#212: Innovation Through Analytics Within the Enterprise with Dr. Tiffany Perkins-Munn

What’s more sexy: analytics or innovation? What about combining them! That sounds great, and Thomas Davenport would be so proud if you pulled it off, but the reality is that the idea of innovation through analytics is one thing, while the reality of making it happen is another thing entirely. Dr. Tiffany Perkins-Munn, Head of Marketing Data & Analytics at JPMorgan Chase & Co., joined us for a discussion on the subject!

Links to Articles and Other Resources Mentioned in the Show

Photo by Prateek Katyal on Unsplash

Episode Transcript


0:00:05.8 Announcer: Welcome to the Analytics Power Hour. Analytics topics covered conversationally and sometimes with explicit language. Here are your hosts, Moe, Michael and Tim.


0:00:22.0 Michael Helbling: Hi everyone, welcome to the Analytics Power Hour. This is Episode 212. Right at the top of any senior executives list these days is the need to unlock innovation within their org. How do we do that? And how do we in the data world play our part in that process? Frankly, a lot of talk happens about this in the world, and it’s hard to find people who are really mapping out a plan and executing effectively.

0:00:54.4 MH: I think that’s what we really wanna talk about today is, who’s out there doing that? And Moe, I think about you as a data leader at Canva and I know you’re working in a very fast curing organisation, but do you see this as something that you’re grappling with in terms of your role?

0:01:13.1 Moe Kiss: This is like legitimately the reason that I can’t sleep, other than babies.


0:01:19.8 MK: Pretty much what cases me up at night. In work thought stuff. So that’s perfect.

0:01:25.5 MH: And Tim, my other co-host here, you’ve been in the analytics space for, I don’t know, forever, 20 years-ish now. And I know you’ve been a part of every layer of this, so I’m excited to see what you think about this as well? And I’m Michael Helbling. But to get this show off on the right foot, we wanted to invite a guest, someone whose career has been built around doing this at scale.

0:01:50.9 MH: Dr. Tiffany Perkins-Munn is the managing director and head of Marketing Data and Analytics at JPMorgan Chase. Prior to that, she held senior research and analytics roles at BlackRock, Citadel and Merrill Lynch. She was recognised by CDO Magazine as one of the top 100 global women leaders in data and analytics. And today she is our guest. Welcome to the show, Dr. Tiffany Perkins-Munn.

0:02:13.3 Dr. Tiffany Perkins-Munn: Hello, thank you for having me. Happy to be here.

0:02:16.1 MH: Awesome. Well, I think a great way to kick off this conversation would just be to learn a little more about you and some of your background. Obviously you’ve held a number of these roles, and so these are big organisations that you’ve had some pretty significant impact with, so I’d love to just hear a little more about your story and we can take the conversation from there.

0:02:36.7 DP: Yeah, sure. So as you mentioned, I’ve worked across the street, I’ve done investment banking, retail banking, I’m in consumer banking now, I used to work at a hedge fund. So I have had experience, I worked in global markets. And I’ve always been in the data and analytics space, so I probably am one of the few people who know where all the bodies are buried, so to speak. In data anyway.


0:03:07.7 DP: Really trying to make the connection between as an industry, as a financial services, industry, what are some of the similarities or what’s consistent across the different types of financial service businesses? What can we learn from each other? So I’ve been in that space for a very long time, sort of working across the different industries in a very interdisciplinary way.

0:03:33.2 DP: And when I say “interdisciplinary”, I mean taking learnings from my time in market research, so lots of different opportunities to understand what some of the challenges are with data and where some of the innovation is coming from in data, and then to apply that to what we’re doing here in the financial services space.

0:03:55.0 DP: And then ultimately there was a period in my life when, this is gonna sound really off the beaten path, but I completely left the street. So in 2010 my husband was diagnosed with pancreatic cancer, that was in April of 2010, and in September of 2010, he passed away. And my daughter was five, and I was trying to figure out like, what do I do now? I’m working in the city, at the time I was at a hedge fund, I’m leaving my house at 05:30 in the morning, so what do I do in this situation? And I decided to actually leave the street.

0:04:33.4 DP: I ended up doing a franchise, so I worked with a franchise consultant. You want to work with a franchise consultant who has a breadth of inventory, because you don’t pay them, they get paid by the business that you decide to invest in or to buy. And so you want somebody who can introduce you to a breadth of opportunities, right?

0:04:56.4 DP: And then ultimately they came back with Title Boxing club. TITLE Boxing Club is a concept where, I don’t know if you’ve heard of it, but you wrap your hands, you put on boxing gloves, you work out to a trainer on a heavy bag, it can burn up to a thousand calories an hour. Sounds like I’m a walking ad for, doesn’t it? I don’t own it anymore.


0:05:21.2 DP: But so I did, I went into TITLE Boxing Club, and we had over 500 locations, we had clubs in Cancun and internationally. It was a wonderful opportunity and a very unique opportunity to once again use data and analytics in a different way. So now here I am with trainers and food plans and point of sale systems and operational models that are extremely completely different from what we have been doing in financial services.

0:05:52.4 DP: And I am like the ultimate boss, there’s no one above me. It all stops with me. So really figuring out how we could be really critically thinking about the analytics. Even analytics around how we chose our location, so where are the most successful locations, how do we build a model to help us predict whether or not a location will be high performing, moderately performing or not.

0:06:21.8 DP: So anyway, I did that for five years, and then after I had done it for five years, I was presented with a choice, actually. Someone with whom I had worked, came to me and said, “I am going to be the CMO of JP Morgan and I need an analytics team, and are you interested?”

0:06:47.1 DP: And I was all in my gym days, I was like, “Nope, I will consult with you, I’ll help you build the team, I’ll find you a great analytics leader, but I am really into this gym.” I had bought multiple territories, so I was supposed to build out these other territories. So that’s how we started.

