JPMorgan Chase & Co. - Jason & Steve's Excellent Analytics Adventure - Inspire 2017

Fast, modern and agile are not terms most people would associate with the large banks... until now. Join Jason as he travels back in time to assemble stakeholders and bring them together to save the future of business intelligence and analytics.



Video Transcription


Jason Mack:
All right, we'll go ahead and get started. This is the Ace Track session. Jason and Steve's Excellent Adventure. Hopefully you are in the right place. Take it down a little bit. Everyone's well caffeinated after the coffee. For those I haven't met, I'm Jason Mack at JP Morgan Chase. I work in marketing, I run an analytics team that helps our marketers understand the effectiveness of their campaigns. We try to give our leaders a holistic understanding of our customer relationships.

I want my analysts to always have the best possible tools to let them solve data problems quickly and enable better decisions. Steve Hittle is a leader in our IT organization. He manages the business intelligence center of excellence. Steven is not here today. He was originally planning on presenting this session with me, but Steven was selected to participate in a volunteer mission, so he's actually currently in route to Northern Guatemala right now. Where he's going to spend the next week building a school for rural children. IT's not all bad guys, they do nice things periodically.

It's just gonna be me today telling you about our story and it's a story about building effective partnerships between the business and IT organizations. When Steven and I joined Chase several years ago we both found this landscape of extremes and incompatible views. Within IT we had the traditional enterprise approach to business intelligence based on the software development life cycle model and layers of governance. Our IT groups were custom to delivering a few big things pretty infrequently.

That contrasted with analytics in the business and almost appeared as an underground. Disappointed by those lengthy IT timelines, the business groups figured they had to get things done on their own with whatever tools they had, even if it was just Excel. Although they were working to solve the same problems as their colleagues in IT, more often than not it felt like they were working against each other, so Steve and I thought there had to be a better way.

If we're going to move the culture forward and make things better at JP Morgan Chase we needed to balance everyone's needs. IT needs to have the reliable controls and governance models that you need to achieve scale in a massive enterprise and then the analysts needed to have flexibility and tools to get their jobs done day to day.

Transforming that culture and getting to where we are today was quite an adventure and I'm going to share that story with you today and talk about the issues, technical, cultural, that we had to overcome and share some of our learnings so that you can apply them in your organizations and realize the benefits of driving analytics driven culture. I'll share the challenge we're still working on today and what we're thinking as we look ahead to the future.

As Steve and I were reflecting on this analytics journey that we've been through the past several years at Chase, we realized that it bore a bit of resemblance to one of our favorite 80s movies. Bill and Ted's Excellent Adventure starring a young Keanu Reeves. Who remembers this film? Anyone? All right, excellent.

In it you know that Bill and Ted, they have this phone booth that they have to go back in time and collect historical cast of characters to finish their high school project and ultimately, they didn't know it at the time, but save the future of mankind. We didn't have a phone both and I don't believe, I'm not vain enough to believe that enterprise analytics could save the future of mankind. We certainly did have to partner and work across a massive organization, bringing together stake holders, working to advance their thinking and hopefully, at least, saving the future of analytics at JP Morgan Chase.

Let's go back in time and revisit this analytics adventure. To set the stage it's useful understand that JP Morgan Chase essentially operates under two main brands. Chase, which includes the consumer and community banking business of things, like your local retail branches, credit cards, mortgage, auto and in the JP Morgan brand, which includes the corporate investment bank and asset management businesses.

Each of these lines of business has a high degree of autonomy and to enact change throughout requires a proven record of success. None of them are just going to change the way they do things because you tell them to. You have to prove that there's a better way. The culture at a bank can be best described as ultraconservative given the regulatory environment. Our CEO Jamie Diamond is famous for saying, that any regulatory issue there's five or six agencies that we have that we have to work with and sometimes there's agencies aren't even clear as to who is responsible and who we have answer to.

