that we’re backed by which is, you know, more than just your typical community. It has self-service training areas, it has forums and best practices, and we really have an evangelical community of users there supporting each other. And the great thing about this is as you’re adopting tools like Alteryx, these types of communities really reduce the learning curve, so you’re able to get up to speed faster on these new tool sets. Even though they’re really easy to use, they don’t require a lot of coding, and they’re easy to use and get up and running faster, but still having that community to reduce that learning curve is really invaluable.
And so the way that we’ve broken it down into is four areas. First is it’s really around the discover and share, so before you can prep and blend your data, you’ve got to be able to find it. And that’s the reason why the first part of our platform is that discovery and sharing component that we use to capture your organization’s tribal knowledge. So this is basically a data cataloging solution that’s used to share curated and governed data sets. It lets users find the data that they’re looking for in the easy-to-use Google-like search feature and provides also social features for collaborating on data sets and even lets you drill down into the data lineage of the assets, certify those assets and watermark them. And this is really important for self-service analytics because one is, most people are spending too much time trying to find the data – they don’t know where to go to even find it. And then if they do find a report, you know, how do they know that it’s certified, it’s trusted? So when you talked about that, having a data governance gateway, I really see tools like Alteryx becoming that gateway where, when you do your search, you can automatically and visually see who owns that data, what’s the lineage of that data, how it was created, if it was certified, and how to get access to it, and if you don’t have access to it, you can use the chat features to, you know, request that access. It sends an email to that particular person, and so this is really a good way to productionize a lot of these elements. Next slide.
The next piece is these prep and blend, again, which we’re well-known for, and so, we really view prep and blend as the on-ramp for more advanced analytics. Without writing SQL or any type of code, you’re able to access all your different data, query it – you know, whether that’s structured data, unstructured data, cloud data – and easily integrate all that in memory, shape it, cleanse it, profile it, in order to get your data set ready for analysis. You can also enrich it with third-party data sets. So, we have really good partnerships with companies like TomTom if you’re interested in drive-time analysis, doing spatial analytics. We also work very closely with Experian for household data or [inaudible] for business data. So all of the sudden, not only can you take the data that you’ve got on-premise or maybe in the cloud, you can also enrich it with these third-party sources and really come up with some fascinating analysis. Next slide.
The third piece is this analyze and model component. So I mentioned Alteryx was code-free. Well, it’s also code-friendly too. And so, we offer more than 60 different predictive analytics tools, so when you’re ready to do more advanced analytics, you can use R and Python and Spark-based tools with no coding, or you can actually use and create your own custom packages. So if you’ve got a data science team that is writing R and Python or Scala or whatever, you can utilize that code, build your own packages, and leverage that right within the tool. And again, this is where I think the real value of self-service analytics is, and this is really where we want to help transform the industry from, you know, traditional data analysts and data workers into these, you know, citizen data scientists and doing data science work with really easy-to-use tools. Slide.
Alright, in, and finally we’ve got the last few switches, that last mile of advanced analytics. So if you’re at the point where you are doing data science work, and you’re building your models, the next challenge you come up with is, “Well, how do I get those models into production? How do I manage them? How do I keep them up-to-date?” And this is where our deployment capability comes in. And so, according to our research in the customers that we’ve talked to, less than 50 percent of models ever make it into production. So you’ve employed these data scientists to build all these models, but they’re really never making it into production. And so, we’ve built a solution that’s going to help you build your models, and then deploy those in real time using RESTful APIs.
And so you’re able to get those models and put them directly into web applications and mobile applications faster and easier, because traditional methods just aren’t working. It’s a long, drawn-out process. It can take anywhere from 12 to 20 weeks to deploy a model, and often costs more than $250,000 to do. And then you got to worry about how you keep them updated. So again, we’re looking ways to automate this whole process and take out a lot of the intermediary steps. And so, without really throwing the code over, because the traditional process of what’s happening now is you’ve got a data scientist who’s building his models, and they deploy them, and they throw them over the fence to a web developer who has to take all that R and Python code, rewrite it into some sort of web application or mobile application, and again, it just takes too much time.
And so, there’s no more throwing code over the fence for someone else to do. We’ve automated that process and have a way to manage it at scale. And so, those are really the four areas that we look at when it comes to end-to-end self-service platform for data analytics. And so, it’s, you know, discovering and sharing the data with ease, prepping and blending it, doing the advanced analytics, and then having a way to deploy and manage it at scale. Go ahead. So with Alteryx, you’re able to, you know, talk about the analytic governance and being able to unlock your data in a way that’s secure and offers both code-free and code-friendly ways to do all your analysis, so if you do have data analysts that may not know the semantic, you know, SQL languages to query a database, you can use a drag-and-drop tool that pulls all this data in memory to do their analysis.
