Jennah James has been in data and analytics for over a decade with a plethora of experience across numerous industries including supply chain management, education, and media. In this podcast Jennah discusses how to determine what level of centralization or decentralization of data / business intelligence your organization should have and how to scale this at your respective organization.
Jennah James has been in data and analytics for over a decade with a plethora of experience across numerous industries including supply chain management, education, and media. In this podcast Jennah discusses how to determine what level of centralization or decentralization of data / business intelligence your organization should have and how to scale this at your respective organization.
Follow on social media @salientinsights:
🔗 TikTok https://www.tiktok.com/@salientinsights
🔗 Instagram https://www.instagram.com/salientinsights/
🔗 LinkedIn https://www.linkedin.com/company/salient-insights
🔗 Youtube https://www.youtube.com/channel/UCkINGLfE1woF6ukCixg0ckQ/
🔗 Website https://salientinsights.com/podcast
#dataanalyst #data #dataanalytics #datascience #excel #sql #python #powerbi #businessintelligence #careerdevelopment #jobsearch #recruiting #datarecruiting #datastructures #datascience #salientinsights #tableau #snowflake #databricks
Those are not concepts that are outside of the capabilities of the most creative mind. They're not, I don't believe it. Like you said, might need trained, but it's a trained mentality to think you're not capable as well.
Thank you for joining us here again on the podcast. We're having Jenna James, who's had over a decade of experience in data analytics. She has a plethora of experience across multiple industries from supply chain management to education and into media. Jenna James joins us here as a vice president of data analytics at TMB, a media company whose major brands include Fail Army, Reader's Digest, and Taste of Home, correct? You got it.
Those brands are including social, web, streaming, and print forms of media. So again, thank you, Jenna James, for joining us. And we just look forward to having a conversation with you. Awesome. Thanks for having me in. So from our previous conversations, we kind of went down a bunch of different rabbit trails. And some of the places that we landed on was just the idea and notion of structuring business intelligence and how to build out those teams. I know in our previous
Uh, podcasts with Wyatt, kind of conversation cut started to go that direction. I kind of had to beat it back because we we were going to talk. So, uh, if you wouldn't mind kind of just breaking out, how do you deal with decentralization versus centralization and structuring a BI team when it's appropriate where and how to do that balance? That's awesome. Yeah. I've seen it done a couple of different ways, and I find that more often than not it meets in the middle. And here's why.
When you go full centralized, you say, we're going to have everything in the middle. We're going to have centralized data pipelines, analysis teams, everything from engineering, data science, the whole bit. And you have everything there. You find, you starve the departments a bit. They're looking for more. And your central team, maybe you guys have enough cash and you're really going for it. And you've got just a huge number of people.
that people can sit back and wait for those requests to come in and can still fulfill them in a timely fashion. I've yet to see that totally done. Efficiency starts to fall off because now it's you're, dealing with a Luch Marger behemoth or like a centralized data warehouse or data lake. And then it's all of like the data curation and quality. so it's going to take us twice as long to get out something. Right. Totally. And you've got it to have whoever's picking it up is got to be have, have enough.
context to answer the question properly, to interpret what the business decoder is looking for. It's all of those things. So that tends to, over time, evolve into departments hiring specific people to communicate with the centralized data team. Right? Cause they are like, we need a translation layer, right? To put it in our lovely nerd terminology. We need a translation layer to help us understand how to ask the right questions of data, how to get it back from them efficiently.
So you end up getting those subject matter experts planted within a department to interface with the BI and data team. And very soon, all of a sudden, wait, we've become hybrid, right? Because now we're pushing more things to them. They're building some of their own tools, which can work out great. All fine. Well, some people then go, well, we need the inverse. We need everything in-house. And in some really large organizations have seen this done, where they're like, well, I want the data warehousing. We're the only one that knows how to do it. We'll hire a couple of engineers.
We need a couple of BI guys. We need some great data scientists. We're going to put together this great group. And we're going to have it just for us in pick a department, right? Performance of X, right? So we're going to have this whole thing. Well, sooner or later, our business goes, hey, some things are getting real expensive. Why has everybody got their own data warehousing contract, their own BI contract? Even if we've done some of this centralized,
gets expensive because you don't have that economy of scale. So sooner or later, they start going, well, could we at least join data warehouses? Could we all be under the same BI tool? Could we all talk about these same things? So you end up, once again, back in a hybrid environment, right? Because you've joined certain things together and you have certain things held in the business units. So I do think it just makes more sense to approach that, where it's probably going to end up anyways, and go towards that. But then you get into...
