#177: Design Thinking, Empathy, and the Analyst with Hilary Parker

What is a system without empathy? What is a show summary without an attempt to overly distill the discussion to the point of sounding like nonsense? On this episode, Hilary Parker (who you may know from the Not So Standard Deviations podcast or elsewhere) joined us to discuss what we can learn from the design process (as in: actual designers) when it comes to analytics and data science. Among other things, that mindset highlights the importance of the analyst empathizing with stakeholders. Tim got very uncomfortable. Michael said he understood Tim’s discomfort.

Articles, Books, and SDEC Talks Mentioned in the Show

Photo by pine watt on Unsplash

Episode Transcript

[music]

0:00:05.7 Announcer: Welcome to the Analytics Power Hour. Analytics topics covered conversationally, and sometimes with explicit language. Here are your hosts, Moe, Michael and Tim.

0:00:21.9 Michael Helbling: Hi Hey everybody, it’s the Analytics Power Hour. This is episode 177. Hi Moe Kiss. How are you doing?

0:00:31.6 Moe Kiss: Hi. I’m pretty good. Yeah.

0:00:33.5 MH: Nice. Or how are you going maybe is a better way. How are you going?

0:00:35.9 MK: How are you going? Yeah.

0:00:38.1 Tim Wilson: Wow, 177 episodes.

0:00:39.6 MH: I know… I just try.

0:00:42.3 TW: Finally, yeah.

0:00:44.1 MH: Tim. Welcome, welcome Tim Wilson.

0:00:45.1 TW: Happy to be here.

0:00:46.4 MH: So these are my two co-hosts, it’s great to have you. I’m Michael Helbling, and we’re getting another episode started. And we’re getting back to… It’s something we’ve talked about a good amount of time, but I’m kind of excited to talk about it the way we’re gonna talk about it today. So let’s jump into it. There’s a lot of discussions about how technical an analyst or data scientist should be, what technical skills they should learn, and they usually lead us into this realm of thinking about what different tools or skills or programming languages and things like that.

0:01:14.3 MH: But you know, maybe we should also be considering mindset, how do we approach problem solving or for the challenges we run into that are both technical and not so technical. Anyway, I think to have this conversation, we wanted to bring in a guest who could help us sort through some of this, someone with a lot of experience working with a lot of different data science teams and solving a lot of different problems across a lot of different things, and we think we found the perfect guest. Hilary Parker is a data scientist. She’s worked at Etsy, Stitch Fix, and also the Biden presidential campaign. She has her PhD from Johns Hopkins University, and is a distinguished podcaster on the Not So Standard Deviations Podcast, and today, she’s our guest. Welcome to the show, Hilary.

0:02:01.7 Hilary Parker: Hello. Thanks for having me.

0:02:03.9 TW: First question, how did you land Roger when I wound up with these guys?

[laughter]

0:02:07.2 MH: Yeah. Tim does like Roger ’cause didn’t you take his R course Tim?

0:02:13.9 TW: That was literally my very first introduction to R ’cause there was a whole group.

0:02:18.8 MK: Same. Yeah. I did it too.

0:02:20.6 TW: Actually, Moe’s sister, Michelle was like, “Oh,” part of the group, it’s like, “Let’s all take Roger’s course at the same time.”

0:02:25.7 MK: Stop it.

0:02:27.3 TW: I was on the second one, I’m like, “Hey, are you guys like, I’m struggling with the second one,” they were like, “Oh yeah, I bailed.” There were like three or four people, and I’m like, “You fuckers. What the hell?”

0:02:38.2 MK: I never knew my sister agreed to do that course. Are you kidding me?

0:02:40.3 TW: Yeah. It That was years ago. But I’m not bitter.

0:02:45.3 MK: Anyway.

0:02:48.6 MH: Anyway, good side diversion. But anyways, Hilary, welcome to the show. Maybe just tell us a little bit about some of your background just to get us started, and then we’ll get a little more into the deeper questions around sort of some of the other things.

