CONTRIBUTOR
General Manager and Editorial Director,
Techstrong Group

Synopsis

In this Digital CxO Leadership Insights series video, Mike Vizard speaks with Jay Henderson, VP of product management for Alteryx, who explains how acquiring Trifacta will advance analytics by simplifying data management.

 

Transcript

Mike Vizard: Hello. Welcome to the latest edition of Digital CxO Videocast. I’m your host, Mike Vizard. Today we’re with Jay Henderson, who is the vice president of product for Alteryx. They are in the process of acquiring Trifacta, and they’re going to build out some interesting stuff in the cloud; machine learning algorithms and analytics all coming together. Right, Jay? So walk us through that a little bit.

Jay Henderson: We’re really excited that we’ve announced our intent to acquire Trifacta. We saw some really big trends happening in the industry, and I think bringing together the technology from Trifacta and Alteryx really helps us address those. First, we saw the center of gravity for data really being the cloud, and as data moves to the cloud we need to bring analytics and insights to the cloud. So that was a big factor, the cloud.

Second, we really felt passionate about trying to help businesses unlock insights from all of the data that they have stored. I think bringing together the two solutions is going to really help with that.

Third, it lets us expand into a new audience. Trifacta really serves the needs of data engineers and IT, and that goes together well with Alteryx serving the needs of the business and analysts.

Mike Vizard: So one of the things that leaps out at me about these folks is they are making extensive use of machine learning algorithms and we hear a lot about AI these days. How do you think AI will change the way people think about analytics? I’m asking this question because not everybody always trusts the data that they’re being presented with to make a business decision.

Jay Henderson: For us, we spend a lot of time thinking about how we democratize the analytics, and how we take some of these really sophisticated algorithms and put them in the hands of a much more casual end user. So as we think about how we’re going to get mass adoption of machine learning and AI, really the answer is to make the AI more transparent, so that you can really understand what’s going on there and make it much easier to use so that the data scientist can focus on some really high-value business solutions. Then the more casual business owner and analyst can do some things on their own.

So we’ve been focusing a lot on making the very sophisticated analytics accessible to the more casual end user. That’s really apparent in the solution we’ve got around auto insights, which will highlight data and highlight insights in your data automatically.

Mike Vizard: When you think about that, is the future of analytics going to be driven by analysts who are using an application, or is it going to be much more embedded within all our applications, so that every end user kind of has access to the latest and greatest analysis in real time, and is making decisions based on the analytics data that has surfaced within the context of some application they’re already using?

Jay Henderson: I think you’re going to see both techniques used going forward. I think certainly as you want to provide access to a business user, doing that in line with the other things that they have and the other tools that they’re using is really important. But I think there will continue to be a role for specialized analysts and to have purpose-built tools for them. As well, I think there will be a role to have tools for data scientists to use. So I think it’s really about making sure that the tools themselves have been built for the person with the right kinds of skills, and that are appropriate for the skill level that that used has.

Mike Vizard: Do you think we’ll get to a point where people will trust the data more? I’m asking this question because a lot of times when I talk to digital CXOs they’re a little skeptical of the whole analytics thing, because they know where the data came from and they’re aware of all the governance issues and the things that can go wrong with that data. So then they become a little reluctant to make a decision based on data that conflicts with their gut. Sometimes they only want the data that confirms their gut. So how much are we going to trust the data and how much can they be a, quote/unquote, fact-based driven business?

Jay Henderson: For me, I think the way that you start to build trust is you can add capabilities into the way the tools and the solutions work to help establish that trust, to provide the ability for the analyst or the business user to investigate the insights being shown, to really look for the underlying causes. So I think there are things that we can do in the technology as we’re building it, to help build that trust, to help them really understand what’s driving an insight that might be being surfaced.

So really, there’s going to be a balance between helping guide an analyst toward an insight, and helping them understand the drivers behind that in order to establish the trust. So really getting better observability into the things driving the particular insight. I think that’s going to be really a great way for the tools to help build trust in some of the capabilities they provide.

Mike Vizard: How well will we close the proverbial loop here on these decisions? A lot of times folks want it to be highly automated, to the point where the analytics is driving the business outcome automatically. Or do we still want to kind of look at that data and make a decision, and then create some automated process that goes with it? I guess I’m asking how prescriptive might it all get?

Jay Henderson: I think some of that will vary, depending on the business problem being solved. I think some business problems do lend themselves to highly automated decisioning, things like personalization, where real-time responses become very important to the way that personalization works. I think those sorts of business problems very much lend themselves to automated decisioning.

I think there are other decisions where you’re going to see much more humans working with the algorithms, man and machine working together, where the machines can help guide the individual to the right decision, but still those decisions are going to require human interaction and actual human thought to make the final decision. But the machines and the algorithms can guide the analyst or the business unit towards that, give them context to make a better decision, maybe even recommendations or options for decisions. So I think we’ll continue to see kind of a range of the way that works.