Then on the same token, if you have data scientists who are using R and Python, they can still use a tool like Alteryx in a code-friendly way – and the results that we’ve seen with our customers are tremendous because we’re able to provide those repeatable workflows you can take, tasks that take, you know, weeks or months and literally get them down to minutes, without exaggeration. We’ve got a number of case studies on our website where you can learn more about that and some of the time savings that we’re seeing. But, you know, lastly, it’s going to work with your IT organization because it is scalable and break down those silos that I talked about and do it in a governed way. And that’s really what the Alteryx end-to-end platform is all about and why we’re different.
Eric Kavanagh: Alright. That’s all good stuff. I have to say, Wayne, I think you’re really onto something with this data governance gateway is, I think, how you described it. Because we’re in this really interesting world right now in which data warehouses, which have been the trusted source for four decades now, are not really able to keep up with the times and keep up with all the different data sources and data varieties. It’s a fairly rigid system a data warehouse tends to be, and so what I see Alteryx delivering here is really what you could call the next phase in analytic maturity, because they’re allowing you to use all these different sources, but because they have this martialing area with data governance policies baked in, now you really get the best of both worlds where you can have many different data sets, but you have governance, and you can also [inaudible] use all kinds of information and service all kinds of different analysts to get their different perspectives on what’s going on in the business world. But I view this as a fairly significant step in the evolution of analytics for the enterprise, but what do you think?
Wayne Eckerson: No, absolutely. The data warehouses, the repositories of a single version of the truth as they were, and I think it just ignored, you know, organizational dynamic and the roles that people play. And I do see these two worlds of BI or analytics, as you call them. And in most companies, they’re going in opposite directions, and they don’t talk to each other, they don’t trust each other, but really they’re very synergistic, and we just have to get them to acknowledge each other and kind of work together. And tools like Alteryx that incorporate the governance through the data cataloging capability, where stewards can manage the data set and certify and watermark them, which is something that I’ve been talking about for a couple of years now in my classes. Very few companies have been doing it, but it gets so much traction and now I hear it’s everywhere.
And so, the way to blend these two worlds together because, you know, you have your cake and you eat it too. You can let the power users go do what they need to do. Go find the new insights on demand, and then, you know, but you keep it from getting out of control. You keep it from creating the Tower of Babel with some standards which require some governance. And the goal really is to create a culture of governance where people want to go through the governance process. They want their reports/data sets to get reviewed so they’re be consumed more broadly. That’s the goal, and that’s really IT’s new role in this new world. I always say their role is to facilitate, not dictate. And that’s a big mind shift for most of IT professionals who have been used to being in a shared service that did everything for the business. Now the business is doing for themselves, and IT really just needs to be the people, as Josh said, putting up those guardrails.
Eric Kavanagh: Yeah, I think the guardrails are key because they allow the free play, if you will, of analysts [inaudible] do different things, but not get off track. And if I understand—
Wayne Eckerson: Exactly.
Eric Kavanagh: —you correctly, Josh—
Josh Howard: Exactly.
Eric Kavanagh: Yeah, you were kind of talking about how, I’ve actually been tracking Alteryx now since before it was called Alteryx many years ago – I think it was called SRC or something along those lines – and a Wal-Mart was the first customer. And one of the really cool things that you guys talked about way back when was the ability to really understand business processes and workflows. And if you have that strong understanding of workflow and business processes, then you can do a number of different things. First of all, you can deliver a much-perfected user interface if you don’t cloud the options available to the user with extraneous information. Two, you can also streamline processes to better understand where there are choke points or control points. And I think that’s probably part of the magic of why Alteryx has been able to deliver this very governance-friendly, but user-friendly type environment that enables all kinds of different information sets and analytical use cases. Would you agree with that?
Josh Howard: Yeah, I mean it’s, you know, I would, Eric, and a lot of this is just putting these types of tools into the hands of business users and giving them a way to do their work in a business-friendly way that’s easy to use and it’s friendly. I mean, if you think about something like data governance, we’ve been talking about data governance for two decades, and as IP storage, we’ve tried to push this down to the business, and it just never gets adopted, never gets any kind of traction, because it’s not built for the business users, right? It’s IT-led, IT-driven, and it works for IT, but it doesn’t work for those business users. And so, we want to take those same methodologies but apply them to a business-friendly toolset, and that’s our approach with, you know, the data cataloging solution and metadata management.