Do you embed? How do you get really good insights from this team and how you structuring it in a way that's really productive? So embedding can really be a challenge. And I've seen that done a couple of ways. I mean, there's a thousand ways to interpret that. I'm sure you've seen it done a few ways too. You can have, like we're talking about an analyst or two or a...
translation layer within the business units or within a central BI team, can say, split the company into segments for however many people are on your team. And then each person's responsible for their own area of the business. And they're going to know it deeply. And I think that to quote me and girls, the limit does not exist, right? Like you can never be embedded enough. It will never be enough. It will never be.
an outstanding data analyst or higher, and then also be somebody who knows exactly how every KPI should look and is in the mind of the marketer. Like those are two different positions. You've got to work together. You can't be totally in their mind. So it comes back to people knowing enough to work with each other and knowing enough to meet each other halfway. And, you know, we're all split a thousand different directions, but
We have to meet each other halfway. I'm lucky to work with a lot of people that do do that, but it's always a struggle for everyone is data needs to learn enough to know the business needs. Business needs to know enough to meet data halfway to make sure that we really understand those requests and can be embedded as best we can. But it's a bit of a fallacy. So if I was to kind of like dig a little bit deeper and say, okay, how do we frame questions for the audience? People who are listening in and they're in a different.
environment than you are, right? Maybe a different sized company, maybe it's three people in the whole BI team, or maybe it's a substantively larger team and you're a 60 to a couple hundred plus BI team. So what are some of the questions you would encourage them to ask? Maybe it's company size, like how do you frame, okay, here's a good way to determine how much to decentralize and how much to centralize.
oftentimes I think, and you could correct me if I'm wrong, the inclination is to centralize, because you want more control because you're getting asked questions all the time. And then it's like, well, so-and-so produced this data and how does it relate to your report? And you're like, great. Now I have to do a reconciliation of this random person stuff. And if I centralize it, then like that becomes easier. Right? So there's always that tendency to be like, let's centralize.
for control and then that makes it so much easier for us to manage and mitigate ad hoc inquiries or questions. But what are some of those questions that you would frame on how to determine how to create that proper balance for your each organization, right? Cause this might be a timeless interview where you have somebody that's in what an occurrence state now, but then the next state they're in, it's like, okay. So what are those timeless questions? Yeah. The first is what are you measuring and why and how?
that to me kind of is the crux of everything. So, and a lot of that might be a question, not for the business stakeholder in front of you, but for the finance team or for the most senior leadership. What are you measuring? What are you trying to achieve? Where are we going? What are the core initiatives? Not just this year, if you're talking about structuring a team, we got to look in five years, what are we trying to do? And I'm all about
that you got to roll with the punches. know we can, we won't get to this. I will leave this for one of your amazing future podcasts, but we can talk agile versus agile becoming rigid, despite its name over time, right? We can get into all of those things, but we got to keep flexible. However, you do have to know like, okay, we're our revenue stream is X, right? You mentioned in my intro, we've got a lot of revenue streams at TMB. So we can't measure one thing.
And I get that some businesses can, more power to you. That's awesome. But I can't and serve the business the way that it needs to be served. But what are the things that I'm measuring? What is that range? And what are those KPIs? How am I measuring it? Where am I getting that information? Because while that sounds like I'm starting with the data rather than the business, they're so interconnected on what are you measuring and how.
That then you want to say, okay, well, if I know that I've got three things we're measuring and they are neatly held within certain data sets, then maybe I can have total subject matter experts within the BI team that are serving that way. Maybe I can do that. But if there is diverse as our data sets are, that's not reasonable. Even for one person to handle some of our departments, it's really too much. They need the camaraderie and support of their
their team members in order to be able to properly serve each department. And when you're a small team, which most of them are in some way or another, you always want to be bigger. You always feel small even when you're not. know, the, want to make sure that you're serving in all the ways and relying on one person to think about every eventuality is also an interesting challenge. So some of the ways that I think about it.