Now I will say that one of the things we see with our customers is that when they do automate the analytics that they’re doing, when they start to schedule the data being gathered or the insights being run, we do see a lot more value being driven to an organization than when the analytics are just ad hoc. So I do think operationalizing the analytics is going to be important for organizations to drive value out of their data.

Mike Vizard: As we go along here, do you think that the analytics itself will also become more federated across multiple organizations? We have a lot of conversations these days about supply chains, but I wonder. Historically when I look at it, a lot of the analytics was, “Here’s your report. You go figure it out.” But is this becoming more of a team sport?

Jay Henderson: Yeah. Look, I think as organizations have so much data that federation is actually the most likely outcome. I think there is so much data and so much need for analytics that you’re not going to have it centralized all in one spot, and you’re probably also not going to push it out into the business and have it _____.

So I think the more that we can get teams and groups collaborating within an organization, whether that’s the business user, the data analyst, the data engineer, that’s really going to be the future of analysts, helping facilitate collaboration across those groups, having some decisions made centralized and others pushed out into the endpoints.

It’s part of why we’re so excited about this acquisition of Trifacta. The combination of Alteryx and Trifacta really will be one of the only analytics platforms that can service the needs of IT and data engineers, analysts and business users.

Mike Vizard: One of the things about the platform that’s kind of interesting is that it does address data management. I think it’s one of the dirty little secrets of IT that we’re not so good at data management, and a lot of that has to do with the IT department didn’t create the data in the first place. Some end user created it and then the IT people store all that data, but they don’t really have a lot of insight as to what data is more is more valuable than the next. Are we going to get better at data management? If so, how?

Jay Henderson: Look, I think there is a whole new generation of tools that are part of the modern data stack, like Alteryx and Trifacta, that are helping deal with some of that, and frankly, I think helping deal with some of the messiness of data management. One of the things that both tools are really great at are combining different data sources together.

I think for as much as IT tries to centralize data storage or consolidate different data sources into a single repository, one of the realities that we’ve seen from business users and analysts is that they always have some extra data source that they want to have combined or that they want to blend together with the data that comes out of those larger repositories.

So both Trifacta and Alteryx are really good at blending data together from multiple sources. I think some of our research shows that a typical set of analysis involves data from over a dozen different data sources. So some of that will be a balance of sort of just getting better at bringing data together for analysis, and balancing that with some of the centralized repositories that might be being created by the IT team.

Mike Vizard: Do we have enough data literacy these days? I’m asking the question now because if we lean more on these AI models, the AI model was trained by a certain set of data, but if the data changes we need to update the AI model. And it’s not clear to me that everybody understands the relationship between the recommendation or the outcome driven by the AI model and the quality of the data.

Jay Henderson: Look, I think we’re in a really interesting point in the evolution of analytics. We’ve got some very sophisticated technology, some very sophisticated algorithms. I think we need organization skills and their process to catch up with the technology.

I think there are some things that we can do as technology providers to help make that easier, to help coach people through some of the data literacy, and to help lessen the burden and the need for skills and process. Ultimately those are the three things that come together to make successful insights. It’s people, process, and technology.

Mike Vizard: Speaking of that, do people really understand the relationship between data scientists, data engineers, and your typical analyst? We need to think about how those teams are structured as we head into 2022.

Jay Henderson: I think every organization needs to be thinking about what skills they have, what skills they need, and really understanding across those different roles who in their organization is going to be playing those roles. I think in smaller organizations you’re going to see people with a mix of skills and a mix of roles, with maybe one person being both the data scientist and the data engineer, or maybe the data scientist and the analyst.

I think what will fit and work for your org is probably going to be a little bit different than maybe what you see everywhere else. So to me, being thoughtful about the roles and the skills that you have in your organization is just as important as picking the right kinds of technologies to help enable them.

Mike Vizard: So what’s your best advice then to all those digital CXOs out there that are trying to make sense of all this stuff? What should they be kind of talking to their teams about? What is the best point of departure from here?

Jay Henderson: For me, it comes back to the people, process, and technology. I think for the people, you really want to be looking at what skills does your organization have. What skills does your organization need? And what are you going to do over the next year to progress that organization to be more where it needs to be?

On the process side, you want to be looking for: hey, what sort of process can I surround those people in the technology with, to make the analytics that we’re doing more repeatable? How do we create really repeatable sets of analytics?

Then on the technology side, you really want to be thinking about creating a modern analytics stack, looking for technologies to help you with extract and load, so building your data warehouse, looking for technologies to help you with transformation and analytics like those provided by Alteryx. Then really thinking about how those tools and technologies are going to be driving insights that are actionable within the business. So for me, it’s really about balancing the people, process, and technology.

Mike Vizard: All right, great. Hey, folks, if you’re still trying to manage your business off the back of a spreadsheet, I think you may be extremely disadvantaged going into 2022.

Jay, thanks for being on the show.

Jay Henderson: Thanks a lot.

Mike Vizard: All right. Take care.

Show Notes