0:03:02.3 HP: Yeah, yeah. So like you mentioned, I have my PhD in biostatistics from Hopkins, that was actually the School of Public Health. So they’re in the news all the time these days with COVID, obviously. So I was there before it was cool, I guess. I don’t know if that’s the right way to put it.

[laughter]

0:03:19.7 TW: For anyone who thinks that there’s a long time between when we record and when this comes out, now they can tell if people are like, “What? COVID? Johns Hopkins? Is there data from there?” Then they’ll know.

0:03:32.6 HP: So yeah, I went there five years, that’s where I met Roger, to answer your question, that’s how I got my excellent co-host. And then from there I transitioned into tech, and so I’ve been working in tech since then, so that’s since mid 2013. Yeah, I love the companies I’ve worked at, both Etsy and Stitch Fix. Took this turn and have been working more in the arts and just like a totally different application than medicine, obviously. And then the Biden campaign was obviously totally different, but.

0:04:07.1 TW: It seems like there are a ton of biostatistics, that seems like an area that a lot of people who wind up then being kind of prolific R users, data science educator types… I mean, not formal educators, you actually went full into the digital… It’s funny listening to Not So Standard Deviations, and I was like, “Oh, she literally made it all the way into the world of digital analytics somewhat.” Which was that an incredibly frustrating… What are these people thinking they’re doing with their data, or did you not wind up in some of the cases where… I don’t know.

0:04:52.9 HP: You mean like when I started in tech?

0:04:55.3 TW: Well, yeah. ‘Cause were you starting to work with digital data and were you running into people just literally taking aggregated data and drawing conclusions and thinking there was causality there, or?

0:05:07.1 HP: So no, because I think, especially at Etsy, when I first came in. I seem to join companies like right before they go public. I should say the two companies I’ve worked at. But I think generally, when I join companies, they’re looking to hire to beef up their data science efforts. And so Etsy was a really good first landing place for me because there were some key engineers early on who were… Really wanted for Etsy to be very data-driven and do tons of A/B testing, and they had already an A/B testing platform, and so I think there was so much discussion and wheel-spinning with that, where the engineers had gotten to the max of where they could with these discussions without the formal training in statistics that they were looking to bring someone on who could help everyone understand what was going on and how to make decisions in those context.

0:06:07.8 HP: So, I was very much welcomed with open arms, and so I was pretty… I kind of thought it was how everyone operated, but I landed in a place that was pretty sophisticated, hence trying to hire me. Not to be like, “Oh well. Obviously, [laughter] only sophisticated places would hire.” But you get the idea. I wasn’t coming in as the first data scientist. I was coming into fairly mature operations. And I would say though that it’s not like it hasn’t been a frustrating experience, just in different ways than what you’re suggesting.

[laughter]

0:06:44.5 MK: But it sounds like a positive relationship with stakeholders is something that you’ve talked about quite a lot previously, and…

0:06:53.1 HP: Yes.

0:06:54.4 MK: In your experience seems to be that you’ve chosen the companies that are ready. Do you kind of feel like there is a lot of frustration for people that work with data when you go into a company that’s not ready, do you think maybe we should all just pick our battles and make sure we’re going to a company that is kind of ready to have those conversations? Is that the best way to go?

0:07:17.3 HP: Well… [chuckle] I mean, someone needs to do it. This is actually why I am not worried… I don’t know if worry is the right word, but the fact that people are now getting MBAs in data science, and there’s just gonna be hoards of people going into the workforce with this training, I’m like, “Hurray, go to those companies that need to have their first data effort.” Like, I would very gladly let you guys set up the infrastructure and everything. And then, I can come in when the job gets fun, which is for… I don’t really mean that, but [chuckle] it’s… So I think it just… It really depends on your personality, right? And what you want… I think there’s probably… If you were to go into a company which has just never used data in that way, not sophisticated, then you would be the VP of data by the end of your effort.