So, I mean, you said it could be kind of a more data perspective of what you're measuring, but if I heard you correctly, I almost feel like I heard you. What is the direction and alignment and vision of that board of directors, the owner, whether it's private and where are they trying to go and what measurements and that may even cause, is that going to come from teasing it out of those individuals and figuring out, okay, if you're trying to accomplish this, what are the things we need to know?
that we're doing to hit these milestones, to hit these metrics, to make sure that we're in alignment with where he wants to be. Is that right? You're totally right. That's I think that you can't have business separate from business goals. So you can't have business data separate from business goals. So if your data team is going to supply your business with the insights that they need, they need to all be rowing in the same direction. And again,
Might be three directions if you've got three very distinct kind of revenue streams going on, but you need to all be aligned on, hey, we're really trying to increase, subscriptions are common in so many industries, so I'll use that. We're all trying to increase subscriptions by 25 % over the next three years, and we're going to try eight ways to get there, and our data team's gonna be locked in on how we support those as we even experiment. Again, coming back to that flexibility.
But yes, you've got to understand what are you truly trying to achieve, not in a department, but as an entire business. So if we're going to basically say it's alignment with the business and then it's also getting the right questions on what to measure and then producing those, the fracturing of centralized versus decentralized is going to come down to more of like, if you're ever in an organization where
the goals of the company require basically two systems. You're like, hey, centralization is pretty easy to consolidate that. But if you're an agency that has 10 different revenue streams and you're talking about 10 different networks of software applications, bringing that to a place and they're all operating in some capacity of difference. Yeah.
then decentralization has got to have some measure and degree of influence. Is that? That's very accurate. And I think I'm always partial. If your hybrid is in the middle of centralized and decentralized, right? It's there. I almost think of it as the wrong side of a magnet, right? Because you'll never get exactly landing on that hybrid mark, right? It's going to jump to the side. And if I'm going to have to jump to a side, I'm going to go on the central side.
Because decentralized, if you do have singular business goals that all tie into the data, decentralized is going to mean duplication and over cost. And data is a cost center. Rarely are we truly making money, getting closer and closer with some of really cool products we've worked on. Other companies, data teams are doing awesome stuff, but in general, still largely thought of as a cost center. You don't want to be adding more costs needlessly.
So I'm going to go more on the centralized side because you can run multiple streams within a centralized system. And so you can go that way. But I think the closer you get to hybrid and try to achieve that unachievable magnet jump, then it behooves the business best because you're trying to strike the balance of serving them where they are. That's a big topic I think you and I will jump into here.
serving the business where they are and reducing cost and duplicative business activities that are just kind of silly while also providing the insights needed. So one of the things that we've kind of lightly discussed is this notion that business intelligence and data kind of supersede almost any other kind of department within a company. And that's because they service everybody.
Right. And so then there's this always ever evolving change of who is business intelligence and how they operate and all those types of things. So would you mind, we kind of talked about some of the fallacies that can, and like some of the pitfalls that people can run into as a result of the, and this kind of ties into our previous conversation, but some of those shortcomings, those places where people can fall as a result of
business intelligence or data and analytics being so diverse? Yeah. So I guess that there's a lot of fallacies around embedding itself and saying that everyone can be fully in those different departments to the level that a marketer needs to. So a marketer is going to be thinking about their campaigns, about the KPIs they're hitting and how they're doing it.
a data person should try to be tracking with them and knowledgeable, they're never going to be at the complete level that you would want somebody who's a quote unquote embedded resource, right? Like there's that aspect. But then there's also the other side that I think you're getting at, which is if one person can't keep track of everything happening with across the company either, there needs to be this balance of
systems, documentation, and multiple people that are supporting it and coming to the table with the business user, right? So the business user can't be completely unaware. I think that there's this fine line between self-serve analytics and sending people on a wild goose chase, right? Like there needs to be, we're providing an opportunity for them to come to the table with us, but we're trying to meet them where they are. That being said, that's also really challenging.
Everyone I always make the argument it comes back to how people learn math as a kid like okay some things click with people I've had people demand line charts for Things that don't always make sense for it But it's what clicks for them and I've had things that I thought Were really easy to understand in a stacked bar chart and people come in and be like, I think you're duplicating all the data I'm like, that's just the total that's sitting at the top of each bar because it's totaling up everything that's in it, know, so people are really
really different on how they interpret this stuff. So I think it's when you're trying to create a singular tool for a range of people, you're always going to have challenges. And that's the other aspect of this, of having central decentralized teams, hybrid workforces. You've got to understand that you're going to need the volume of humans that can communicate with your stakeholders to help them interpret and use the tools that you're putting out there. It will never be a one and done. Here it goes.