That results in very little risk tolerance and defaulting to historic norms just the way we've always done and the strong preference for good process. When Steve joined Chase he came into the IT organization take over an analytics team that was taking months to complete projects and not really meeting the business needs. He wanted to improve what his team was doing. He inherited an IT status quo that is based on the SDLC process of, "Whose seen this before?" Right? We're going to go and get requirements, get a bunch of sign offs on that, go design something, go get the sign offs again from people again, build something, test it, get sign offs again. Maybe six months later deploy something into production that probably no longer even meets the needs or the business has moved on.

This waterfall approach only seems to lead downhill, right? I was recruited into Chase to build an analytic practice for our marketing organization, give them dashboards, enable better decisions across all those different lines of business, all the marketing channels, and across all of our customer relationships. When given that task and started digging into it my reaction was a bit like Ted's here. A bit shocked figuring out where I'm going to get all this data to build what I need.

I also found, what I refer to as the analytics underground. We had these business groups that felt neglected by IT and they had problems and critical questions for which they needed answers and they needed them today, not in two weeks or two months or six months. Then questions are coming in faster than answers could be deployed through the traditional sanctioned approaches. How had the business responded? We saw them creating their own Excel and SAS data marts. They liked their analysts could come up with their own answers without waiting on IT but this kind of build database under your desk mentality had a significant downside in that there was a complete lack of consistency in results. No two groups could give the same answer to the same question.

We needed a better answer. Steve needed a way to get faster results from his COE team and although there were internal processes he needed to change, he knew that traditional tools were a major limiting factor for us and he wanted to test new software to prove that their team could do better results in conjunction with the business partners. I wanted better option for my analysts to be able to answer questions faster. I wanted to make it easy for them to get to the data they needed, no matter where it exists and solve problems without the need for sophisticated programming.

We started looking around and through the tablet community we kept hearing about this tool called Alteryx that seemed to have the promise of being able to connect to any kind of database you want, build things in hours that typically takes weeks or months. We looked at Alteryx, got a demo and our reaction was a bit like this, "Whoa," right? How many of you felt that way the first time you saw it, first time you got your hands on the tool. Really blows you away with what is possible, so we're super excited about it, what might be possible.

It's really cool that Alteryx offers free trial and encourage analysts to download today and get started but using internet in a large enterprise like Chase feels a bit like this. Go and get my free trial and oh cyber security policy, "You can't download software." Nobody was going to forging their own path. From the beginning we knew we have to win the support of others, bring them into our phone booths, so to speak, if we're going to be successful in bringing Alteryx into the organization.

The first person we had to bring into our booth was IT risks. These were the gatekeepers and some of their concerns are around, how are we going to ensure, we're going to give this tool to business analysts. How are we going to ensure that their logic is correct. Part of the discussion that we had to have with them and win them over was to make them realize that this is already happening today whether or not we want to realize it, right? This is happening in Excel, it's happening in SAS. Using this tool will potentially give you more controls than what we're doing today.

There were also questions around who should own it, so we bought a few licenses to get started and the analysts that wanted this better way were quick to adopt Alteryx. IT had permitted it, we got it through IT risk, got all the sign offs, got it packaged, brought into the organization. The ones that have the most pain with the data, individual analysts were quickly embracing it. Then we found that there was a strong crossover, that people who were good with Excel and Tableau picked up Alteryx very quickly. At this point, IT as an organization still largely ignored it, right. They permitted it to be brought in but were convinced that we've got these other enterprise reporting systems. We've invested millions in them so it must be working and didn't really want anything to do with it. So it was growing, from the analysts base.

Which led us to start to realize something, if we just give a tool to lots of analysts and let them all go do their own thing, that's starting to bring us somewhere familiar. In our first year I was the first licensee of Alteryx at the bank. I think we're up to 60 by the end of the first year and if we just kept buying designer licenses and giving them every analyst who goes their way, we weren't really going to be any better off than in the days of Excel and SAS.

Sure we would have a better tool, but we haven't solved the overall problem of working together ensuring consistent results. We did need some amount of central coordination. We needed a server environment. The analysts, of course, would have been perfectly happy with designer alone but passing around package workflows, data sets, would have only been a marginal improvement over Excel. We needed to answer the question was, "Who is going to run this?" Is this going to be something run in the business, run out of IT, who's going to manage it. At Chase the business is not allowed to run servers so we had to engage IT management to get a server environment set up.