You know, when I talk to a business user, I’m never talking about a semantic data layer, and how we’re helping manage, you know, metadata. But, you know, on the back end, that’s essentially what it’s doing, those types of things have been within IT for a long time, but for the business user, it’s all about how to find data faster, how to get your job done faster, and providing that information in an easy-to-use interface that they’re accustomed to using, just like in their consumer lives, right? They want a Google-like search interface, they want a social collaboration element where they can network with other users in that organization to break down those data silos and capture that tribal knowledge. And so, we’re just taking a different approach of how we work with the business, but doing it in a way that’s also IT-friendly.
Eric Kavanagh: Yeah, and I got a great question—
Wayne Eckerson: You know the other thing—[multiple people talking at once]—Josh, that struck me in your presentation was, we’re in the age of platforms now. I think we’ve moved past the age of tools, and it’s fine, but the platforms, right? And so, I’ve been covering BI for 20-some-odd years, and in the BI space, we’ve gone from tools to analytic platforms where, you know, one product essentially deports every mode of analytics for every type of user, right? From reports to prediction on a common architecture and self-services. We’re also seeing the same thing on the data assembly side, or data integration side where somebody’s putting together these platforms that ingest data, add it, catalog it, repair it, transform it and make it available to users to download and analyze. And now, what you guys are doing, is taking the next step in many ways and combining those two platforms into one, so it’s a combined analytics and data platform, which, you know, makes sense. That’s the future: convergence. The only thing I don’t see in your platform is your basic reporting and dashboard tools or capabilities, but maybe that’s embedded in your analytical module.
Josh Howard: Yeah, we do batch reporting very well. We’ve got a very robust solution there, but you did hit on a point around dashboards, and we do see this as an opportunity for us to grow. We’ve always traditionally had really good partnerships with Tableau, Power BI and Qlik, but we’ll continue to do so. But what we’re finding is our analysts, our customers, they don’t want to wait ‘til the end of the workflow and that cycle to see their results, alright? They want to see the results as they’re working in real time, and that’s really the direction that we’re going, and with we know what we’re labeling as inline visualitics so that you’re seeing your data as you work, and you can iterate on it and see that in real time rather than waiting ‘til the end and publishing it to a visualization tool or a dashboard to see those results. And so, it just eliminates the need for balancing back and forth in order to get your insights.
Wayne Eckerson: Yeah, well, that makes a lot of sense. And you guys are known now for ease of use. You know, you use the company Tableau on their rise to fame and fortune. You’re right there with them, and who better to kind of take the lead in this converged platform space because you’ve got your foot in both the analytics and the data management. So, we’re beta testing to see how you guys fare in the next couple years.
Josh Howard: Yeah, and you know, I do think it’s interesting, and I’m glad to be a part of this space, and it’s really been interesting to see, get a look at, you know, the data integration space, the business intelligence space, and the advanced analytics space and really see those converging. And, you know, I think platforms like Alteryx are going to really help a lot of those business users excel and enable those users to get access to their data and do that analysis, you know, and get to those insights faster and easier.
Eric Kavanagh: Yeah. All the [inaudible] here, and I agree with you, Wayne, that how it really does make sense, and I think, yeah, there’s a question from an audience member I’ll throw in here. It’s very germane to the conversation. It’s about DataOp. For those of you who aren’t familiar with the term—
Josh Howard: Next slide.
Eric Kavanagh: —it’s really come on strong in the last nine months or so. It started with one or two vendors, then three and four, then five and six, and now lots of folks are talking about DataOp. That’s basically the data management side of DevOp. So what we’re seeing is a lot of focus on really trying to understand what different tools and what different technologies are touching data as it moves through its life cycle and how does that affect your analytical view. And it seems to me that Alteryx actually kind of solves the DataOps problem by focusing on this platform approach before DataOp even became a term. But I’ll throw that over to you, Josh, first, and then you, Wayne, for commentary. Josh, what do you think?
Josh Howard: Yeah, I think it’s an evolving space. You know, we try to be data agnostic, and so being able to access data – whether that be within your firewall, in the cloud, unstructured data, structured data – so because we know this is going to continue to change, you know, and I’m sure Wayne would agree with this, and so would you, Eric. If you go back, you know 10, 15 years in this space, I mean, there was only a handful of databases. We’re now up to over 400 different database types. And so, we’re just never going to keep pace with that. And so, there’s always going to be something new and shiny for an organization to adopt. And so, we just want to be agnostic and use our open technology and APIs to be able to seamlessly integrate with whatever you already have in your organization. And also see the second piece to that over on the DataOp side is really with more and more workloads being pushed to the cloud and new cloud technologies and machine-learning technologies are really pushing us into this new paradigm, and I really think that’s where, you know, DataOps is going to go. And we’re going to see a lot of interesting things happening in that space.