Enjoy that dashboard. Have a great day. We're done. You're always going to have some interaction. Dashboards change, business tools change, people change out. You've got a new person coming in. There's always going to be an aspect of that. Yeah. And I think to that point, if you are able to just drop a dashboard and then have no feedback, guaranteed you're not looking at your usage reports or adoption reports because
Right. If there's no feedback, then it's not useful. Yeah. I find that, and maybe you can tell me if I'm wrong, within the organization, what ends up happening is the things that are the most useful to the business are the ones that are getting the constant feedback. Like it's just like, man, can I stop supporting this? Right. Like, yeah, And what a challenge that is. You totally nailed it. It's like, we created this awesome thing that's so useful for so many people, but again, kind of goes back to my comment on.
how different people interpret things differently. One person's gonna come to you for the same tool and be like, this is so close. You guys did great, this so close. I just need this one tiny little tweak. But the next person's like, don't change a thing, you finally got it right. And the next person's going, how can anyone interpret this garbage? It's got all the data, but how am I supposed to see what's going on? And you're like, okay, the last two people are thrilled, but you don't wanna end up with three tools that everyone's using slightly differently, because you get in hot water so fast.
People interpret nuanced data really, really quickly. You think a dollar's a dollar or a user's a user. And so often, there's so many nuances that everybody interpreting needs to understand and needs to know those pieces. I'm certainly guilty of doing exactly what I'm asked. It's actually a terrible thing to do. Hey, I need this report. I need x, y, and z on it. I need you to pull it for these dates.
This stuff I'm calling this person knows it. They so know exactly how to give me a request. How easy. I'm in two seconds. Send them off to a great. Have a great day. Take care guys. thanks for getting me that quickly. Two weeks later. Hey, why'd you send them those crazy numbers? What? What do mean? What happened? Like, well they they put this filter in this filter and then it said this number like, my gosh. Okay, they totally didn't know how that is supposed to work.
Yeah, it's always amazing how a business user can come back to you and say, like, you start to think they're really savvy because they're asking all the right questions and they're working with you really, really well. And then all of a sudden they're like, my report's broken. And you're like, what happened? And they're like, no, all the numbers are wrong. And you're like, okay. And then it's like, you have like, you left a filter on when you were researching something and you just clear the filter and they're like.
Oh man, you're a genius. Thank you so much. they that you swear like they you're just like this beyond godsend brilliant person. You're like, was a filter. Yeah, totally. Totally. And some people are like, I can use a dashboard full of filters. Like I'm good. Like I don't need walk through as you hold their hand through something and they're like, ease up guys, little control freaks. All right, easy data team. And then the next person data teams like, Hey, you're so capable. Enjoy that. They're like, what am I supposed to do with this?
So it is, you'll never get it right in some levels. So you have to accept that too, and just take the good. So at what level is your perception on self-service, right? So some people look at it as this unicorn that's never attainable. Some people are like, no, it just requires a ton of training and requires a perfect and pristine data definition and semantic layer design, and you can achieve it, right?
I mean, on the spectrum, as it relates to kind of in concept of like, servicing the larger organization as a BI team, mean, what's your perception there? Athletics. It's so I'm a sports person admitted, but what is perfect fitness? Are you ever going to achieve it? What have you spent every waking minute on it? What have you spent all your sleeping minutes also trying to achieve that? Well,
With complete dedication, unlimited resources, still most people aren't LeBron, right? Still most people aren't. Even people that have invested an incredible amount into achieving those kinds of goals. But there's a lot of Olympic athletes out there. They're pretty incredible. Most of what they're doing, rarely do we have people in many Olympic sports. Those are your total next level people. They're real anomalies.
But even Olympic athletes, relatively rare, and they're going for one single goal. BI and self-serve analytics rarely has one goal. They're choosing, they're trying to serve one team like this, one team like that, one team like that, and they rarely have every resource available to them. And perfect inputs, all of those things, right? So is it something to strive for? Do I think we all should take care of our bodies in whatever way?
works for you in your life and the other demands on it. Yes. Yes, you should. And yes, you should strive for that goal. But you got to be okay that it's never going to be perfect. That's kind of my view on it. Will you have anomalies that can make it perfect with crazy resources and the right genes and right setup? Sure. So using your athletics analogy, for those who are familiar with it, it's kind of like the Barry Sanders and the Michael Jordans of the world.
can play across three different sports professionally, but even at the end of the day, they still had their one, you know, Jordan never really made it as a baseball player. but he was the king of basketball, right? And so, so I, I love that analogy. I think it really fits on, you can only serve so much and you can only do so much. so.
