We’ve extended Inspire Early Bird Pricing until March 1. Register now and enjoy 20% off conference passes and 10% off training passes. P.S. Don’t forget to bring friends! When you sign up for five or more tickets, you get an extra 20% discount on conference passes. Learn more now.

Alter Everything

A podcast about data science and analytics culture.
AlteryxMatt
Moderator
Moderator

Supercharge your data storytelling capabilities and automate your reporting with Alteryx Auto Insights! Katherine Roach, leader of the Global Cloud Acceleration Team at Alteryx, takes us behind the scenes of Auto Insights, sharing highlights, use cases, and most-loved features. 

 


Panelists

 

 


Topics

 

 

AlteryxMatt_0-1674589835485.png

 


Transcript

 

Episode Transcription

MEGAN: 00:00

So who here has gotten stuck creating endless variations of data reports? Or what I like to call acting as a report jockey? Or maybe you've had to spin up a simple explanation for your complex analysis so that your boss's boss's boss can know exactly what the report says? I'm guessing a lot of you can relate. I know I certainly can. My name is Megan Dibble and currently I'm the data journalist at Alteryx, but before coming here, I was an analyst using Alteryx at a company where I faced situations like these. By the time I made my job transition, Alteryx had just acquired a new product, which is now called Alteryx Auto Insights. It accelerates the process of going from data to a report to insights, and it automatically showcases the why behind the findings of your analysis. So for this episode of Alter Everything, I sat down with Katherine Roach.

KATHERINE: 00:51

Hi guys, so I'm Katherine. I currently lead the global cloud acceleration team within customer success. That's a lot of buzzwords in one sentence. So what it really means is all of our new market cloud products, such as Auto Insights, Machine Learning, Designer Cloud, how can we make sure customers are successful with those, and they have the right tools and resources available to support them?

MEGAN: 01:14

Awesome, well, it's great to have you on here, and where are you based out of, Katherine?

KATHERINE: 01:19

I'm baked out of Australia, so you can get a whole 30 minutes of the Aussie accent on your podcast.

MEGAN: 01:25

Score. In this episode, you'll hear how Auto Insights can save you not only time, but can help you retire your report jockey status in favor of more exciting data analysis that will help drive impact for your organization. Let's get started. So you mentioned the product Auto Insights, and I would love to hear about how you got involved with that product and what that product is. If you could tell us more, that'd be awesome.

KATHERINE: 01:56

Yeah, absolutely. So my journey to Auto Insights actually started back when I was 16, funnily enough, when I decided I wanted to become an actuary, any young girl dreaming of her older life.

MEGAN: 02:07

At 16?

KATHERINE: 02:08

Yeah, at 16. What a dream. You can tell I'm so much fun. So I studied and worked for a number of years at EY in non-traditional actuarial work, and that helped clients navigate operational and strategic challenges using data and analytics, whether that's predicted modeling, machine learning, et cetera, and that really gave me that understanding of what it's like in the trenches and how hard data can be, but how powerful it could also be to organization. And although I really loved my time there, working on heaps of different client problems, I really wanted to do something that made, I guess, what I felt was more of an impact. So like any millennial having a quarter life crisis, I joined a startup, and that was the beginning of my journey. So I joined Hyper Anna, which is now known as Alteryx Auto Insights. Back then, it was a Sydney based startup. In 2018, I joined just after their series A funding. And when I joined, I was actually like the client engagement manager, and what that means was I essentially was there to deliver a proof of concepts to new customers, showing them how the product works, and how it adds value to an organization. And within a few months of joining the team, it was major innovation, which for me meant defining a new way of working with clients, forging a new part to success, which we'd, I guess, never seen before. And so when I was went along, started to head that function, we can fast forward to November 2021, and that's when Hyper Anna was acquired by Alteryx. And I can get into that later in terms of you know how the two products sit together. It was a no brainer. It's not often that I get startups get funding, it's often that they make it to an exit event, their loan get acquired by a company where the product, the values, and the culture so aligned. So that's really how I started my journey at Alteryx Auto Insights.