0:08:07.1 HP: And so, if that’s your career goal, I think that’s a really smart way to go about it. Especially if… The person who hired me at Etsy didn’t have a PhD in statistics, she’s a brilliant leader, she’s now CTO of the DNC, but she didn’t have what maybe would be thought of as the needed credentialing at, even some place like Stitch Fix there was is clearly like hiring PhDs was a priority, and so, I think that there’s lots of opportunity in that case. But is it the type of job that would bring me joy? No, [chuckle] unless I was super passionate about the product and I’m starting to realize as I get further into my career. I think the circumstances where I’d wanna do that is if it were my company or something that I had a huge amount of ownership and then I was so invested in and like, “Oh, I really wanna do this stuff and I need to build it up”, versus being hired to come in and shape up an operation, so yeah.

0:09:08.2 TW: But you seem to like to actually keep your hands pretty dirty while also trying to figure out what is the… What are the opportunities for process or systemic change, is that a…

0:09:24.0 HP: Yeah.

0:09:25.3 TW: You’ve done really cool projects where your hands were, you were definitely doing it, but then a lot of your musings and even the opinionated analysis paper that I’m sure we’ll chat about here a bit, were around what are things that are kind of wrong, which to me is kind of different from, “I’m gonna teach an organization how to use data effectively.”

0:09:49.4 HP: For sure.

0:09:50.6 TW: I don’t know.

0:09:50.7 HP: Yeah, yeah, I would say that I don’t think you can really be, like I hate this term, but be like a “thought leader” without doing the work yourself to some degree, it’s just like you won’t come up with the right ideas if you’re just so zoomed out that you’re only watching these organizations work. And then, yeah, I derive a lot of joy from building things. So the system building stuff that I’ve talked about, both in the DevOps sense of how do you… Like the opinionated analysis development paper you’re suggesting. And then also in the more product sense of like, “Hey, here’s how we could build a product that collects the right data, and that leverages that data in a way that’s really interesting.” And so, I enjoy that all a lot, and yeah, I also am just the type… Like the way my brain works, the way I’m psychologically organized, however you wanna put it, I do just like having things, making things clean and organized. So I think [chuckle] that’s probably something most data people like, but setting up databases with the right schemas and making sure we’re recording everything and have everything set up to work really well later on is just fun for me, so yeah.

[music]

0:11:09.4 MH: Alright, let’s step aside for a minute for a word about our sponsor, Observe Point. Just like we strive to have confidence in the quality of this podcast, Observe Point gives data professionals confidence in their data and insights. Isn’t that right, Moe?

0:11:24.9 MK: Yeah, that’s right. So we’re still looking for a platform that can make sure this show is high quality, but Observe Point has thankfully got that problem solved for website data collection. So their platform really enables teams to automatically order the tags on the site to make sure they’re in place and capturing things correctly.

0:11:44.2 TW: There’s nothing quite as soul-crushing for an analyst as the dive into the a user pathing analysis only to discover that the tagging was dropped from the shopping cart for the past two weeks and no one noticed.

0:11:55.6 MK: So with Observe Point, once the platform is set up to check the site, you can rest easy as a data analyst, knowing you’ll get alerted immediately if something goes wrong.

0:12:04.5 TW: And the results of each automated audit is recorded in the platform, so the team can track whether their data quality processes are improving or degrading over time. It’s time series data about the collection of data. [chuckle]

0:12:19.0 MH: Is that meta or what? Obviously, you’d love to learn more, so go ahead and request a demonstration over at observepoint.com/analyticspowerhour, there you can learn all about Observe Point’s full data governance capabilities. Alright, let’s get back to the show.

0:12:36.8 MK: So okay, we keep… I’m so interested by you and your career, but [laughter] we are here, we are here to talk about this topic of design thinking. And it was really interesting listening to you previously discuss this and how sometimes you think like the word data can almost just be interchanged. So do you wanna give a bit of an overview, I guess, about what you mean by design thinking and how it applies to the space of data?