And maybe I have maybe that picture I gave with Michael Jordan. You can make one department happy maybe with self-service. Maybe. But doing it organizationally diverse is a little bit too far. It is. And you're going to have those that have embarrassment of riches on resources and can...
get a lot closer than others and that's great. That's awesome. of riches. Break that down for me. How do I say? You have so much that you almost feel shame. The embarrassment of riches, of like, my gosh, have so much that you really feel obligated to give it away because not everybody has all of this. So if you've got an embarrassment of riches within BI,
Are you serving your data science group? Are you serving your other business units? You hope that that embarrassment of riches is coming from everyone having it. but just if you have just so many resources that you can really pour, you know, one person per dashboard doing nothing but building and maintaining it. I've heard about companies that have that. Awesome. That's great. I think in most of those cases, they still run into errors. can't fix incidentally.
The danger there is that, and maybe this might hurt somebody's feelings, but you're going to end up getting cogs and the best BI and data people would never be satisfied with doing one thing. They're too curious and too like, so you end up getting a lot of people within the organization, but that doesn't necessarily mean that they are going to be that like hungry to learn figuring out how do we add value here? What's this piece doing over here?
And if all you're doing is servicing one dashboard, you're going to get a lot of complacency. yes, couldn't agree more. And I really don't like the glorification of longevity. I think there's incredible value and people staying in one place for a while, but it needs to be moving up or around or side. Like there needs to be progression. I'm totally with you there. And I think that there was a time and a place, in our history, where, companies
treated people who were fulfilling the same activities every day. And there was a necessity for faithfulness on both parties. On both parties. Excellent point. so, but now, I mean, that might be different in other roles, but in BI, there's no way. There's no way. You're going to be stifling the value you can add to the organization by remaining in one place and one industry, working with one set of tools.
Yeah, there's a great book by David Epstein. did, he's done several, but the one I'm thinking of is called Range. And it talks that most people are not your Tiger Woods that had a golf club in his hands at six months old. Most people aren't that, and they're better for it. On my team, I've got somebody who was a screenwriter for 12 years. Can you imagine what he's like telling data stories? He's phenomenal.
You know, I've I've got civil engineers. I've got people who were very high achieving baseball players. They totally get it on a different level. They tell stories, they get teamwork. There is so much benefit in a range of backgrounds. I actually view it as too bad when I see companies that want to hire or really focus on, well, if you don't have retail experience, I don't know how you'd possibly do this role. Maybe that's true in some areas of the business, but it's not true in data.
in data and I think in a lot of tech roles at those companies, that company is better served by those that have a wide range in their background and they bring that to the table and add. Well, I mean, I think the number one skill in BI or data is one's ability to learn, which is what I would kind of measure as intelligence. Right.
I'm an average guy. Like I truly believe I'm an average guy who's just like acquired skills to learn. Right. And I think it is a skill like, and the more you exposure you get and the more hungry you are to figure out new things, the more you refine and develop that skill. And that's why, especially like for me, when I'm looking at add those perfect team members, they're going to have crazy backgrounds. Like there was one guy who I
brought it, we interviewed and brought onto a team and he was like composing ballets. He was a chef for a restaurant and, on the strip of Vegas, he was, mean, like just these crazy like careers and he was in them for like five to seven years. And then he'd been in beta and BI for like two to three years. And it was just like, you know, you get that diversity and all of a sudden you just see like a totally different type of caliber of candidate.
I completely agree. look for that in candidates and I people that'll come and say, I'm diehard data, I eat, sleep and drink it. I dream about it at night and I'm like, okay. But we're not as benefited from that. I find when I actually, hike a lot and I am out there and when you let your brain rest in a completely different context, isn't surprising to anybody.
sleeps or does activity or meditates anything, you you find, my gosh, that problem I've been obsessing over and not letting myself not think about for weeks. The second I stopped thinking about it, I solved it. Oh yeah. I can't say how many times I've walked away from something I was working on to come back and it was like, man, I feel like the hour ago, me was an idiot. Yeah, Completely. So I think that. Go ahead.
No, I was just going to say that the best way to solve a problem is to walk away. I think Einstein and think, what is it, Michelangelo were kind of noted for their interest of like shifting every several hours. So they would like practice music or art or whatever every so often. And I think it really creates this Renaissance type of, and your ability to kind of come back and it's like, you have a fresh perspective. Totally.
