Hi and welcome to the Data Strategy Gurus podcast. Once again, we're at the Click World Conference in Las Vegas, and we're joined by Kevin Ryan, who's the Director of Business Intelligence at Walter Kluger.
Kevin, can you explain what a Director of Business Intelligence does?
Sure. My job is sort of two main priorities. One is to take advantage of, in an ad hoc fashion, any question that might come to my team related to sales, marketing, product development. Basically, I'm like an order taker. If someone has a question about turnover or a question about finances, they come to my team, and we process those requests. That's about half of the work that we do.
The other half is much more strategic in nature, and that involves me and my team helping to design and implement a data strategy that can survive and be pragmatic across all the divisions within Walter Kluger. So one part is very tactical, and the other is more strategic in nature.
Yeah, I'm interested in the technical part where you say you get these ad hoc questions, and sometimes you get the same question five times again. You're trying to learn the self-service because there's intelligence. How do you take these types of questions and keep up with the speed of the questions that are coming in?
Yeah, I mean, in a practical sense, you can't necessarily engineer an automated solution for everything. That's a reality. But I will say that when we do start clustering requests that come into my team that seem to be fairly repetitive, that's when we start automating. That's when we start really trying to push the analyst out from my team and into our stakeholder organizations.
That really is a philosophy that we're trying to implement around the citizen analyst. If BI is centralized, it's my opinion that it's difficult scaling out to really accommodate the strategic needs of a whole organization. So whether that's technology, process, or training of individuals, it's something that we've tried very hard on my team to start implementing, and it's been quite successful.
Yeah, so you start in a centralized way capturing the questions, and when you're looking into scaling that, you take that equipment with the business analyst and put that, correct? Because you had to be ICCs a while ago. People have been trying to centralize, decentralize all these motions, and nothing seemed to work in a certain way for some organizations. The BICC did work because you have centralized the knowledge on how you set everything up on a technical level, as well as on the company industry level.
Absolutely. And I think that centralization versus non-centralization gets to one of the real struggles that I do have around culture. Right now, the culture of my organization is that analytics questions can be centralized, and there's a lot of resistance that I find in my job. It's around creating that culture where folks are more curious about accessing data, answering their own questions, providing insights, and then the inherent learning that happens when you self-discover a problem in the data.
When you find something in the data that is apparently yours, the momentum that brings to any individual or organization is really impressive.
Yeah, an interesting point—the data culture these days. What are you trying to set up to create that real data culture where people don't look at it as technology and get afraid of doing something with the data, but really get the value out of that?
Yeah, um, you know, I don't remember the guy's name. He's a social influencer, but he draws...
This he has this anecdote about dissemination of information. What I found is that putting things out there, letting people volunteer to be, you know, "I want to learn more about data and analytics," or "I'm a classic marketer," or "I've been doing the same sales job for a very long time." Asking people to raise their hand who wants to learn a little bit about SQL, who wants to learn a little bit about how we onboard data sources, how do we then visualize data. Getting those early adopters, getting them bought into that culture, then they sort of—I don't know, it's something that I've just found to be starting to work well because eventually that momentum reaches a tipping point that I'm hoping. Yeah, it should be much easier to change a culture than from scratch. So we're close, we're getting there.
Yeah, you're trying to find your champions within the company, on the business side. It's always—for me, always a business IT—try to do that separation and then try to find these people and help them in getting those skills. Then yeah, at one point, the champions—from my experience, it usually involves a vice president or someone with some kind of stature in their name. And the reality is that those people might support it, but to really build or change a culture, you need to get people—analyst level, manager level employees—that can get in with data, answer their own questions, and that's how you really change that culture. Slowly but surely, it's been working out.
Yeah, and then you said as well the strategic level—what you're right, you're right. So that is less stakeholder-driven and more around the hard problem after next. So right now, what we're trying to do, for example, is—none of my stakeholders is asking for segmentation based on customer lifetime value, for example. Okay, yep. But it's an idea that we're working on such that over the course, when the request comes from a stakeholder, the hard work's already done. So we're sort of like trying to foresee. And it's just—a lot of my job is engaging with stakeholders and understanding. What I'm hearing now is that Revenue might not be the best metric to be evaluating a customer relationship. Are there ways to do differentiated treatment from a customer experience perspective, from a marketing perspective? So what that leads to in my mind is understanding—being able to put a metric on a specific customer and all their relationships, and have that be able to differentiate how you treat that customer relative to other customers. So it's not being asked for right now, but that's a lot of what we're trying to do: push the organization in areas where really advanced metrics and capabilities can answer hard questions that are coming down the pipe.