0:13:02.2 HP: Yeah, for sure. I mean, so this, the idea of design thinking kind of represented a breakthrough that Roger and I had early on in our podcast, Not So Standard Deviations, which was… So the first episode of that podcast was Roger kind of posing this… He was the one who had the idea for the podcast and he was kinda like, “I have this question to start us off.” And it he was, essentially, how do you tell people what to do? So much of the statistics training is around what could go wrong with these various things. And I guess he had a student in one of his Intro classes come up and be like, “I feel like I know what not to do or the errors I can make, but I don’t actually know how to build an analysis.”

0:13:49.3 HP: And he was like, “I went back to my desk and sat down and tried to like, ‘Let me figure this one out.’” And essentially couldn’t. And then we started talking about it, and it’s like, “Oh, this is… This isn’t doable.” This isn’t just something you can figure out and teach in an Intro class and have that be that. And so we were circling around that topic for a long time. And what does it even mean for an analysis to be successful or correct? Or however you wanna define… Even talking about it being done is not actually something you can… I think it’s the type of thing that you probably think like, “Oh, I get it.” But if you actually started to talk about it and started to define it, it would get difficult fast.

0:14:35.1 HP: We sort of had this breakthrough. And just by coincidence, I had done a design sprint, which is the this sort of specific implementation of design thinking with someone at Stitch Fix. And was like, “Wow. This really is scratching an itch.” And then I started to read books about it and was flipping out. I was just like, “Oh my gosh. This is it. This is what that missing… How to do it right. This is how we can define it and do this.” And so the concept of design thinking is… Essentially, the people who’ve created this field or… I don’t know. Professional designers are like, “Oh, the way that this… This is a different type of intelligence than scientific thinking or artistic thinking.”

0:15:20.8 HP: It’s kind of this third thing. And it’s this much more constructive state of mind where the way that architects work, the way that web designers work, you’re actually building something functional. So it’s not just emotional expression like art, but it’s this kind of functional building of something. And so, basically, when I say design thinking for data science or this concept, it’s the idea that you’re actually in this constructive state where you’re building out an… Usually building out some sort of argument for some sort of decision. Or you could be building a recommender system. There’s more product-focused applications of this too.

0:16:01.9 HP: But when designers work they know that someone gives them the design brief, and then they have to work with that client. And they have a lot of… I wouldn’t call it process. But there’s a lot of… I mean yeah, process around like, “Okay. I’m gonna give this client three options. We’re gonna talk through it. I know that when I get the design brief they’re gonna… ” It’ll be like a jumble of words that don’t make sense. And we actually need to parse what that is in order to deliver what they want. They’ll ask for something, but they actually want something totally different. And so I think that that’s where I’m like you could just interchange these words. ‘Cause I think if you say a data analyst instead of a designer it’s the same thing where you’ll get a request for a specific number and it’s like, “Okay. Let me use zoom out. What are you actually trying to solve? Let’s talk through. Let me understand your context. Are you comfortable with spreadsheets or do you want graphs or… ” Those are simplistic, but it’s like how can you tailor this information and guidance on decision-making to this person so that they accept it and they’re able to use it.

0:17:11.3 HP: So that’s what I… That’s a very long-winded answer to your question. But when I think about design thinking, that’s what I mean. It’s like there’s obviously a lot of technical aspects with how you do analysis and the statistical tests, the properties of those. But actually pulling that together into a useful argument, or deliverable, or whatever you wanna call it is kind of a totally different skill set. A totally different mindset and type of intelligence, many would argue, than understanding the asymptotics of some sort of statistical test.