Yeah, I mean, one of my first jobs out of college, was working for a nonprofit and we're doing fundraising and we had huge events. And so there's five of us in the office and we'd be there and it'd be 10, 11, 12 at night. And I was the junior member of the team and I'd start going knocking on doors. Like, we got to stop. We're going to be back here in seven hours. Like we've got to stop. Like, what's one more thing if I just get this one more thing done? I how long have you been trying to do it? Well, I've only been on it for like 45 minutes. I go take you five in the morning.
It'll take you five, go home. Like, and we'd get us all together and we'd get out and we'd come back in the morning. They're like, it took me six minutes to do it. Like, yeah, like it's normal. There was also fatigue and also everything else in there, but it's the same. These things are all highly related, right? Yeah. Anybody who's done any form of development knows that that comma or semicolon that you couldn't figure out took for four hours or even a whole day.
You walk back, you just walked away and when you came back, was like, really? Really? Exactly. It's, it's that, it's that kind of thing. And some of that, I love your point about learning though. And I, it kind of goes into these other things that we're talking about, cause they run into often people that are in other areas of the business. And, know, we're starting to talk, what, what do you do? What are you working on? You're working on this. And I'm like, data.
cells really hard. can't, just don't, my brain doesn't work like that. I can't, I don't do that. I need somebody to tell me what it means. I can't interpret this kind of thing. And I do have a reaction to it. I really struggle with that sentiment because I don't believe it. I with brilliant creative people throughout TMB. will say so many, and many in other walks of life that I've interacted with. I paint, I like to build things, know, like, so I have,
a bit of that in my personality too, certainly not to some of the professional levels I see, but this I can't or that's a different side of the brain that I don't have or I don't work on. I really struggle with for a few reasons. One is I just don't believe them. I think they're all capable and I think it's a mental block. We have a. It's trained mentality. It is a trained mentality. Absolutely. It's totally learned thing. When they were a child, they didn't say like, that that math
puzzle you want me to do versus the letter puzzle, I'm incapable of that one, but I can do that one. Like, no, they didn't. They went and played with it as a kid. You've got kids, you know what they're like with that. But I see so often also that I don't think it's sustainable. I truly don't think it's sustainable. Seven, eight, nine years ago, I had someone I absolutely adore in the industry tell me, analyst is the new typing pool. And I went,
Kind of know what a typing pool is, but I'm also a little too young for that. But you go and look at it, you know, you can imagine nobody types. That was below people or that was somebody else's skill set. So people dictate what they wanted to say in a letter or something of the sort. And you'd have this literally like a bullpen situation of everyone typing and they would be the typist. Well, and that was the typing pool, right? Like they were all there. Analysts are kind of treated like that.
I've walked into rooms where it's a room full of analysts and they're taking everyone's requests. It's dying. And that's got big implications, both for the business stakeholders who need to have the baseline skills to do it themselves. They have to be able to type. That doesn't mean they have to know how to type in German. It doesn't mean that they need to be a thousand words a minute. Like, come on, there's, there's reasonable aspects here, but they're going to have to be able to analyze. And it also matters for the analysts. The analysts cannot be content on
I can build a dashboard. I can put together three data sets with joins and put them out in front of people. Not good enough. Absolutely not good enough. You better be doing some predictive analysis. Get comfortable with statistics. Brush off that high school notebook if you even got it in high school, meaning didn't. You've got to be ready for that. You've got to be ready to move into, because the world's changing. A lot of people are aligning it with
chat GBT and AI revolution. always remind people AI has been around a long time. Gen AI is relatively new. But a lot of this has to do with hardware. We're not even at Gen AI. Yeah, totally. We're not. We're not. So people have got to be ready for that. But unfortunately, we still have, at least in the US, a school system that prefers to teach calculus over statistics. Calculus is used in some.
areas and is fascinating. Love it. Statistics is used every day by everyone, whether you realize it or not. And yet we're uncomfortable interpreting basic probabilities. And we've got to get to a place where we're more comfortable with statistical significance, with z scores, with standard deviations from mean. understand that that only applies to bell curve data.