So you're educating them. Well, "educating" is a hard word—always be sure to help them understand what you can do with data. For me, it's kind of strange that a marketer would know anything about customer lifetime value, but what you're saying is exactly like that. So sometimes it's a bit strange that the people from the tech side know more about that and can help them in such a way. Yeah, I think what it really is about is understanding—if you've got a set of data and it contains usage data for a particular customer, a stakeholder might look at that and be able to say, "Oh, that's interesting. That tells me what just happened in the—"
Last quarter or the last month, someone with understanding of technology can look at that and say, "We can pull in other variables around retention, and now we can do things like predicting a customer's decision to retain or not retain based on a particular data signature in their usage or things like that."
And that's how the technology—the technologists' understanding of our data—kind of coordinates with our stakeholders to come up with some interesting analytical solutions. So yeah, they do understand what you can bring together, and that's not necessarily known by the people currently.
Correct. It's just understanding, really, on a time-based perspective: can you help me understand about the past, or can you help me predict things about the future? That's really what I'm trying to push towards. Data is obviously interesting being retrospective, but there's a ton of power in perspective and being able to be predictive.
And you were a B2B firm, so a lot of the traditional playbooks kind of lack that idea of churn, the idea of cross-selling and upselling. But it really is just the same math. If you have the data, you can answer the same types of problems. And then again, it's about getting a culture to adopt those recommendations and change for the better.
Yeah, you mean in B2B it's not the same when you talk about churn and everything, and in B2C it's very prominent to do cross-selling, upselling?
Yes, yeah. I mean, it's... you know, a subscription business, a monthly subscription business like Netflix or something like that, they might have something like four percent per month where they have to... they're constant. So it's a constant problem. Our businesses, we retain 90-plus percent per year, so there's not the pressing imperative that's sort of built into your business where you need to... but it's obviously... retaining customers is the cheapest way to do marketing, to grow your footprint and things like that.
So yeah, it's amenable to all the great tricks of the trade that exist in B2C, you know, BI shots and things like that.
Yeah, and your... well, business intelligence—it feels like, "Hi, this is from ages ago." Are you looking at artificial intelligence to apply that in your B2B space?
I mean, so look, I have a pretty high threshold for what I would consider AI to be. And I think for me, it's mostly around a self-learning component—right? So being able to put, say, a decision engine out into the market and have the decision engine surface recommendations for customers and then learn from when a recommendation was adopted or not.
So we are scratching the surface around that, I would say. But AI more so in... you know, finding, taking an entire corpus of data from, you know, our customer service agents and putting that through a natural language processor to really thin out the thematics that are existing, and then applying metadata to that and being able to analyze that—that's, I'd say, the extent of the AI that we're going for.
But you know, definitely trying to be prescriptive in anything that we do. So finding out a customer that we have... we've uh, we have models called Red Alert models, which... it's a data-driven approach to proactively letting a sales rep know that they might have an issue.
Is that AI? I don't know. It's an evaluation of data, and they're making a recommendation and learning from that. So yeah, it's statistics, and then the next level it's AI. It's a model, and then you have...
Yourself learning models like you say, it's a thin line. I agree with you where you would say very often people shout out it's AI, but it's kind of automated decision, right? Right, right. So, um, you know, I have utmost respect for folks that are actually building AI engines, and so I tend to be a little bit—yeah, yeah, I'm following what you're saying. So I just sometimes challenge as well, okay? AI, yeah, just chill me and whatever you're doing. Sure.
So in that respect, in business intelligence solutions and especially in AI where they think it's smart enough for itself, it comes down to the data. What are the type of data problems you're facing on a data quality level? Because AI is gonna be very helpful in supporting the data quality for toolings, yeah? Not where we're just discussing AI at the end of the loop, but look at it to help in the data management as well.