Wayne Eckerson: Yeah, I think another term we use for DataOps is “data pipelines” or “data supply chains,” and we do see a lot of companies coming out, especially in the big data world. You can manage that workload and keep data lakes from becoming data swamps. Yeah, and I would agree that a lot of that is now moving into the cloud as well.
Eric Kavanagh: Well, and you know, so Alteryx made a couple acquisitions. I don’t know if you want to talk about that over the last year or two, I suppose, Josh, and it really fleshed out this platform, in terms of ingesting data and in terms of some of that semantic stuff. And now you really do have this sort of end-to-end solution which enables analytics to govern it. I don’t know of anyone else who has taken quite that focus and approach, and I think it was very clever on your half. But do you want to talk about that for a bit?
Josh Howard: Yeah, sure. And so, it’s been a big year for Alteryx. You know, we went public earlier this year, and we made two key acquisitions that help us, you know, kind of end out our platform. And so, the first one, it was really that data cataloging piece. Again, you know, what we find is what we want to help those organizations govern that data. And so, we actually acquired a data governance company called Semanta, and that has become our data cataloging solution and what we’ve built into the overall platform. Because we do, again, we see governance being a key component to self-service and to enable self-service. And so, again, that gave us all those, you know, metadata management, data cataloging capabilities. And what we’ve done is we’ve built an interface onto that to make it easy to use and very friendly, integrated that with our overall platform.
The second one that we made was a data science company based out of Brooklyn, New York, and that was done in order to build out our machine-learning capabilities as well as the model management piece. And so, what I mentioned earlier was we’ve got lots of data scientists using our platforms and doing very important data science work. However, getting those models, you know, to the last mile was very challenging. And so, I mentioned, you know, the 12 to 20 weeks it often takes, the $250,000 that it’s required to build some of these models. And then, how do you operationalize and keep all these models up to date? How do those models learn? And how do you train those models? And so, that’s a big problem as well, right, the deployment capabilities. And so, those two technologies with the data science side and the data governance side have really rounded out our platform and what we’re trying to do, trying to bring it to organizations, to solve this challenge.
Eric Kavanagh: Yeah, and I’m glad you threw that in there because we had a question from the audience just in about machine learning and AI. And, Wayne, maybe I’ll throw this over to you real quick. To me, there is just so much potential for machine learning to really optimize a lot of the different issues that we’ve struggled with over the years – things like data quality, for example, things like congestions on analytics and helping that discovery side of the equation, right? Because some of these algorithms that keep learning in particular can really go on their own and find some interesting things that could be surfaced for the user. Because one of the challenges, of course, with analysts in general is that every analyst brings their own set of prejudices, their own view of the world. That can be fairly difficult to change sometimes, and so I see a lot of potential for machine learning and AI in the future. What do you think?
Wayne Eckerson: No, absolutely and just basic rules. Those things together will further simplify these self-service tools, make them easier to use. You know, as you said, everything from making recommendations for other reports, for data sets to look at, to adjusting models, you know, calm correlations in the data prep tool. You know, we’ve already had this like Tableau innovated the right visualization for the data set you want to display. So all of that makes these tools much more powerful, makes self-service much more plausible, and helps users use data to drive insight and value faster.
Eric Kavanagh: Yeah, and you know, in the world of enterprise software, obviously there’s so much cool stuff going on, but the bottom line is it always takes time to build technology. So obviously you can go and acquire stuff, as Alteryx has. But when you have experience in a space, you know, there’s an old expression: There’s no substitute for experience. You just know how to do things better, and I think one of the keys to Alteryx’s long-term success here has been that Alteryx was really onto the whole process of using third-party data many years ago. I can’t remember exactly how long, but I want to say six or seven years ago, Alteryx has already baked in the ability to go out and grab data from companies like credit companies, for example, or geolocation data or any number of any third-party data systems. And I think that was the beginning of what we now see maturing in terms of what we call data blending these days, because we didn’t even have that term back then.
But, Josh, I’ll throw it back over to you again. And, me, I think that’s a lot of saturation and experience baked into the Alteryx platform around that data blending concept, which now has just been augmented by ingestion, by machine-learning, by data cataloging, and so forth. I think that’s why we see Alteryx where it is today. What do you think?