0:07:05.5 DP: And then I just realised that I was getting pulled in more and more and more and more and more, so when the time came where I really had to make a decision, should I get out of the lease? Or not get out of it? It was up, five years at least was up, I needed to renew. Should I renew the lease and build out these other territories, or should I go into JP Morgan full-time?

0:07:30.5 DP: I had a very hard decision to make, and ultimately, because I was really enjoying the work I was doing at JP Morgan and I was hoping it would last, ultimately I decided to let the lease go. I sold the other two properties that I had not yet built out and I came to JPMorgan Chase, and at that point I came into the global markets or the investment banking business. And so there was a period where I was gone, but I was still very focused on analytics and then I was back.

0:08:02.0 MK: I’m dying to know, what about the work that you were doing at JP Morgan sucked you back in? There must have been something in you that was triggered of like, “I really enjoy this and I find it fulfilling, and this is interesting.” What were those things?

0:08:18.9 Tim Wilson: You had to have missed the regulatory environment. Is that…


0:08:24.8 DP: There it is. How’d you guess?


0:08:27.6 MH: That’s right.

0:08:27.6 DP: You know what it is, is that because of the type of business that JP Morgan is, meaning like it’s been around a long time, they’ve tried lots of things, they’re often first mover in many places. In this space, especially in the investment banking space, it’s still not a space, just generally speaking, where I find firms are really forward-thinking when it comes to what we’re doing in data and analytics. Because the clients aren’t consumers, the clients are big, huge firms, like a hedge fund could be a client or a retailer, a big retailer could be a client, right?

0:09:07.3 DP: So this was really an opportunity to think about, because ultimately, when you’re thinking about how well does your customer perceive you and you’re thinking about customer as a business, as a corporation, here was an opportunity, I thought, to really pull back the onion. Because guess what?

0:09:26.6 DP: Hedge fund number one is not actually making the decision about whether they trade with you or not, there’s a trader who’s an individual, a person who sits within that hedge fund, who’s making that decision, in collaboration with probably a portfolio manager, in collaboration with analysts who are probably sending them research.

0:09:49.4 DP: There are a group of people who are making these decisions, and in the investment banking space, how can we peel back the onion and get to those people and start talking about this, if you think about it as nested, like if you think there are schools, there are classrooms and then there are kids within classrooms in that nested way.

0:10:09.6 DP: How could we use that, which is more of an educational reference in a hierarchical linear model, which is more of a social science kind of methodology, how could we use that to then think about people who sit on desks, who sit within businesses. Right? And so that was really exciting to me because I saw an opportunity to do something really interdisciplinary, and I saw an opportunity to do something in a space where there wasn’t a lot of work happening.

0:10:36.4 DP: Because sales people aren’t… They’re thinking about their account in aggregate it, and less about how John Doe who sits on the account as the trader is really the one who’s affecting change.

0:10:49.1 TW: So I think… I mean I made the joke about getting back into the regulatory, the regulatory environment. I think of healthcare and financial services as being two areas where there’s a rich volume of individual person level data. There’s so much, if you just put ethics and regulations aside, all sorts of crazy things you could do with it, but you can’t put ethics and regulatory stuff aside.

0:11:15.5 TW: So when you were talking earlier, rattling off a bunch of CPG businesses or retail, how much does that weigh in when you’re on the financial services side of things? How much does that… Is that a constraint to the innovation when you’re using data? Or do you just have to be more careful? I’ve had financial services clients where I feel like we’ve run up against that.

0:11:44.4 DP: I don’t feel that it’s a constraint to getting the work done, I feel like sometimes it slows us down. Because if you build a model and you want to use that model to then target a certain population, there are regulations that say you need to have that model reviewed by a regulatory board to make sure it’s adhering to like fair lending processes, for example.

0:12:08.8 DP: And because there is that additional step, it might make you go a little bit slower, but once you get used to the process of how it happens in the sort of the financial services ecosystem, I feel like we start to build that into our process. So when we talk about model building, we’re not just thinking about the analytics and the data, we’re also thinking about the regulatory pieces, the legal pieces, the compliant client species and how those all fit into the timeline.

0:12:39.9 DP: It’s not as if we’re competing and there’s somebody over there who’s doing it faster, ’cause guess what, everyone has to go through the same regulatory legal compliance hurdles, right? So it’s sort of like par for the course in that business, and you just kind of get used to that. But I won’t say… I completely agree that because it slows us down, it also makes us pay very close attention.

0:13:06.4 DP: We have opportunities to iterate, to think strategically is that what we’re trying to do. What is the… And ultimately to… You know, these initiatives always start with a product typically has an idea or wants to be positioned differently to the consumer, and so we are always, at least in the role that I’m in now as the head of marketing data and analytics, we’re always thinking about what does that one change, even though it’s an operational, technology or data change internally, how does that impact the consumer?

0:13:44.2 DP: What does the consumer to have to do differently, now that we made that? What’s the consumer’s experience now that we made that change? Is it going to be a hardship or more difficult for them to do what they’re trying to do on the website, for example, now that they made that change? So we’re always…

0:14:00.4 DP: I feel like it gives us ample opportunity to kind of flip the question on its head and make sure that we are always thinking about the consumers and how they are engaging with us and what that particular switch might do to change their engagement.

0:14:18.4 MK: Do you feel like the buy-in that you get from, say like leadership, is also potentially different? Because I work at Canva, which is total opposite, very, very fast-paced tech start-up, throw something to the wall, see what sticks, and then build and iterate. It’s so fast. I tell people on my team they get whiplash sometimes, which I love.