MEGAN: 03:56

That's awesome. And that is pretty rare. So that must mean that Auto Insights is pretty special, right?

KATHERINE: 04:02

Very special.

MEGAN: 04:04

What makes it so special? Can you explain just what the product is and what it looks like?

KATHERINE: 04:09

Yeah, absolutely. So if we think about organizations today, they have heaps of data, and we all know that data is probably one of the most important things that companies have. But there's a lot of problems that encompass data and to the business a few key challenges in unlocking the power of data, right? So one of them is getting messy data into a format that you can perform analysis on, ETL Prep Blend create that perfect data asset. That's what Alteryx Designer is all about, we love Alteryx Designter. But once you have that perfect data asset, the challenge really becomes in reporting, finding insights in that data and in the personalization of those insights to each and every person that needs them. And so what the business really needs is advice, and not more dashboards. And this is really the problem space that Auto Insights plays in. You can think of it as your personal AI powered data analyst. So once you have that beautiful clean structured data asset from your design and workflow, or database, then Auto Insights uses its smarts to consume and process that source data. So there's no need for an analyst to do anything other than point Auto Insights at the right data and let it get to work. It will go through and it'll start to find insight automatically. It will present and explain them to business people in simple, clear terms. Instantly, people are able to understand what's happening, what's important, what's changed, what's going wrong, what's caused whatever we're seeing, and much, much more. So they're spoon fed these explanations of what it all means, and aren't just presented with numbers on a page.

KATHERINE: 05:47

So it's really answering important questions instead of just giving a number. And of course, it also makes it easy for business people to be able to probe further, right? Into what they're seeing. So users can effectively ask Auto Insights questions and get answers back quickly and automatically. And it's the same kind of thing that users would normally get by talking to an analyst, that back and forth. But now, it happens without anyone really needing to be involved. So as you can imagine, it's much more than just a dashboard. It has machine learning behind the scenes, natural language processing, to make it really, truly feel like that personal data analyst.

MEGAN: 06:26

I've had people ask me, I already have a reporting solution, or we have dashboards, why would we need Auto Insights? But that I was a really great summary of all the extra things it does, and having machine learning behind the scenes working to answer stuff that you didn't even know to ask is really cool. That's my favorite part about it.

KATHERINE: 06:46

And you make a good point, right? It's I already have a dashboard, right? Why do I need Auto Insights? And then that's a question that we get a lot. And to be honest, all of our customers have existing dashboards. Dashboards are fantastic at giving you the what, all beautifully displayed on one page, and sometimes that's all you need, but there's a lot of the time where the business now has the number. Great. But now I have a bunch of follow up questions. And if I want to find answers to them, I get lost in pivot tables or rabbit holes or multiple drill down. So as an analyst in my past roles, I can tell you when someone comes to me and says, but why is this number this number? I think into my seat and think, oh dear, please don't make me answer this question because it takes a really long time, because the dryness of why, they will change every time their data updates, day to day, week to week, month to month. Why will always change. So there's no easy way that you can just put it into a predefined dashboard. And that's really where Auto Insights can help. It's getting you that personalized why, and bringing together the what of a dashboard and the why of Auto Insights to make kind of that empowered, data driven decision.

MEGAN: 07:56

Yeah, I mentioned in my quick introduction of myself earlier that I was a data analyst before this role. And so here you talk about the product and all the benefits and everything, it makes me think, gosh, wish I had that back then, because sometimes depending on the business situation, there are urgent needs and people come to you and they're like, we need the today or tomorrow. And it's like, wait, no, you don't understand how much work it takes to make a dashboard, or do we have the data in good shape or whatnot. So like being able to automate the more tedious parts of it, and get something in front of business users that has not only valuable information and summary-- what they're looking for, but even more than that, which I would not have been able to think about and just a couple of hours time, but the algorithms were like, no, you might want to know this. Here's the why. Here's this. That would have been great. For sure. And yeah, I think another part of it that I enjoyed using and getting to know, and I've been teaching myself as I write these blog articles is just the logical way it's all built out so that you can type in your question, and then it'll generate responses, and it's very, just, visual and intuitive. So that's also nice to really get to that realization of, oh, we have this data point that's changing things. So if someone comes back to me and they come back and they're like, why is there this spike here? Here's the answer. It's already here.