Their initial reaction was not as positive as we had hoped. "Blow them up," as Napoleon said here when he sees Bill and Ted in the booth. There's something about a piece of software that's fast growing, flexible, and business friendly that seemed threatening to a command and control style of IT.

We had to win them over, we had to bring them into our booth. How we won them was planning and working through processes that were familiar with them. Instead of fighting them from the beginning, right. Working within the frameworks and then what really helped was focusing on the outcomes we were trying to achieve for our analysts not the tactical decisions of how it would be set up, configured, managed along the way

An interesting thing happened as the analysts that used Alteryx. We started to notice they were actively seeking each other out to ask questions and share learnings and define best practices as they were working with the tool. This community was not something we had generally seen before in analytics. With traditional tools analysts would tend to actively avoid sharing knowledge because they're knowledge of a proprietary programming language they viewed as like their value proposition to the bank. With something Alteryx we found that the analysts had a better understanding that their value is not holding them to a little piece of knowledge, but, in fact, spreading it and driving the culture forward if you want to become analytics and data driven. I don't think a lot of people think that working at a bank sounds fun, but for these analytics teams that had Alteryx and were doing this kind of stuff it really was.

So instead of waiting for IT to tell them how they should be using their technology this community were active participants in their own future. As the Alteryx user base grew they were better prepared to work with IT as Alteryx was adopted more broadly and we helped to fuel this by getting analysts to share their work on internal lunch and learn type sessions., internal virtual user groups, and departmental meetings that brought together analysts from across different lines of business to show and discuss their work within the tool. It's a lot easier to discuss specifics within an organization.

Now we've got IT's full attention. Some additional questions started to being needed answered. Things like, "How is this tool going to fit into our existing technology stack," "How are we going to support it with this new develop model?" And, my favorite, "Why does it cost so much?" How many of you get this question? We needed to the win support of the line of business CIOs, who were concerned with questions like those as well as "How can we onboard and gain efficiencies bringing multiple lines of business on at once?"

How do we do that? Winning over the CIOs is all about demonstrating value, so we would actively push analysts to share their use cases to continue to build the momentum. There's a really useful framework for doing this, it's very simple, right? We had them define the problem, describe their ideal outcome, highlight what the particular challenge was. Maybe it's something about the data that shows how you solved it in Alteryx, including how long it took you and then talk about the benefits, so fine frameworks. It doesn't have to be formal. We would give them the structure and ask them, "After you've done something, write up an email and share it." We do have internal blogging and social media platforms and they had used that but even something as simple as an email and you start to collect those so that when that pricey subscription comes through the next year, you built up a library of wins that makes it very easy to demonstrate the value of Alteryx to your CIOs and continued to use.

We were basically developing a new paradigm in how the business was working with It. Previously the business felt they were being imposed upon to simply check boxes like, "This isn't giving me any value." In our new approach, right, it's a true partnership where were defining those problems, setting up a true proof of concept for pilot, right, we'll buy a license, get started. Providing the support, mentoring, and training so an analyst can get up to speed quickly and that builds trust that keeps the analysts happy and productive. They don't just feel like they're checking boxes.

You know how everything's going great right up until the time it isn't. In the past several years at the bank things like the formation of the Consumer Financial Protection Bureau, the London well incident at JP Morgan Chase resulted in this major focus on the controls environment and that impacted all aspects of the bank not just operational areas or trading areas, it affected analytics. Now there was this new person called a control officer that we found out that we had to win over and bring into our booth.

This person is concerned with things like, do you have a well defined process and then, as I look through the process, where's the risk and then what controls can be implemented to mitigate that risk. To address control officers concerns we had to align processes across different analytic teams and one of the things we were able to do is leverage Alteryx in many cases to build automations that reduced a lot of the manual touch parts that were still present in analytics processes. Frankly it was a painful process to go for and, at times, it felt as though this intense focus on the controls environment was preventing other work from getting done, but we know we're in much better shape today as a result of going through the process.