0:14:43.0 MK: But one of the things that I can see as being like such an advantage to this is like number one, because there is a lot of regulation, I imagine it has leadership attention. And because it has leadership attention, therefore you might get buy-in to data projects that otherwise you might struggle to get the attention of, like the attention of the leadership team. Is that an accurate…

0:15:06.1 DP: Yes and no. Because in the final analysis budgets matter. Right? So yeah, you could have a great idea. How much is it gonna cost? What are the resources that are needed? Are those resources available? Do we need to go pull those resources from somewhere? All of that is part of the decision criteria that goes into whether or not a great idea.

0:15:32.3 DP: It’s 200 and I don’t know, 20 or 40,000 people at JPMorgan Chase, the whole enterprise. You could imagine that great ideas are like everywhere, especially as people are in innovation mode, they’re thinking about what’s next, what’s going to be cutting edge, what’s going to be forward thinking. So it really depends on…

0:15:53.3 DP: And then sometimes you come across people who are risk-averse, so some leaders are going to be more risk-averse than others. Some are going to be like, “Go out there and do whatever, use whatever you need. Fine. We’ll find the resources.” And others are going to be more like, “That doesn’t seem to be in line with our goals for this year, we’ve already identified our goals, let’s stick with those,” unless it’s something that’s really out of the box. So you get a mixed bag, I’d say.

0:16:21.9 DP: But in general, I think because of the role that I have in marketing, data and analytics and it sits, I have a big team that works on all of these efforts and we sit very close to leadership, and I feel that they are very supportive, understanding, they guide us, help us figure out.

0:16:45.0 DP: ‘Cause as I mentioned, there’s so many new ideas coming down the pike, initiatives coming our way, how do we prioritise, how do we figure out… If it’s a quick analytics exercise, that’s one thing, but if it’s like we need to do test and learn for some initiative over several months, that’s another, right? So how do we prioritise those competing demands.

0:17:07.0 DP: And sometimes it’s like, do we have the data? Is the data or this… Or is this ask, or is a data exercise part of this ask? Does this ask mean that we need to pull disparate data sources together in a way that they aren’t pulled together now in order to execute? Or is this an ask that where we already have the data pretty centralised, with a couple of tweaks we can have easy access to it?

0:17:34.0 DP: So all of these are sort of questions that help us determine how we should focus, what we should focus on, and whether or not senior leadership across the organisation is going to be really supportive of the idea.

0:17:50.4 TW: So do you have the expectations of what the data can do? Like if you’ve got… Do you run into leaders who are like, “Go fast, this is… ” They’re kind of in love with the idea of innovation through data, but they have unrealistic expectations that they think that oh if they say “go”, that you have all the data you need and it’s centralised and it’s just gonna be checking in a box?

0:18:17.3 TW: How much time do you and your team need to spend setting expectations that it’s not enough to just have enthusiasm, it’s not enough to have a billion rows of data, ’cause it may or may not be the right data, how much of the work is trying to sort of coach and educate and set expectations?

0:18:41.6 DP: I would say it’s part of every single initiative. It just is, right? Because think about it, the further you move away, if you are a senior leader, the further you move away from where the work is actually happening, like the execution of the work, you lose the ability to fully understand the nuances, the detail, what’s involved, who has to do what. You lose all of that. It’s just the nature of sort of moving up through the ranks.

0:19:12.9 DP: So for us it’s really important that we always start at a position where we are, especially if you’re not… Let’s imagine like a CEO of a company who has never been in data and analytics. I think it’s different when you’ve been in data and analytics, but just imagine the CEO of a company never been in data and analytics, the company is really large, so there are lots of hierarchies and levels.

0:19:36.4 DP: And so helping… Part of what we do in data and analytics, I feel, and this is not just for JP Morgan, but any role where I sit, is that we are the objective truth tellers. And so in order to do that, it’s very important that we not just, we don’t just take a good idea and run with it, but we really set out the parameters.

0:20:00.2 DP: Like, “Here are the elements that are involved in the execution of that task. Here’s what it will give us. Even though you think it’s gonna give us X, it’s really gonna give us Y. Here’s what it will take in terms of time, in terms of data, technology, systems, etcetera. And here’s the length of time, because all of these things have to subsequently be socialised.” So really working through when any time we build a plan for any initiative, it includes all of those elements. Because for me, it’s very important to take people along on the journey.

0:20:32.9 DP: And I say things like, “I’m not sure how much you know of this already, but I’m gonna walk through it anyway, so that we’re all on the same page. That we’re all aligned.”


0:20:45.8 DP: Because you’ve been in situations where you get halfway through the project and people are like, “Oh my God. I didn’t know that’s how you were doing that?” And you’re like, “What do you mean you didn’t know that’s how I was… ” Right? So to avoid that, and by the way, it still happens even with this process, but to avoid that, I really try hard to bring people along on the journey to make sure we’re laying out the parameters in a really transparent, objective way. To keep updating them.

0:21:11.8 DP: It’s almost as if you have to build your own communication strategy in a way for every initiative, so that with every initiative comes, “Oh, okay, I’m talking to the team that’s building, I’m talking to technology, I’m talking to these senior leaders, I’m getting together with these senior leaders,” like there’s a strategy for how the elements and the process and the development of the plan is executed, and you wanna keep that as part of your broader sort of what you have to do in to be successful list, if you will.

0:21:43.2 TW: I love that as like a disarming way to like, “Oh, you probably already know this.” They probably don’t already know it, but if you ask them, “Do you know this?” they’re gonna say yes, ’cause they’re gonna feel like they should know it. So you’re like, “You probably already know it. Humour me, let me run through this.” And then we don’t have to reveal that you didn’t know it and yet you now, you now know it.

0:22:03.3 DP: Right.

0:22:04.0 MH: I love that as a strategy.

0:22:05.5 DP: Yeah, I find it to be very effective.