KATHERINE: 09:28

Here's why. yeah.

MEGAN: 09:30

Yeah, exactly. I don't have to panic or anything like that.

KATHERINE: 09:34

I know, I distinctly remember the sinking feeling of when a request would come in, and it was thinking, you have no idea how long this is going to take me. So yeah, I'd like just imagine back to uni or my first days in like, EY, how much time I could have really saved knowing that a tool like this was out there. And it's one of the reasons I joined a startup, right? It's like, you're leaving this extremely safe job within a big four as an actuary to drive some startup in the data and AI space, and most startups fail, right? I truly could understand and believe the problem that they were solving for. And yes, the execution early on wasn't exactly like solving for miracles, but now that I see the product today, I wish I could have taken it back into my old life. It's just phenomenal.

MEGAN: 10:21

Yeah, totally. Could have saved some late nights, or sometimes as an analyst, your job kind of like devolves into basically like a reporting bitch, like people just come to you with request after request. It can get really repetitive, and not as fun as actually delivering value and actually getting into the nitty gritty of the data and why stuff happened, and having those discoveries. So I'm wondering what Auto Insights can do for a business. What value can I add once [people?] start using it?

KATHERINE: 10:56

That's a great question. And again, one that we get quite a lot, and it's really like, how can you think about the different stages of going from data to insights? So going from the data that you have to getting an insight or a decision into the hands of the people that need it. So you can think about it as that process. So the creation of data or the creation of information to go to the business, well, that creation process, and then putting the right information in front of the business. Are you putting truly insightful information in front of the business? And then finally, making sure that the business can actually interpret and make decisions off that data. So that's the pathway from data to insight. And auto insights can add value to each one of those stages of going from data to insights. So in terms of the creation of information, it's typically a very manual and time consuming experience to the BI and analytics teams. And the benefit of Auto Insights really comes into the automation aspect. So really to the analysts, you don't need to start from scratch. Auto Insights ingests and make sense of your data, so there's no need to do the drag and drop, spend time slicing and dicing your data, creating a dashboard after dashboard. You can kind of let Auto Insights do the work of those mundane tasks for you, so the analyst can really spend time on that higher value work. So it helps to automate the creation of the data information that goes to the business.

KATHERINE: 12:21

And then it's about putting the right information in front of the business, right? And this is really talking about the difference between either reactive insights and proactive insights. And what Auto Insights does is it provides deeper, smarter, more powerful insights to the business, instead of just a number on a page. So Auto Implant provides what's happening? My revenue's increased. Why is it happening? It's a particular department or sales promotion or demographic that's really driving that change. And then it will tell you something that's maybe unexpected. It provides these types of data stories that will allow you to essentially discover more about your data than what's just a number on a page. And you can really personalize it, right? It's about like, how can we get personal information into the hands of the user? So now you've gone from showing high level number on a page with a chart and maybe some commentary, if you're lucky, to showing a number that you care about the drivers of that number, and a host of supporting data stories with the ability to drill down and slice in any way you want. And now that's really information that you can make a decision off. So now that you've got the data, you've got the insights that you want to provide to the business, it's about making sure that they can read and interpret that information. So we know that millions of dollars is spent by organizations every year in terms of uplisting their data capabilities from investing in systems to capture the right data, to talk about Alteryx, to investing in BI and analytics teams. But one of the key challenges they face is that lack of data literacy in the workforce, and which can be frustrating to have just spent millions of dollars in trying to procure all this data, but the people that need it the most, the front line of business, they're not able to actually make sense of it and articulate the decision that they're making from it. That can be frustrating. So what Auto Insight does is it makes those kind of powerful insights really simple and easy to understand with things like natural language surrounding each insight. So the businesses can uplist their data literacy, and have a real impact using data to make a decision.