What does it look like today? Alteryx is now in use across nearly all those lines of business that I had on the slide earlier. We're seeing our analysts are getting to productivity faster and that's both analysts that are new to a group, new to a department or existing analysts that are just new to a data source, introducing a new data source because they have Alteryx. Everything is not perfect. Shocking, right? There's still some challenges that we're working through today.

One of those is around enterprise server structures. Things like figuring out what the migration processes should look like between development and testing and production environments and how we should establish version control. It's been nice in the past few releases of Alteryx with burgeoning added on the server and also, on the server run, they did data capture and data lineage. This is something we've been trying to work with Alteryx to solve, so we're very excited this morning to see with the announcement the new tools that will fit into this ecosystem and help solve some of this as well.

We also need to monitor our server environments better, right? To understand things like data connection optimization and performance so in version 11, if any of you notice, there's data connections. Basically it can be moved up to the server and then you can connect through designer onto an established data connection on the server. That's great because we can build very good connectors there and ensure people are getting to the data in the right way and then, as far as monitoring, it's balancing the use of Alteryx server as a scheduled data factory versus the on demand usage that you would get out of exposing gallery. How do you need to size it appropriately, depending on how you're using it.

Finally, new version cadence, as new versions of Alteryx come out a lot faster than we can keep up with. We have a very defined process, we have to go through for its security scanning and getting things packaged and certified and signed off. I still have 10.6 on this laptop now. We just got 11 packaged, so I'll be pushing that out to my users next week or so. Balancing all those great features that Alteryx is bringing is, those capabilities versus the processes that we have to go through internally and making sure our users feel like they're getting the benefits of the new versions that are coming through their subscription.

This is how my team operates today. We have a sampling, this is a small number, there's actually many more than that, but I felt that was pretty representative of data sources at the bottom there. Everything from Excel files, traditional data warehousing technologies like Teradata, all the way to modern big data such as Hadoop and modern MPP data platforms like Greenplum.

We use Alteryx to bring all of that stuff together to do our advanced analytics and data prep and pipe that both to Tableau desktop for Tableau desktop users and as published data sources to Tableau server where my team builds, we have about five dozen dashboards supporting about five thousand, six thousand business users. Alteryx is core to what we do, even though Tableau dashboards tend to be the most front facing thing that the business sees

What's next for us as we look ahead? We're looking to continue to build out this notion of self-service at scale and to do that we're looking to expand governance so things like published and certified workflows, editor available for reliable data sources, giving our analysts guardrails so that they can have repeatable success. Again, some of the announcements from this morning's keynote fit right into this, so very excited to see those new features starting to be built in. We want to continue to make sure that analysts that pick up the tool, we can accelerate onboarding so we're looking to curate templates, work with templates, and preconfigured data sources.

Once an analyst gets Alteryx on their desktop they're good to go from day one. One of the things that we do through that packaging process I mentioned is we can do a lot of customization, so instead of just installing Alteryx I can ensure that certain data configurations and all would go along with that package when it gets pushed down to the users from IT. We're trying to spend a lot of time there to make sure that the package that they get has everything they need and they don't have to spend a lot of time setting up data sources and getting Alteryx plugged into all the things that they need.

We're also trying make sure that Alteryx integrates well with existing processes to avoid that perceived rebuild cost and people avoiding improving things because not wanting to have to redo code or things like that. I think the amusing thing is with that, almost every time that we've cracked open old code and done something with Alteryx. Not only can you improve it, but we actually find errors in the old one that no one realized was there and then we want to quickly want to get to the new one and put the old one away without dwelling too long that.

We're also working through the challenges of we call the big data evolution, so large enterprises like mine are migrating a lot of their data, almost all of it out of traditional data warehouses into various flavors of Hadoop for a variety of reasons, performance, cost and data's always been big for a bank whether it was in the traditional ones or in Hadoop. With a company that does 14 billion credit card transactions a year, but now this new platform presents new opportunities for us and one of the things I'm excited about having Alteryx through this transition has been a major advantage for my team.