0:22:08.8 MK: I just love your point about having a communication strategy for every project, because it’s actually something that comes up with my team a lot, is like how do you actually demonstrate the value of the work that you’re doing to the business? And the truth of the matter is, I was actually chatting to a machine learning engineer a couple of weeks ago, and he was kind of like, “But the work is so good and everyone should be actioning this.”

0:22:28.0 MK: And I’m like, “Cool, that’s nice that you think that, but you need to communicate to the business why they should actually, and why they should listen to you.” It’s not enough to just do good work and have it speak for itself, but the framing of a communication strategy actually implies that you’re thoughtful and you’re going to put the working of like, “Actually, who do I need to communicate with? At what level detail? How am I gonna do that?” I think that’s something that a lot of data practitioners could find value in implementing in their everyday work.

0:23:02.9 DP: Yes. And I think what you wanna do is that you wanna move the way… Well, you know how machine learning technologists and data scientists speak. You wanna move it as far away from that. Speak into the language of the business.


0:23:18.1 DP: Right? So when I’m doing my storytelling, data storytelling seminars, I’m often asking people to not use acronyms, explain things, instead of saying it’s the blah, blah, blah, blah, blah, “Task efficiency methodology.” Nobody knows what that is, right? Just tell us what it is. And I also…

0:23:39.3 DP: And what happens is that when I go speak to people, so as part of this communication strategy, I’m often talking to people who sit in the businesses for the thing that I’m trying to develop, and I start having conversations with them. In those conversations, because they feel very empowered now because I’ve engaged them, they give me lots of examples of things like, “Oh yeah, I need to be able to do X, Y, and Z,” or, “I had a client and this is how it worked, and it would have been so much better if it worked this other way.”

0:24:08.3 DP: Amazing. And then guess what I do? I take those examples and that language and I incorporate it into my presentation, so that every time I go see another leader or a senior leader, I’m talking through the voice of someone who sits in the business with whom they are intimately familiar, and I move it away from “regression models and machine learning and supervised learning” and more into the language of what they are trying to achieve for the consumer, but through the lens of people who are actually doing the work.

0:24:38.0 DP: And I think that’s where the impact really comes, it’s really translating. ‘Cause in the data and data science and engineering space, we do very complex, sophisticated kinds of things that are often difficult to translate. It’s as if you speak a different language. And so you really have to figure out how do you then take that and turn it into something.

0:25:01.6 DP: Imagine the person is a super smart person, super smart, but has no data and analytics background. How do you translate that into something that’s easily digestible, bite-size and understandable, so that they can engage with it? And then you know you’ve won if then you hear them out telling somebody about it and they do it properly, and you’re like, “Oh my God.” That’s a big huge win.


0:25:30.4 TW: Do you see that as kind of a threading a needle of… You’re not trying to turn the business into data scientists, certainly. That’s the extreme. You’re not trying to show all the work and, “No you need to understand what the area under the curve really means.” But isn’t there a need to say, a need to through analogy or though, I also don’t wanna over simplify it and just get to, “Here’s the answer.”? Because we’re dealing in data science with uncertainty.

0:26:04.0 TW: Do you see that as kind of like that’s trying to find the sweet spot where they have enough of an understanding and an intuition about what’s been done and what, how they can interpret that without… You’re not giving them scissors and saying, “Run.” You’re like, “Here are scissors, this is what they’re for, and these are how they can burn you.”

0:26:27.4 TW: You’re confident to go and walk over there with the scissors, but you’re… I don’t think you’re gonna take off on a mad sprint and get yourself in trouble. Does that… Do you see it that way with the data storytelling and the working with stakeholders, that you’ve gotta give them enough of an understanding but not…

0:26:43.2 DP: You’ve gotta give someone enough of an understanding. So along on your journey, on your communication strategy, you will reach a point where you don’t have to tell the underlying story anymore. Because several layers and levels of data scientists have reviewed it, they’ve seen it, they’ve talked about it, they’ve hashed through it, they understand it.

0:27:02.5 DP: Now you can move to more of just the story, but I do think that it’s the progression of things, right? And then you may come across someone who’s brand new to the whole initiative, like someone new comes in and is running this business and you need to go explain what’s happening. Then in that case you end up saying, “Here’s how we made that decision. And here are the… Think about it as upper limits, lower limits, we ended up here. This is considered a high percentage, a low percentage, we came out here.”

0:27:34.5 DP: So talking to… Still talking to them in very easy to understand ways so that they can understand how a decision was made, obviously without making them feel stupid, but not using all of the sort of traditional jargon that we might use when we’re talking to each other.

0:27:53.1 MH: You know, it’s actually sort of a little bit of a slight change maybe, but I loved your perspective, which is, over the years have you seen data fluency of business leaders start to shift? Have you seen progression? Or how do you feel about that? Yeah, I’d love your perspective.

0:28:11.9 DP: I feel like over the last… So I don’t know if you guys remember this, but in 2012, Thomas Davenport wrote an article that basically said, “Data science is the new sexy.”

0:28:26.7 TW: Sexy.

0:28:27.1 MH: Yeah.

0:28:27.7 DP: And I was like, “Great, ’cause I’ve gone been doing it for 20 years, so thank you Thomas Davenport.”


0:28:34.1 DP: And know I can actually get paid for doing it, so thank you Thomas Davenport. But I think that was the start of a broader conversation that had begun to happen across industries, that was about how we are… The amount of data that we are collecting is ever increasing, and how can we use that data more effectively and impactfully towards the benefit of the consumer? How could we make that?

0:29:03.5 DP: I can remember… First of all, I’m one of those people who I’m on every social media platform. I bought all of my Christmas presents from Instagram and TikTok this year just because I wanted to see if it was possible.