MEGAN: 14:27

Yeah, it's like when I've been trying out the platform, it is cool to see the summaries at the bottom where it says in natural language, plain English. Here's what happened. Here are some reasons why it happened. We had an increase here, and it's spelled out for you instead of just looking at some values on a page, or some table of numbers where, it'd take kind of a lot of brainpower to understand that sometimes, so you can get it across faster, for sure.

KATHERINE: 14:54

Absolutely. And that's one of the things that I love is like, the data stories that support each insight. So maybe for the listeners, they're not as familiar with Auto Insights, so you have our main feature, which is called missions. And you can think of missions as almost like a PowerPoint. So it has, within a mission with different pages, and within a page there's different stories. And on each page you can create your own query, and those stories will then supplement that query. And what Auto Insight does is those stories that are generated, instead of seeing like that squiggly line on the page that has like a hundred different lines on the page, that's super hard to read and interpret. What it will do is essentially look at that information, go away and ask loads of different questions of the data, and provide back the data stories that actually tell you what's happening in that squiggly line chart that you don't need to spend that effort to to read and interpret.

MEGAN: 15:46

Yeah, cool. So you mentioned, you know, about how Auto Insights automate a lot of parts of the process, or speeds it up, or whatnot, and as the former data analyst myself, sometimes when people talk about, oh, here's this product. It's like what you had before, but it's automated, there's concerns around, what about my job? Will this get rid of my job? Am I going to be automated out? So curious if you've heard that concern and what your response is to anyone who's a little afraid of the machines learning.

KATHERINE: 16:18

For me, my simple answer is no, right? I think everyone knows how important data is, and the impact that data can have an organization. And the demand for insights will always outweigh the supply. And so I don't think you should see it as this is going to take over your job. It's really about how can we supplement you? How can we save time for analysts being report jockeys, only outputting variations on the same report over and over again, automate those kind of mundane tasks to really focus on the higher value work. And the work that they're probably more interested in anyway, right? That's the fun, juicy, sexy work that analysts want to do, that they don't get time to do today because they're bogged down in the constant reactive Q&A with the business. So I think you should definitely see it as, especially if you have an overworked or overburdened data team, Auto Insights can definitely supplement that, and hopefully elevate to be able to give your team more capacity to do the exciting stuff.

MEGAN: 17:19

Yeah, it isn't fun to just make reports all day, or be like a report jockey. So we talked earlier about existing reporting solutions, and then Auto Insights, how it's different. But could you shed a little light on how it would fit together and kind of the reporting ecosystem, or the business tool ecosystem?

KATHERINE: 17:40

Yeah. So I think it's kind of what we mentioned before is like, the traditional dashboards. They're really good at giving you the what, and Auto Insights is really good at giving you the why. And I think the combination of the two that really allows you to get those powerful decisions.

MEGAN: 17:59

So I'd love to hear about any cool business use cases that you've seen, how you've seen Auto Insights in action.

KATHERINE: 18:07

Yeah, so I'd say one thing about Auto Insights is it's very industry and department agnostic, so it's not necessarily about what industry or what department you're in. It's really about, do you have the right problem to solve with AutoInsights? So some examples of use cases could be around sales performance, understanding where sales has increased or dropped, where a team's hitting their goals, is a good or bad performance, who requires additional coaching, where are the new opportunities for revenue growth, for example? You can think about it from a sales perspective. You can think about it from a finance or procurement perspective as well, the idea of the identification of cost savings, where teams should optimize pricing, which supplier vendor has more contracts across departments. Where's the room to further, I guess, like contract negotiation in those? So you can think about it from a finance lens. You can think about it from an operationalized lens where it's either your call center information about analyzing calls and queries coming from your customers, which products are impacted. Where are the upskilling opportunities of those teams? So you can kind of see that it can be applied to any industry, but I think fundamentally what these things have in common, it's best used in that sweet spot of your line of business. Your line of business managers or team leads who are really on the hook for performance whether that's in finance sales, ops, HR, data, fraud, these are all areas that we work in. But really, that person that's using the tool, their job is to make the team's performance better. And they are typically very underserved with insights. They only get a static dashboard or numbers, but no direct analysis or support for them to make a decision, even though they are the ones that need it the most.