0:17:45.6 TW: So is part of the design thinking around… I’ve noticed there can be a tendency for analysts and even for stakeholders to just kind of… There may be five paths forward but man, they hit the first one and they just start following it. And is part of the design thinking the stop and think? Like frame the problem. Frame the different avenues of pursuit. Think that through. Is there a slowing down part that is trying to put some sort of scaffolding. Saying, “We’re not gonna go answer the marketing question. That’s way too big. And we’re not gonna go back and just give you your one number.” Is that part of it? That’s trying…

0:18:32.2 HP: Yeah.

0:18:33.0 TW: Trying to kind of think strategically? Is that a fair way to…

0:18:36.1 HP: Yes. And actually this was… That specific kind of aspect of it that you’re describing. It was that moment of like, “Oh my gosh. Finally this vocabulary, this language, this… I don’t know. Paradigm that describes this process.” So the idea that the designer’s work is framing the problem. Like that is part of what they’re bringing to the table, is digesting the words that are coming at them, the request, and framing the problem. And then usually once you frame the problem correctly, the solution is very… It flows very obviously from that. And so again, I don’t think… The reason why I’m like, “Oh. Design” isn’t because there’s something special about those applications. It’s just that they’re 10 to 15 years ahead of us. So that’s already taught in school. And that’s this very accepted part of the vocation. And designers talk to each other about it. And if you talk to designers, they have the same frustration around their briefs that they get that we have around getting asked for a number. Right? But when you…

0:19:50.7 HP: There’s an author, Nigel Cross, who we read one of his books for the Not So Standard Deviations Podcast, and he has a lot… If you’re interested in this from an academic standpoint, a lot of his papers are talking through the establishing this as an academic field and as a fundamental type of intelligence with the explicit goal of then being able to get it taught at, in grade schools because it’s like, “Oh, you take your science classes, your art classes, and you could… ” If you frame it that way and establish it like that, then it’s not vocational, it’s actually this kind of fundamental intelligence thing, it’s a very education-focused.

0:20:34.1 TW: Wait, so that’s who’s gonna bump the, no, no, no, statistics and thinking probabilistically needs to get taught in grade school, and that’s the concept, it’s gonna screw us over on that, ’cause they’re gonna win out?

[laughter]

0:20:48.1 HP: Well, I think it would help us, I guess, I’m not sure, ’cause I think that divorcing it from math means that you’re not necessarily competing with calculus anymore for like, “Hey, we only have this amount of time to teach students probability.” And instead it’s like, “Hey, here’s a way of constructing art.” Like the same way you learn constructing essays, those five paragraph essay formats, you could see that happening with analysis, where it’s like, “Okay, here’s how you construct an argument using numbers.”

0:21:21.6 TW: Wait a minute. So I was peeved when my son got told when he was getting into college that the year that he took statistics, ’cause some other class was… And he was like, “Oh, you didn’t take a math that semester in high school.

0:21:29.9 HP: Oh, funny. Yeah.

0:21:31.4 TW: So I was like, oh, so I’ve been on the kick of like, “Oh.” I was annoyed at the time, and now I’ve come to be very excited that math is deterministic and statistics is probabilistic, which is probably an oversimplification, but that’s been my little flight of fancy on the…

0:21:50.9 HP: That’s just weird. Like someone, I don’t know, that seems like a weird college thing.

0:21:56.0 TW: It was weird, but then he like… I mean, he’s getting his CS and math degree, and so it worked out, and the reason he couldn’t take… Yeah, that’s a whole other side.

0:22:06.3 MK: Okay, Tim you’re killing me.

0:22:07.3 TW: Sorry, carry on. I’m killing you, carry on.

0:22:09.2 MK: There is a way too much stuff to talk about.

0:22:12.8 TW: Okay.

0:22:14.1 MK: Okay. So Hilary, you and Roger started talking about this years ago, and I suppose the thing… Well, context first in Australia, everyone who works in data is obsessed with Stitch Fix. We all follow the blog, it’s like the pinnacle whenever we talk about what we want our data to look and work like, and I’m sure on the inside it’s always different. But as you have evolved, I guess your understanding and you’re thinking around this concept, it seems like it’s a skill, it is teachable, but it’s like a skill that some people would naturally be better at. How would you evolve this in a practical sense in the workplace? How are you finding people that are good at this and helping them evolve or flex this skill?