0:29:16.1 MK: Wow.


0:29:16.5 DP: I don’t mind if you track me. Please do. And make sure that when you target me, you target me accurately. That’s me, right? Whereas other people are like, “Oh my God, they’re tracking me.” But I remember when… When I used to go on Facebook or something and I would buy a pair of shoes on Amazon, say last week, and then I’d go on Facebook and Facebook would show me that exact pair of shoes, and I’m like, “Didn’t I just buy these shoes? Like who is behind the scenes working this algorithm? We’re gonna have to have a talk. Come on now.”

0:29:49.4 DP: And now you will see they have gotten smarter and smarter and smarter, not only are they showing me another pair of shoes, they’re showing me like earrings that might look nice with those shoes. So I think the whole system over the years has gotten smarter, which means that we are now collecting petabytes of data.

0:30:09.9 DP: And to the point that was made earlier, all that data is not useful, but people get into their own heads and they’re like, “Oh my God, we have all this data. We should be solving the world’s problems.” But the reality is, a lot of that data, I would venture to say over half of it is noise.

0:30:26.1 DP: It’s noise in the system, just because we’re in this process of collecting everything we know. Like every time we send an email, click on a link, go on to a secure browser, whatever we do, open an app, it’s a data collection point. Some of that will be useful and some of it won’t. And so I think the discourse has changed so that senior leaders now are cognisant of the fact that data is important, but I would venture to say that they’ve moved beyond data is important, to, data is important, but also technology as an overlay is important as well.

0:31:08.8 DP: I feel like there are elements of what we do with products and what we do in technology and how we use data that are coming together now, just across industries in general, in crucial ways to really inform the best use of the data, the pieces that are useful and interesting, how we should run the analytics, what the reporting should look like.

0:31:32.0 DP: Because now we’re talking about we want a quick… We don’t want a deck, we want an automated report that’s pulling through the analytics to give the insights, right? So all of that is sort of coming to fruition in the world today through the eyes of senior leaders who I think have a better picture or a better idea of what that means, of what having a lot of data means, and what some of the pitfalls are but also some of the opportunities.

0:32:01.9 MK: What that makes me think about though, is there is this spectrum, and I saw in one of the presentations you’ve done previously about “crawl, walk, run” etcetera, there is a spectrum of, I guess maturity. And it’s really easy for leaders to get really excited and inspired about the buzzword-y stuff, “AI, machine learning”, etcetera, etcetera, etcetera, and particularly like you said, at the intersection of how it can be built into products and technology.

0:32:35.0 MK: But the bit that I guess is like the core problem that I still struggle with, and particularly as we get more and more data is like, yes you can automate this report, but that doesn’t necessarily help the business understand what they should do next. And I’m like, I just wanna understand from your experience, because you have such an incredible resume, are you still seeing that as a challenge that faces the data community?

0:33:00.9 MK: That particularly with the young people, ’cause the young people coming through, they wanna work on all the sexy stuff, and trying to get them to…

0:33:08.4 DP: Do they wanna work? I wonder if they wanna work.


0:33:16.4 MH: Zoomers getting called out.

0:33:20.8 MK: But they wanna do sexy stuff. I interviewed someone that was basically like, “I don’t wanna have to do any reporting or like pulling insights or like a monthly report,” and I’m like, “That’s part of the job.” Part of the job is letting people know how we’re tracking against performance and that sort of stuff.

0:33:37.1 MK: Are you seeing this still as like a core fundamental problem? Because for me, I just see the more data we get, the less good we do at those earlier steps and the more we try and run towards the advanced methods.

0:33:52.7 DP: For me, a lot of those advanced methods are really just buzzwords, like you said. Everybody wants to feel like they’re moving into AI, some people are and some people aren’t, and that’s okay. I think that really thinking about what that ecosystem of analytics, data, product needs to look like. And then figuring out, is this the right level of… The right level of detail, or the right level of connectivity.

0:34:35.0 DP: For example, you can crawl, walk and run, at the same time. There will be some projects that you will be crawling on, some where you’ll be walking, and somewhere you’ll be running, and that’s okay. And even when I think… So I hire people all the time, by the way, for who I want to do AI and data science. I want them to do deep learning, I want them to do neural networks. But you know what one of the primary questions is that I ask them? “How well can you move around in Excel?”

0:35:08.0 TW: Oooooh…

0:35:11.5 DP: Drop the mic on that one.

0:35:13.1 MK: That’s a goodie but an oldie.

0:35:15.5 DP: Right? I mean, to me, you could be tool-agnostic as far as I’m concerned. The question is, can you be a critical thinker through data? Can you use a tool, take Excel, because it’s a very well-known, well-used tool that even if before you become a data scientist, you will know. You should know how to use it, right?

0:35:35.9 DP: Can you manipulate, wrangle, move around in the data to answer questions? Can you organise the data in ways that will help people to understand what you’re trying to say and do? So for me, when people are like, “You’re hiring a senior data scientist with a PhD from MIT, you’re asking them to do Excel?” Yup, yes I am.

0:35:56.4 DP: Yes I am. Because that gives me information about who they are, what they can do. Some people, believe it or not, are stuck in a programming language, and if you ask them to move outside of that programming language, and I mean, both physically, how they program, but also conceptually, because you think in a certain way when you’re trained on a programming language. It is very difficult for people to do that.

0:36:23.4 DP: So in essence, it’s very important that we are able to be kind of fluid in both the way that we use the data and understanding the technology and helping people to understand what it entails. Once you’ve done… You’re working in Excel and you know you can do reporting in Excel, there’s obviously a next level of reporting and next level that you could be doing.