KATHERINE: 19:54

So for example, we have a customer, right? And they have a global sales team, so you can pick sales managers across all geographies worldwide. And their role day to day is managing sales team and supporting business initiatives to drive revenue, both up and down. So they want to be able to tell their executives what's happening in their revenue, they want to drive the best performance to their consultants on the ground that actually drive in the business. And so the challenges that they've face is that every Monday they have to do an 8 hour report, which consumes their entirety of their day, they have a very detailed dashboard that is the single source of truth for everything, but despite having that information available, it's really challenging for them to navigate that dashboard to find a compelling story in the data. And that's because of the different drivers of performance each week, right? So this week it could be a particular department or client or industry driving that growth. And next week, it could be an entirely different driver of change, right? So in their dashboard today, it's like, how do you drag and drop and slice and dice for every single permutation of what's driving your chain? And so with that limited time to be able to understand the performance, they can only really do simple top down analysis. And that results in superficial insights. They've never really been able to look across all of their drivers across all of their 20 sales segments across the 30 different sales KPIs that they have because it's just impossible today.

MEGAN: 21:23

You don't want to have a dashboard with 20 tabs, and every tab has 20 filters, that'd be too much.

KATHERINE: 21:30

Exactly. And even if you had that, it would change, right?

MEGAN: 21:32

Where do you look? Oh, yeah, yes.

KATHERINE: 21:34

Because this time it would be filtered on this department, because that's what you think drives a change this month. But next month, it's entirely different, so it's a different filter. So it takes an entire day for these people to be able to dig out this information from their dashboard. So how they started using auto insights is really about, how can they share and find data stories within that reporting. And so they essentially set up automated insights in their area of focus that's driving their sales KPIs. So they had missions set up for each and every one of their sales managers so that each sales manager could have their own bespoke view of their business. And again, that was super quick to spin up because it's just simply changing a filter, right? All the insights that come with it are automated. So you can think of that. It was able to detect trends in all of their 30 sales KPIs, why they've gone up, why they've gone down. For example, it would call out a particular department, or customer, or industry, or product, or a combination of these. Ultimately telling them which areas contributed to that increase in revenue. So Auto Insights can share the areas that essentially didn't perform well, but also offers what is a strong performance. And then highlight anything unexpected like outliers and anomalies for them to assess. So this sales team ultimately was now able to not only look over the whole scope of their data, but reduce that 8 hours into 30 minutes looking at Auto Insights. And these are people that are driving the business. They shouldn't be wrangling data.

MEGAN: 23:09

They don't have 8 hours to spend on that.

KATHERINE: 23:11

They don't have 8 hours. They're trying to make money. These are the types of benefits that Auto Insights can really bring to organizations.

MEGAN: 23:19

That's an awesome amount of time savings. And did you see any other benefits come from having those insights?