Things like having the in database tools for Hive and Impala connectors is really giving our analysts the best of both worlds so they're able to leverage the easy to use Alteryx interface but then also have the power of a large, several hundred node cluster to process the data when they built workflows using the Hive Spark and Impala in database tools. Finally, the data security involved with that big data elements, still needing to set up and maintain access roles and monitoring for private data and limiting data movement.

There's three key lessons that I hope you can take away for your analytics adventure. The first is, partnering early with your IT and business stakeholders because you're not going to get far alone, second is promoting a community because you can really scale analytics once it goes social and, finally, don't be afraid to ask "why?" Always challenging assumptions and pushing to improve in the right way.

Alteryx has been an amazing adventure for us and I hope our story's helpful in making your analytics journey most excellent.

Questions?

Crowd question:
[inaudible]

Jason Mack:
Mm-hmm (affirmative).

To clarify, the question was around why wasn't IT involved. IT definitely was involved from the beginning, we had to get their permission to bring it in. We got permission, got it packaged, we started using it but they didn't, when I say they weren't involved. They were involved in supporting Alteryx as a service standpoint. Like running a server on, engaging it that way, but definitely were involve from the beginning, right, and taking the connections to all the environments and all.

Crowd question:
[inaudible]

Jason Mack:
In ours we have a model where business users can publish to development environments and to these, one of the things we does create a special environments that are sanctioned for production date but business users can publish to then we have traditional quote "production environments," right where business users will complete, after completing their testing, submit a ticket and then IT might reach it there for them. We have a model where they can do all of those.

Crowd question:
[inaudible]

Jason Mack:
Mm-hmm (affirmative), exactly. That was featured as like adapting, instead of saying, "No, we're going to do this a different way." is like taking a tool that's going to make it better and adapting it to the current process and then pushing to improve where you can, like I mention this can build environments where we're allowed to touch production data there.

Crowd question:
[inaudible]

Jason Mack:
It's much faster, exactly. It's heck of a lot faster.

Crowd question:
[inaudible]

Jason Mack:
We look at a lot of different tools, so if you can name any data prep tool up to buying Alteryx, I probably tested it. I think, one of the things that we liked about Alteryx is very much, we've been using Tableau prior to that and because they have this designer and server model, I presented it as a very low risk option, right? Some of the other ones were it has to sit in your data center, it's a big project from the start and with this we said, "Let's just buy a few licenses, try it, and if we don't like it we just won't renew a year later."

That seemed very reasonable to them. We tried it and then we went down that process of building and documenting those use cases, so we kept buying more and said, "Alright, now we need a server environment," and grew it that way.

Crowd question:
[inaudible]

Jason Mack:
I think they would like to be able to push stuff to production, but that's just one of those things that's not really negotiable currently. We're pressing, getting closer and closer to that and I think if we can build automations within Alteryx that do the checks, things like that, we reduce their manual touch points but it's just one of those things that you're never going to overcome

Crowd question:
[inaudible]

Jason Mack:
What's that?

Crowd question:
[inaudible]

Jason Mack:
Mm-hmm (affirmative), exactly.

Do you want to give people the mic or I could repeating questions.

Crowd question:
[inaudible]

Jason Mack:
IT doesn't, the business is doing the checks themselves, right? The business owns responsibility for the data and the calculations and everything that they're building. IT is just migrating the information to make sure that they're meeting their audit requirements.

Crowd question:
[inaudible]

Jason Mack:
Separation of controls, exactly. Mm-hmm (affirmative), yep, purely compliance.

Crowd question:
[inaudible]

Jason Mack:
Yep, we do. In Alteryx and that's what I mentioned, the building of data connectors up to the server so most teams, if you have a team, you will use a service account and get that set up with access to the database and then you'll set that up within Alteryx. It used to be called alias data sources and now it's called manage data sources when they moved to server in 11. That's something I mentioned we package, make sure our users have, so everyone on my team gets the preconfigured data sources with those functional ideas that they need for the different environments.

Crowd question:
[inaudible]

Jason Mack:
Yep and those departments would have their own ideas, sure.