0:36:48.2 DP: But I really like people, my personal preference for people who I work with are people who can roll their sleeves up, get really into the data, understand it, get out of the system, move into another system. But whether they’re training somebody on what they need them to do in another system or using a specific different system, they know, they are able to critically think through data and they know that, so that’s generally what is most important to me. Yeah, all of that goes together, by the way. Data, analytics, reporting. It’s all part of the job.

0:37:25.1 TW: I will absolutely judge the crap out of a media agency when they send a spreadsheet and it’s like this data is horribly structured. ‘Cause to me, I love the same question, like you ask around Excel and what do you think of VLOOKUP and pivot tables, and if somebody’s like, “Oh, they scare me,” I’m like, “If you understand VLOOKUP, I don’t care if you don’t know SQL yet, ’cause that’s just a joint.”

0:37:53.9 TW: The moving between the languages and thinking about data and watching people do horribly risky things in Excel just because they’ve structured in a way that’s grossly inefficient and it’s not gonna have traceability. Okay, topic for a whole other podcast.

0:38:09.7 MK: Yes. [chuckle]

0:38:12.0 DP: I agree though.

0:38:14.3 MH: No, I love that though.

0:38:16.2 TW: I have a burning question that’s a total shift of gears that… And it goes a lot back to the Thomas Davenport and competing on analytics and driving all this excitement and executives falling in love with the ideas of data, and then you said sometimes people just wanna be… They wanna be, “What’s the business goal for this project? To do AI.” You’re like, “Well, that’s terrible. That’s not the goal.”

0:38:39.6 TW: But again, in a large enterprise, when something gets the layer of somebody had what could be a really cool idea to try, and data is involved, sometimes my experience mostly being on the outside of it is that ball starts rolling and because of the time and the inertia to plan for that, if it gets to a point where it’s like, “No, this actually doesn’t work. We thought that we could use this, this and this for personalisation and it would drive upsells on this other stuff.”

0:39:14.8 TW: Do you run into the, that it’s hard to stop trains? Like it’s got the label of fancy data stuff on it, and it’s got a lot of people who’ve had to be involved, and it’s got regulatory approval, but it’s not actually working, and it’s really hard to say, “Nobody’s done anything wrong. This just isn’t actually a great and cool idea.” Stopping those trains sometimes seems to be like a huge challenge.

0:39:46.9 DP: Well, that’s the politics, because there’s some person who owns that initiative, like “owns”, who is completely committed to its success. And that is really more of a political minefield that you kinda have to maneuver. As I started, when I started this, I pretty much go in saying to people, “My goal is to be transparent and objective. I am the bearer of truth, and I am sorry if my truth does not align with what you have built or what you’re trying to do.”

0:40:25.1 DP: And by the way, you can take my truth and say, “Screw you, Tiffany. I’m throwing, flushing that down the toilet.” But I will have done my part, which is to tell you my truth about this particular initiative. So yes, I do think that people get really caught up in its success and are more likely to have confirmation bias, which is like only paying attention to the things that kind of support its success, but there’s nothing really we can do other than state the truth.

0:40:56.1 DP: Like, “Here are the objective facts, it’s not gonna work.” Doing pilots to help people see that you thought that it was gonna turn out one way, it turned out another way, probably indicative of an opportunity for you to roll back and not actually execute against the full initiative. So I think that’s pretty prevalent.

0:41:13.6 DP: But that is a critical use of data, I think if people stop… I find that people are constantly, constantly spinning stories with data. Like I’ll go into a situation and they’ll say, “Oh my God, this thing is running so well. This program is working.” And then I’ll look at it and I’ll say, “Wait a second, this doesn’t suggest that it’s running so well. This suggests that it’s doing mediocre at best.”

0:41:42.9 DP: And so having someone who can really come in and interrogate the data in that way and offer a different perspective, I think is very important, but also empowering data people to not be politicised, to be the bearers of truth, to state what’s objective and be transparent and just call it like it is, and have them not worry about losing their jobs. Because that’s the world that we’re in. So for me, that’s the…

0:42:10.6 DP: Look, okay, I’m not young anymore, I’ve been at this a long time. I feel very comfortable in my spot saying, “I’m the bearer of truth.” Some young kid may not feel that way. But I feel like the onus is on us to instill that level of empowerment, to empower them to hold hands with the data and to let the data drive the outcome.

0:42:33.0 TW: But you also said, you earlier were referring to expand and put words in your mouth, you’re not trying to tell somebody they’re stupid, you’re not trying to make someone look bad. That’s kind of the nuance of like, “It’s the truth. I’m not here… Let’s figure out how we can do the right thing without throwing some under the bus,” ’cause now it’s not really productive, and chances are they were trying to do the best that they could.

0:43:02.7 DP: Yeah. But you know what? Back to the political statement, you have to really understand, especially in these large organisations, how to work the politics. So that means that if there’s a big meeting with everyone, when you’re talking about this research and you’ve realised that it’s not what you thought it was, you don’t use that meeting as an opportunity to say, “This is not what it’s supposed to be, it’s wrong, it’s… ”

0:43:28.4 DP: For every initiative, believe me, there’s gonna be one or two uber senior leaders who are driving that thing forward, especially if it’s expensive, right? You simply put a meeting on the calendar with them, they’re smart, sit down one on one, walk them through what you found, learned, that you now understand.

0:43:47.9 DP: And I find that nine times out of 10, they come around themselves. They start asking people to pull back, they start asking people to investigate a little more, and before you know it, some of those big initiatives actually kinda die on the vine. But then it means that in addition to data scientists, data engineers, AI leader, whatever, now you have to be like a political strategist.


0:44:11.9 DP: That’s what it is.

0:44:15.0 TW: Yeah. I love that.

0:44:15.1 MH: Okay, we do, we have to start wrapping up.

0:44:16.7 MK: No.

0:44:17.9 MH: No, I’m sorry.