Those are not concepts that are outside of the capabilities of the most creative mind. They're not, I don't believe it. They are within those capabilities. Like you said, might need trained, but it's a trained mentality to think you're not capable as well. And we can train to think that we are capable. And that comes back to our, do you structure a BI team? Well, people got to meet each other in the middle, business side and data side. So we better be prepared to do that. The analysts better be prepared.
to speak the business language, to understand what they're trying to drive at, to not pick a really pretty dashboard that doesn't tell them anything interesting about how to make a decision for their business. And the business user better know enough to interpret and to work with the data team to get to things that give them those insights. I mean, it's gotta come together. Right. Those things that drive them to produce the greatest value for the organization. That's exactly right. Exactly. So, I mean,
To that point, I mean, I know one of the things we've kind of talked about is like dealing with stakeholders that don't exactly know the cost of what they're asking because they don't have maybe the best comprehension of data or knowing how to ask for what they're wanting. And we've kind of discussed and talked about that. Would you mind kind of sharing your thoughts on that? Yeah, I think that
By and large, people want to understand, and it might be some of the mental blocks we were just talking about, but we got to get a little bit more comfortable with how data comes together. And if we're not collecting data on something, I can't have an insight for you next week. We got to collect first and have a critical mass there. And sometimes there's a block on like, well, but I need to understand this thing. Can we start collecting it? And you kind of go.
Okay, we got to, we got to take a step back and really look at kind of the foundation and how this is coming together. The question I ask people to ask, I try to remind myself to always ask is what are they, what's the decision they're trying to make? And we business use case a lot and that's good, but it can be interpreted a lot of ways. What decision are you trying to make? Well, I want to know whether I should stop this campaign or continue it. That's a decision. Now let's talk about how we're going to evaluate that. And it might be more than one KPI and that's okay.
No, love that. Decision. Yeah, because normally I'm like, what are you trying to accomplish? Right. But even then you get, well, I'm trying to pull this down into an Excel sheet and then, so that's what I'm trying to accomplish. And you're like, no, no, no. And so you end up falling several times, but what decision are you trying to make? It's like a jumper wire. It like totally short circuits.
that notion of like, are you trying to accomplish? Cause you can end up going through a series of 10 different questions of, why are you trying to pull that down, that Excel sheet down? And they're like, Oh, well, because I have this other Excel sheet. Well, what are you trying to accomplish with that? Oh, well, I do a VLOOKUP into that. Okay. Well then why do do that? Right. And it goes on and it keeps going. And then finally it's like, Oh, well, cause I give this report to so-and-so VP. What did they do with it? Oh, well, they make this and this decisions based off of it. And you're like, but.
If you're asking the question, what decision is being effectively made? I love that. That's so, that's a great way to reframe things and short circuit some of that what you're trying to accomplish. Yeah. Yeah. And I've certainly been in those same situations of like, I just need this piece. And sometimes, especially if that piece is relatively easy to get my knee jerk is give them the piece, give them the piece and let them go in peace. Right? Like, let, let's just let them move on. but it doesn't.
typically help the business. You do better when you ask the, what decision are you trying to make and get to the root of it. And so often like you got to the point, well, I'm V looking it up, V look up thing it, verbing that onto this other thing. And then I'm doing this and then I'm doing that. And you get to the end and realize, wait, we already have something that does that for you, but you've done all of this other work. it's, you know, those questions are important too, but the decision cuts through to what's the end goal there.
And I've found a couple of times like, gosh, an end goal is being made, but every data person knows averaging averages. Good way to get crooked real quick. I've found more than more often than Using averages across different greens, right? Yeah. got asked that one recently and I'm like, no, we don't want to do that. Yeah, totally. And that's been, and you realize like, my gosh, that this person didn't realize that this was already averaged and couldn't be again.
even if they are aware of the dangers of that. So it's what decision you're trying to make and then backing into what you need from there is so important. And some of it's hard because you'll say things like, well, we need to monitor how this is doing. We're not making a decision right now. We need to keep an eye on it. I get that. So then it takes a little bit more digging. But again, this comes to data people trying to make sure that we're meeting
the business stakeholder where they need to be and the business stakeholder coming to meet the data person where they need to be and trying to make sure that that's a as clean of a handoff as we can. It's never going to be perfect. Yeah. Well, thank you again. Well, thank you again for just joining us and being able to just talk about the structure of BI, some of the how to build a team, how to interface with the business kind of perspectives on self-service. then all the way down to just how to deal with stakeholders who may not know.
how to communicate or understand the level and depth of the requests. And so really appreciate you taking the time today and look forward to potentially having you on again. Thank you. I'd love to. Thank you for having me in. I'm looking forward to seeing more of your episodes come out.