Crowd question:
What do you guys use to share best practices across organizations, do you have a center of excellence or a wiki or something along those lines?

Jason Mack:
Yep, a lot of different things. We do have, Steve runs the center of excellence. We have an internal platform we call it our internal social media platform. It's like a forums type of thing and is relying to interests so we have an analytics one where people can both either ask questions, post comments, things like that and you flag them as such. If someone posts and tags an Alteryx question it comes to my inbox, I can answer it, things like that.

We do lunch and learn sessions. Where we have people demonstrate things that they've done and then we also, just as simple as creating [sherping 00:30:34] pages, things like that. We outline, "Here's how to do things." A lot of the Alteryx training I hold internally off of platforms. Because we're so distributed, JP Morgan Chase has 250,000 employees all over the world.

Doing stuff in person can be difficult so we do a lot of virtual user group meetings, so we'll get together on a WebEx with cameras and everything and talk about either specific topics or just bring a doctor session kind of thing. Like come with your problem and see how it helps out. We do a lot of different things to achieve that.

Crowd question:
Hi. In Arback, I feel like we're five years behind where you are, almost exactly, but I feel like a lot of the data analytics on people that we work with are very tethered to the idea that the only thing that they can use comfortably is Excel. How did you change that culture? Could you describe more deeply how you changed people's minds about the new tools?

Jason Mack:
My group, it's probably going to be different for each one, so my group were already Tableau users. That's how we found Alteryx was through that community. They had already evolved beyond Excel, but I certainly do a lot of demos and show Alteryx to finance groups, things like that, that are very heavy in Excel. I think it's just really a matter of demonstrating the speed with which you can get things done and just the hard limits.

Anyone whose done stuff in Excel they know not just the row limits, how difficult and clunky it can be to get things done. Just showing them the benefits of a better way of doing it helps a lot, but our team it was an easier sell because we had bumped up against the limits of Tableau and needed the tool to make it a better. It's a lot easier.

Crowd question:
[inaudible] and server course. Did you face any problems from procurement and understanding how this fit into the tool set?

Jason Mack:
No, not really. The way it works internally for us, once we've approved the vendor and this specific version of the software and everything it gets loaded up to an internal catalog and anyone can go and buy it as their department financially signs off for it. Until, at our scale, until you reach tens of million dollars to spend, procurement doesn't bother to care what each line of businesses is doing because, again, they each have their own P&L, so once it's approved and if they want to buy it and they can justify to themselves, they can get it.

Questions on that side?

Crowd question:
As you start looking at big data, are you seeing that your Hadoop cluster is a source or a destination for Alteryx work flows?

Jason Mack:
It's a source, for sure. We use Hadoop for a lot of things. A lot of our, say. web data is the obvious use case there. Chase's website is, probably, I think I've heard among the top 50 in the world for traffic. It's definitely the number one banking portal, so we get trillions of records every week into Hadoop environment and then I can use Alteryx.

I expose that data through Hive tables and then I use Alteryx in database tools, because I don't want to move all that data. It would take forever, it probably would never work. I can't copy that into my laptop. We expose that through Hive tables and use the in database blending to get the stuff down from trillions of records to the 10,000 records I really need to build the Tableaux dashboards that's gonna explain how well the particular ads were performing on the homepage or things like that.

Crowd question:
You didn't push downs to the cluster?

Jason Mack:
We're usually, Hadoop would be one of our sources and then we're publishing stuff up to Tableau server because my team, again, the most external thing we do is dashboards, right. That we give our leaders and all, so were not pushing analytics back to the cluster. We're pushing them up to Tableaux server from one of the sources.

Crowd question:
You mention that in database processing right but if you think about that in the Hadoop, well that's pushing down your in database processing to the cluster. You guys are doing that.

Jason Mack:
Yeah, exactly, yep.

Crowd question:
Okay.

Jason Mack:
Both Hive and Impala. We use Microview's Caldera. We have a 250 node cluster.

Crowd question:
You talked a lot about empowering your analysts, right? Have you proliferated at all to other people in the company or the common man or citizen analyst as well?