0:44:18.8 MK: No.


0:44:20.7 MK: I have been… I’ve had this question since the very start of the episode.

0:44:23.5 MH: Well, you can blame him for asking the last question, which normally Moe get’s to ask. I’m sorry.

0:44:29.8 TW: I feel like you can’t break precedent, you gotta let Moe ask the question.

0:44:33.2 MH: Okay, fine. Moe, as per tradition.

0:44:35.7 MK: So at the start, Tiffany, you were talking a lot about seeing the same problems across multiple different businesses, both within companies you worked with and clients and that sort of thing, and perhaps you’ve caught me on a particularly cynical day, because there was part of me that was listening to you being like, “How are you still this excited about it when you do see the same problem over and over and over?”


0:45:02.4 MK: And the same is true of the solutions. When you start talking about some of the exciting things you can do with data, you do look around and competitors are doing similar things, and everyone in the market is kind of going in the same direction. For you personally, how have you kept so excited and interested when sometimes you are dealing with that same problem or the same technology over and over?

0:45:28.5 DP: Yeah. So I think for me from a personal perspective in my career development, I’ve moved beyond, “What does the data say? Is the technology right? Can we do this?” And I moved more, I’ve morphed myself more into a, “How can I tell this story?” And so now, when I hire people, for example, I hired what I call “data journalists”. Because now I’m looking for something different because I personally am doing something different.

0:45:56.5 DP: It’s not about… It’s like John doesn’t believe in that initiative and we have to convince him or whatever, or that thing is running off the rails and no one understands and they have to stop it. And for me, it’s not like, “Oh, well, we have data to support that.” It becomes like, “How do I tell an interesting story to John, to help him understand that this thing is working, or this thing is not working, or this analysis is meaningful, or this analysis isn’t meaningful?”

0:46:26.4 DP: It really becomes more of a storytelling endeavour. And that’s what really excites me. It’s less… To me now it’s less about the data, it’s more about when I bring all of that stuff together, with the personalities, with the goals of the broader firm, how do I tell a story to get this thing started or to stop it if I need to, if it’s something that we need to stop?

0:46:48.9 DP: So it has morphed for me personally, into a different type of endeavour that where obviously data technology systems are a critical part, but it’s around the ability to engage partners, colleagues, other senior leaders towards my goal. So if I really feel like this is something great and no one sees it, how do I galvanise and get everyone and use data and tell the story to get everyone on board? So it’s that kind of exercise for me now.

0:47:26.4 TW: Oh, that’s great. Love it.

0:47:28.5 MH: Alright, now we do need to wrap up. What a great conversation though. A couple of things that really stood out to me from some of the things you spoke about, one was how you talked about using regulation. And a lot of times we think of that as like, “Ugh, that sucks. That’s regulation, it slows us down.” But to actually make it a benefit by being thoughtful and considering the end user experience, I just thought that was brilliant.

0:47:52.2 MH: I also thought the way you described the non-linear nature of progression in terms of “crawl, walk, run” as something that can happen all at the same time, that’s really something I think that really stood out to me and I really like that point. And then I also, I love that idea, and we talked about a little bit about that comms strategy for every initiative, and I thought that’s brilliant as well, sort of just being fully functional in that.

0:48:17.4 MH: There’s a lot of more stuff, but those were three things that really stood out to me. So thank you so much, Tiffany, for sharing with.

0:48:22.5 DP: Absolutely. This was fun.

0:48:25.3 MH: That was awesome. It was a lot of fun for us too, I think, from the smiles and laughs and nodding we were seeing us do.


0:48:30.4 MK: And interrupting.

0:48:32.2 MH: Yeah, of course. Yeah, I was like, “I wanna ask the next question.”


0:48:37.0 MH: Alright, but that’s part and parcel of this whole thing. But one thing we do love to do is go around and share a last call, something of interest that maybe might be of interest to our audience too. Tiffany, you’re our guest, do you have a last call you’d like to share?

0:48:49.4 DP: In April I will be speaking in Orlando, Florida, at the ANA’s Measurement Conference. I’ll be both moderating the conference and doing my own session, and I think I’m gonna do a session on storytelling. So if anyone is interested, please by all means join.

0:49:05.9 MH: Awesome.

0:49:06.9 TW: Nice.

0:49:07.7 MH: We’ll include a link to that in the show notes so people could go find that on our website. Alright. Moe, what about you? What’s your last call?

0:49:15.8 MK: Okay. It involves several things, because one article spurred me on to another thing, but I got sent this article from The Verge, which is the title is, We Might Be Wrong, But We’re Not Confused, how Tomer Cohen, chief product officer at LinkedIn figures out what works best. Which was a really interesting read.

0:49:37.8 MK: The funniest thing is that someone sent it to me purely for like this one paragraph that he mentions, which is about data strategy, and then that got me onto something that Jeff Weiner, the Executive Chairman at LinkedIn had written about vision to values. And it’s this thing, like anyone that has hung out with me recently knows that I have been painstakingly over months working on our vision and our mission, what our team values are, and all these sorts of things, and I kept having this feeling that none of those things were truly connected.

0:50:10.9 MK: And for the team, it’s like, I don’t understand how our team values that we’re meant to live every day connect to this really long-term lofty vision statement. And Jeff’s work on vision to values, honestly, just like… It made everything click into my brain of what all the different pieces are.

0:50:31.8 MK: So that you can use those everyday values to then work towards your operating model and your strategy and your vision, and it just clicked. So I will share the links to those for anyone that is also panting down the same path as me and needs some inspiration.

0:50:50.6 DP: Thank you.

0:50:53.2 TW: Nice.

0:50:53.7 MH: Nice. Alright Tim. What about you? What’s your last call?