No, I think that there's—you know, and I haven't seen a lot of—so whenever you do, when you roll up your sleeves and you do kind of programmatically trying to cleanse data, right? There's inherently a threshold with which you're willing to say, um, I'm gonna crank up the interpretation of this algorithm. It's going to change my data, and it might make some change in the data that I'm not going to be able to find. But the balance of using an intelligent approach to a problem with data quality—so in that regard, I think that there's a tremendous amount of potential for AI in data quality, where it can sort of learn from a data set and make those modifications that need to be made.
On a small scale, we're fortunate enough that the data problems that we have are around, frankly, classical organization of silos—around how do you, you know, how do you define retention? One business to another one as a customer in that window of being a retention risk. And those are easier problems to have, but you can't throw a technology at them.
Yeah, no, it's really the semantical meaning of your KPI off of what churn means. Yeah. And do you use a kind of framework for that or anything that helps in getting that right?
So we take two approaches. One—and really anything related to changing the data strategy of our organization—we tend to go top-down and set up just committees of, you know, how are we going to define a customer? How are we going to—who's going to be responsible for making changes? But then the other part that we do is, again, that sort of grassroots, decentralized approach. We're creating a lot of pilots, specifically around data quality and data stewardship, where you know those decisions where you can't have an elegant AI-driven solution to data quality, you can outsource it to your business analysts and have them—and bring them into the fold of data quality, you know, and making it more of a team sport.
We've done a lot of work in that regard with our partners at Thailand around using their kind of non-code-based, drag-and-drop-based solutions to quality to be able to do, you know, outsourcing some of the stuff to stakeholders. Because that, to me, is a concern in scaling my organization or the data strategy across the whole division. It's going to need the amount of people that you have to do it is going to stay lean, and it's going to have to be very efficient. And it's going to—
Here’s the formatted transcript with punctuation, capitalization, and paragraph breaks:
---
Have to rely on stakeholders to participate because, you know, doubling the data responsibilities of the acting—you know, it's difficult to then say, "Well, I need you to double the size of my team as well." So we're sort of preparing for that.
Yeah, just to fix the data quality issues here.
Oh, for sure. Yeah, for sure.
So you're using Talent already, you're using ClickView. What do you think about now they're coming together? Um, will that make your life easier? Processing things?
What I love about ClickView and Talent—it's to do so. I'm essentially trying to build a data strategy for an entire enterprise—help build a data strategy for an entire enterprise. ClickView is an amazing product because it allows you to make a quick buy versus build decision. You know, if you have several different data sources and you just need BI out of the box, they are there to support those organizations. And I think that it's a brilliant model to follow.
Um, because at the end of the day, like, once data is in a lake or a warehouse, it's agnostic technology, right? So how it got from a source system into a data warehouse—Click Talent, Power BI, Azure, you know, writing custom scripts in Python—it doesn't really matter. But once it's there and it's well-documented, that's so—that to me is really where there's a cool place for a company like Click to exist right now. Because I think there's tons of directors of BI like me out there that are struggling with, you know, TP sites and struggling with email attachments and things like that.
You know, they're not massive Fortune 50, Fortune 500 companies, but they're upscale. They have data. Yeah, and that's where Click is a fantastic—just walking around the floor and seeing some of the demos that they have right out of the box. That's the brilliance of their positioning, I think.
Yeah, and I see the partnership they have. I saw some partners as well doing version control, yeah, on ClickView as well—on Cognos and Power BI. Yeah. And then you think, yeah, they know what is missing, but that is tackled by the partnerships.
No, it really is. And I think that, you know, companies like Click didn't exist when I started my career, where you could just say, "Just procure some funds, you know, write the check for the software, a little bit of the services," and you could be up and running with visualizations in a matter of weeks. Yeah, it's just that—that to me is a very cool differentiator, I think, for lots of different organizations out there.
For sure. Yeah. Great one.
Um, Kevin, besides data music connectors as well, what is your favorite type of music or band?
Um, so that's a very, very big question. Um, you know, I think right now I'm listening to a lot of Grateful Dead, listening to a lot of Dead and Company, trying to decide whether to go see Dead and Company in Philadelphia in June.
Um, but I run the gamut. We listen to hip-hop in my house. We listen to classic rock. Um, you know, we listen to dance—you name it. But right now, heavier rotation with Grateful Dead and Dead and Company, I'd say.
Okay. Great one, Kevin. It was nice talking to you.
Very nice. Thanks for your time.
Thank you very much. All right. Okay. Cheers. Cheers. Thank you very much.
Foreign.