KATHERINE: 23:28

Yeah, definitely. Because these insights are in the hands of the sales managers, what they had in there is essentially their projected revenue as well. They're not just the revenue that's realized but they had a field for projected revenue. And so what they could start to see is through the outlier detection, there was particular geographies that were actually spiking in terms of potential revenue. So what they could do is make sure that they had their account teams really targeting those areas. I think, for example, it was like Germany, and Germany saw a huge spike. And so what they did is they really highly focused on Germany so they could really realize that revenue. And that ended up making that person one of the top sales performers for the quarter. And he said, one of my top three reasons why I've been doing so well might sales is because of auto insights, which is an amazing thing for them to say. So that's the persona. I think in terms of the type of use case, we always talk about reoccurring impact. So meaning that you are required to do this on a constant basis, right? Like you're always required to monitor it. Maybe it's in a daily stand up, fortnightly, monthly reporting, whatever it is, it's in some type of regular cadence, and you're really trying to focus on understanding why something has happened. And then finally, I think the most important thing and thank goodness we have Designer to solve all of our problems, is you need to make sure that you have the right [play?] instruction and time series based data to make Auto Insights work. So it's really powerfully driving up the drivers of the trend so that data needs to be transactional in nature. It needs to refresh regularly. And we, naturally, Alteryx makes sure that your data is in the right shape so that Auto Insights can perform its magic.

MEGAN: 25:14

Gotcha. So do you see users then, they would use Designer to prep the data and then send it to auto insights for the reporting?

KATHERINE: 25:24

Yeah, absolutely. And that's where I'd see the, when the acquisition happened, that's why I was so supportive because I used designer in my prior life and I think that the two tools just blend together so well. Now obviously with built in native connections so that you can prep your data in designer and then push it straight into Auto Insights, but it really extends the kind of ecosystem of Alteryx from just the analyst community and then being able to prep and blend their data and make a perfect data asset to actually allowing their stakeholders now to self service insights about the data that they've just created. So it really allows you to have a much broader reach across the organization with Alteryx.

MEGAN: 26:06

Yeah. And then my last question, I was curious if there were any hidden features or lesser known parts about Auto Insights you wanted to point out, or some kind of cool story in the way you've seen it used or anything like that.

KATHERINE: 26:20

Yeah, that's a really good question. I'd say that it's a very rapidly changing product. And so we're constantly releasing new features. So I'd say things like magic missions, where before, you would land on a mission and create it from scratch. Now auto insights takes its best scans of what you want to look at. So it creates this magic mission across all of your data to show you the key things of what happened, why it's happened, and if there's anything kind of unexpected. So I think Magic Missions is like a really cool new feature that's just been launched. Yeah, not too sure about the hidden one. Maybe I know the product too well. So I think no, I didn't.

MEGAN: 26:55

No, I mean, I've used it. I got to say when I started using it, it was very easy to learn. It was, you know, all laid out and nothing is really hidden, and it's very clean design and the cloud and everything. So it's a little different than Designer where there's so many tools in Designer, where there's tiny menus and all these little hacks you can do, but I guess auto insights is, the whole point is making things super simple. But yeah, I did see the Magic Missions the other day. And I was like, oh, what's that? So exciting new development.

KATHERINE: 27:27

Absolutely. I think that I like to think of Alteryx as that Swiss Army knife where you can do anything with Alteryx Designer and it's designed for an analyst. Whereas for Auto Insights, it's really designed for that business user. The people that have a lower level of data literacy so they can read and interpret and understand what's going on. Yes, it makes the analyst's life easier, but it's just different audience, different designs. So nothing too hidden.

MEGAN: 27:53

Well, at the end, I did want to mention some resources that we have on Community for anyone who's curious about Auto Insight that wants to learn more. We do have an interactive lesson on Community about Auto Insights that will walk you through the product. We've also done, sort of, blogs about it and that can help you dive in a little bit deeper on things we talked about, like Missions. I think we have one that backs it up further and just talks about, why would I use this? Or what are some use cases to get people thinking about what are the possibilities? Because they're are so many. And it's been awesome to have you on the show.

KATHERINE: 28:35

Awesome. Thank you so much, Megan. I really appreciate it. Thanks for having me.

MEGAN: 28:40

Thanks for listening. To try out Auto Insights and find the resources we mentioned in this episode, check out our show notes at community.Alteryx.com/slash podcast. See you next time. [music]

 


This episode was produced by Maddie Johannsen (@MaddieJ), Mike Cusic (@mikecusic), and Matt Rotundo (@AlteryxMatt). Special thanks to @andyuttley for the theme music track, and @mikecusic for our album artwork.