0:50:58.1 TW: So mine is on the data visualisation front, so credit to Drew Giltmiller who, co-worker, former co-worker of mine, pointed me to Chart.Guide, which is… I know there have been chart guides all over the place, but this is like to me, one of the best. The poster, you have to put in your email address, but there’s an interactive chart guide, but the poster is so spot on with the dos and don’ts of visualisation. It’s got a pie chart and then it’s not recommended.


0:51:30.0 TW: I tweeted that, and then I will gotta give credit to one of our loyal listeners, Mark Alves, that spawned a little Twitter thread, which ultimately wound up with The Economist’s visual style guide being shared, which I also hadn’t seen. Which, if I ever am in a position again where I get to do a visual style guide or a data visualisation style guide for an organisation, that will be the model that I absolutely enjoyed going through as well.

0:51:57.2 TW: So I’m still getting value out of Twitter, dang-it, and that was me sneaking in a twofer on the last calls, but kinda briefly. So what about you, Michael?

0:52:10.1 MH: Well, actually mine sort of springs out of Twitter as well, so it’s someone I follow, and I actually may have mentioned him for another reason before, Mike Taylor. He does a lot of stuff with media mix modeling. But he’s been writing a lot of essays about memetics in marketing, and they’re actually kind of fascinating, and I think he’s turning it into a book. But I just found them very interesting.

0:52:31.8 MH: So as marketers think about meaning and things like that, I think about it sometimes. I’m not a marketer, but I do think about that stuff and sort of like that connection of the presence of meaning and those kinds of things. So anyway, it’s fascinating. I think there’s more thought experiments and essays that he’s writing, but I think he is putting it into a book format, but we’ll include that as well, kind of interesting to read some topics.

0:52:54.7 MH: Okay, you’ve probably been listening, and I guarantee if you’re like us and if you listen to the show, you probably are, you’ve probably had a few things come up and you’ve been like, “Oh yeah, that resonates to me too.” We would love to hear from you. And the best way to do that is through the Measure Chat Slack group, which you can join at join.measure.chat. And also on LinkedIn or at our Twitter. So please feel free to reach out, we’d love to hear from you.

0:53:22.8 MH: And of course, no show would be complete without a huge shoutout to our producer, Josh Crowhurst, who does so much behind the scenes to make the show happen. So thank you, Josh, for all you do. Once again, Dr. Tiffany Perkins-Munn, thank you so much for coming on the show. What a delightful conversation and so much wisdom shred. Really appreciate you spending the time with us. Thank you so much.

0:53:49.1 DP: Yup. Again, thank you for having me.

0:53:51.5 MH: Alright. And I know that I speak for both of my co-hosts, Moe and Tim, when I say, no matter what stage or what regulatory environment your company’s in, remember, keep analysing.


0:54:07.1 S1: Thanks for listening. Let’s keep the conversation going with your comments, suggestions and questions on Twitter at @AnalyticsHour, on the web, at analyticshour.io, our LinkedIn group and the Measure Chat Slack group. Music for the podcast by Josh Crowhurst.


0:54:25.1 Charles Barkley: So smart guys want to fit in, so they made up a term called “analytics”. Analytics don’t work.

0:54:30.3 Kamala Harris: I love Venn diagrams. It’s just something about those three circles and the analysis about where there is the intersection, right?


0:54:40.9 MH: Alright, so here we go.

0:54:43.4 TW: Michael, that is the most efficient…

0:54:47.9 MK: You’re on the ball this season.

0:54:49.6 TW: Wow, easy.

0:54:49.5 MH: Now I’m gonna screw it up.

0:54:49.6 TW: You’re jinxing it so bad right now.

0:54:51.4 MH: I’m jinxing it.


0:54:55.7 MH: 2023. We’ve had no technical glitches and we’re ready to go eight minutes in.

0:55:00.8 TW: The day is young. Alright, let’s get… Try this again. Here we go, five, four…


0:55:09.7 DP: So they introduced me to things like Dunkin’ Donuts and I was like, “What do you do after you open a Dunkin’ Donuts?” Do you work in the Dunkin’? I don’t know. Do you just eat donuts all the time?


0:55:21.5 TW: I remember from those commercials from the ’80s, the time to make the donuts, just rolling all the…

0:55:27.4 DP: I mean, it was a great business model, but it seemed like it would be problematic for my health, to be honest. I would be bringing home donuts and coffee to everybody in the house. We’d all have high blood pressure.


0:55:42.6 MH: I love that as a strategy.

0:55:44.0 DP: Yeah, it’s… I find it to be very effective.

0:55:46.9 MK: I just love that.

0:55:49.8 MH: Yeah, it’s very holistic.


0:55:52.0 MH: Oh, sorry.

0:55:53.5 MK: I’m having such a rough day.

0:55:54.8 TW: Hey Josh, enjoy editing the over-talk by the co-hosts.

0:55:58.9 MK: Sorry.

0:56:00.9 MH: Go ahead, Moe. Sorry.

0:56:04.3 MK: It seems… Is my internet shit? Is that the problem?

0:56:11.5 TW: No, no, you’re fine. You’re fine.

0:56:16.3 MH: It’s, I’m the problem, Moe.


0:56:20.4 TW: Rock flag and be a political strategist.


Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Have an Idea for an Upcoming Episode?

Recent Episodes

#243: Being Data-Driven: a Statistical Process Control Perspective with Cedric Chin

#243: Being Data-Driven: a Statistical Process Control Perspective with Cedric Chin

https://media.blubrry.com/the_digital_analytics_power/traffic.libsyn.com/analyticshour/APH_-_Episode_243_-_Being_Data-Driven__a_Statistical_Process_Control_Perspective_with_Cedric_Chin.mp3Podcast: Download | EmbedSubscribe: RSSTweetShareShareEmail0 Shares