Jason Mack:
Analysts exist within every one those lines of business and even within like particular product teams and things like that. They will tend to find us, when they hit the limits of their current processes so I have encountered the past three years or so that we been using Alteryx, I haven't had to actively sell it to any group. Once it's in they hear about it through Tableaux, they see stuff we do with our internal communities and say, "Hey, can you show us what it can do?" Then they'll make the decision in their group if they want to wind up adopting it.

We do give it to a lot of those analysts. We sit at the corporate level in marketing, but a lot of the ones in the line of business have picked them up as well just depending on what their needs are.

Crowd question:
Do you guys have anyone dedicated to community or culture? Dedicated to -

Jason Mack:
Around Alteryx?

Crowd question:
Just in general, analytics, self-service analytics.

Jason Mack:
Around Alteryx, it probably may. I think we the analytics community, there's individuals in each group that care. Data geeks that love doing this stuff right and self-select into participating in a lot of this. It's not anyone's job to say, "Here, go promote a culture of analytics." It's happened naturally.

Crowd question:
Thanks.

Crowd question:
How many years have you guys been using Alteryx?

Jason Mack:
Three years. I think I bought the first license in very early 2015, late 2014. Something like that.

Crowd question:
You talked about the downsides to the waterfall approach. How it could take a while in the BRDs and testing are a lot of fun. Have your analysts, when your guys deploy, build a workflow and deploy it. Are you guys using that same approach? What are you guys doing?

Jason Mack:
We take more of, call it agile. We don't do scrums and things like that. Because a lot of what my group gets called into is to hard problems other groups can't solve or new data sources, things like that. We take approach of let's get our hands in the data, try to figure stuff out, fail fast sort of thing. If we say, "No, this is not going to give us what we need, what we're trying to solve."

I'd say it's more like an agile methodology, but not officially.

Crowd question:
Thanks.

Crowd question:
How big is your Alteryx cluster in terms of the server enterprise environment. Is it two node, three node, or?

Jason Mack:
We have multiple server environments. For my groups, mine is a relatively small cluster. It's just a two node, eight core setup. There are multiples to each department that has adopted it, getting their IT groups to support their server environments as well. We have several, but that one's mine.

Crowd question:
[inaudible]

Jason Mack:
That's something we're trying to figure out. The future, the challenges we're working through.

Crowd question:
That's what we're working through. I think when [inaudible 00:38:26] talks about it, right, say if you are doing something every week, every month and it's a mature workflow. It's not changing. Then you want to promote that into your regular SDLICT fold and, maybe, using [inaudible 00:38:38] how to do that. Is that [crosstalk 00:38:40][crosstalk 00:38:41]

Jason Mack:
We definitely do that. We use Alteryx as a rapid prototyping tool for that kind of stuff.

Crowd question:
Excellent, right.

Jason Mack:
Instead of just going to IT with a vague, "Hey, we need this to be done." and waiting for that process. We'll go in, we'll figure it out, we'll build an Alteryx flow and say, "Okay, this is gonna take the data that we're buying from a vendor or getting out of some other process into the format that we need where it's analytics ready."

We've been running it for a month or two now. We're confident in it. Here's a very solid set of requirements that you can pick up and control [inaudible 00:39:09] whatever the enterprise class three ETL tools are. Give that to them and let them do that process there. We definitely use Alteryx as a rapid prototyping tool to help IT. They're happy about that as well because we're not giving them half baked ideas that wind up failing and everyone's unhappy with each other.

We know exactly what we want. That's the fundamental challenge with analytics and IT. IT wants, "Tell me exactly what you need and I'll go build it." In analytics, it's this whole journey and adventure. I don't know where I'm going to go or end up when I get started. Because as I get in and explore the data I might take a different track.

We're out of time or good?

Speaker:
Still a few more minutes. Any other questions?

Thank you, Jason Mack.

Jason Mack:
Sure, thank you.

^Top

ERLEBEN SIE
DIE WELT VON ALTERYX
SELBST.

Los Gehts