Josh Howard: Yeah, I mean, necessity is the mother of all invention, right? And so, you know, it was our customers that were, you know, we, you know, originally doing spatial analytics, and that’s really how we started, was doing spatial analytics. And you know, taking data like TomTom and doing drive-time analysis, you can see, you know, uploading that data with, you know, home data from Experian. So that was really where we started, and what we found was, you know, our customers needed a platform for blending all that data together. And wouldn’t it be cool if we gave them the tools to do it. And so, that was really the impetus of Alteryx.
And you know, what we’ve found is, you know, over the years, is that data prep is really that first step in your analytic journey. So you know, it takes 80 percent of a data scientist’s time, you know, doing predictive analytics and data science work is actually spent doing data prep work, and less than 20 percent actually doing analysis, and so that’s what we’re trying to overcome. And so, data prep is that first step in your analytic journey. So before you start doing any kind of reporting, advanced reporting, predictive analytics, all the way up to cognitive analytics, you still got to access data, you still got to prep and blend it and pull it together. And that’s what we’re solving with this platform. And enabling those users to do all those things in both a code-free and a code-friendly way.
Eric Kavanagh: Yeah, and I love that concept, too: code-free and code-friendly. Because the fact is you do have a lot of code jockeys, which can add tremendous value, but there are lots of business users who are frankly turned off by code. They are intimidated by it, and who can blame them? So, Wayne, I think that’s also a nice feature, a nice approach. There’s code-free and code-friendly, right?
Wayne Eckerson: Oh, absolutely. Yeah, that’s how you get more and more people onto self-service.
Eric Kavanagh: Yeah, and self-service, I think, is the next big step, and I really like what we’ve discussed today, so it’s about how really thinking through your processes, your work flows, your data life cycles, and so forth. And baking those policies into the platform, to your point Wayne, there are some issues around standardization, you do lose a little bit of flexibility, but once people understand the methods of the madness, you wind up really shepherding the process forward such that in-users understand they now can get what they want. They don’t have to wait on IT, and it changes the nature of how IT and business people work together, I think in a very positive way, because now IT can serve as the enabler, they don’t have to be a gatekeeper on technology as much as they used to. There’s not as much support, ideally, if you have some standards. So you wind up fostering greater collaboration because that’s the whole goal, right?
So for closing comments from first Josh and then maybe Wayne.
Josh Howard: No, I mean, you know, I agree with everything you said. You know, it’s important that we give both IT and the business users the tools they need to be successful. So, we think that IT shouldn’t be in the business of creating reports. That should be left to the business user who has that context of the business and the data that they’re using, but do it in a governed way, and something that’s going to work for IT as well.
Eric Kavanagh: Alright, closing comments from Wayne.
Wayne Eckerson: Yeah, IT’s role has changed from one of doing it all to facilitating self-service and really being the champions of culture of governance and getting the users to want to govern their own output, for their benefit and the benefit in the organization. I mean, IT’s role is— I feel sorry for IT, you know, because sometimes they do have to go in and build it, divisions in business antics like legal and HR typically, I’m not going to do any of that. And certainly if you want something that’s cross-functional enterprise, who else is going to build it but IT? But in general, yes, IT has to change to thrive in this world of self-service. They have to be in a more supporting role rather than [inaudible].
Josh Howard: Yeah, and I think with the next evolution with the centers of excellence and where these projects aren’t being led by IT or the business, but rather a centralized organization. You know, we’re starting to see the rise of the chief data officer and these types of projects falling in that realm where they both have the governance perspective as well as the business perspective. I think that is a best-case scenario for creating that data and analytic culture, and I’m excited to see what comes of it.
Eric Kavanagh: Yeah, we had a couple last-minute comments from attendees coming into the chatroom and also the Q&A. I like this comment: Govern the output, there’s no ambiguity as to who’s self-service report is correct.
Josh Howard: Yeah.
Eric Kavanagh: Yeah, that’s good stuff. It’s all about collaboration, it’s all about working together, and, you know, Josh, you mentioned too, the importance of having users talk to each other, and that’s something that Alteryx focuses on as well.
So, folks, we went a bit long here, but we started a little bit late, so I want to thank you so much for all of your time and attention today. We do archive all these webcasts [inaudible], so feel free to share them with your colleagues.
And with that, we’re going to bid you farewell. Thanks again to Wayne and, of course, to Josh from Alteryx. We’ll talk to you next time, folks. Take care. Bye-bye.