0:23:03.5 HP: Yeah. Well, actually, this is part of why I care so much about framing it this way because… Again, in a design world, there’s these parallels to the way that we think about it where there’s kind of like, “Oh, that person’s magically good at communicating with stakeholder.” There’s just kinda like, “That’s a special person, they’re good at this, most people aren’t. Let them do their special thing, let’s try to hire more people.” And in design you get kind of this like, that’s an amazing designer. He just comes up with these gold ideas, so mystical, it’s so mythical. But actually when you study it as a type of intelligence, you can start to understand, okay, here’s how to build up the skill. It is a skillset and you can build it and you can change the neural pathways you have by exercising them and solidifying them. I don’t know all the right terminology, but it’s like your brain isn’t just set a certain way, there’s neuroplasticity, where you can… By thinking in a design sense, you’ll establish those pathways so the easier you’ll be able to do it more quickly.

0:24:09.6 HP: So again, parallel to what we’re talking about in the design world, a lot of Nigel Cross’s writing is about, here’s how to train people this way. This isn’t mystical, you can train people, here’s some evident… He had a really good example of cab drivers. I actually can’t remember this if it was him or a different book, but cab drivers in New York, their brains that do spatial reasoning are really huge, ’cause they can navigate through the street. So it’s just like that’s a sign that you can learn, this isn’t just set in stone.

0:24:41.6 HP: And so I think that, to your question of, how does this implement in the workplace, I’d say the reason why I like talking about this is ’cause by starting to adopt this myself, I saw myself become a better data scientist and a better analyst, and just having the framework for me understanding in a framework, goes so far. I just feel safe and I know what to do and what to say. And so I want other people to feel that too, and it’s just kind of like, “Hey guys, guess what. There’s this super helpful way to frame this that will help you do your work better.” But in terms of practically speaking, I mean, this is I think probably still in the realm of a somewhat radical idea, so I can’t… I can’t sit here and say, “And then all my co-workers at every place I work adopted this, and we’re all better for it.” I think in some ways, the reason I talk about the podcast and give talks about it is ’cause you have to find people who are open to this, and so… Yeah.

0:25:40.2 TW: It feels like there is… There’s sort of two thresholds, the first one is that somebody has to actually recognize that there’s something there. I mean, a super simplistic example is data visualization. It matters, and I feel like there are analysts who are gonna dig their heels in and say, “The data is accurately represented on this chart. I’m not a designer, I don’t have design talent, it makes sense to me. So no, this data visualization stuff doesn’t matter.” And if you can’t get them to recognize that, “Oh, when those other analysts show their stuff to stakeholders, they have much better… ” They have to first realize that there is value, and then once they realize that, then it’s like, “Okay, can I learn it and can I teach it?” And I feel like that may be in every profession, but I think there’s a challenge for… If an analyst doesn’t say it is problematic, and I will say this is, I get irritated when somebody says, “Oh, Tim, you’re just good at that thing.” And therefore, they don’t even try, and I’m like, “Well, no, it’s because I realized it was important and I practised it and I sucked at it for a long time, but I read some stuff and I tried to get better.” So it does feel like…

0:27:00.7 MK: Tim, I do love how I have had that exact same example, and somehow when you describe it it’s like this negative thing and like, “Oh, Mike, your team is being lazy,” whereas normally someone says to me, “Hey Moe, you’re good at this thing, can you help me learn how to do it?” And I’m like, “Oh, I don’t know how to do that. I feel like it comes naturally to me, but let me try and help you learn it.” Whereas Tim you’re like, “They’re just lazy.”

[laughter]

0:27:27.6 MH: Not just anybody is the quintessential analyst here, so…

0:27:33.5 TW: Oh, come on. Really? Somehow…

0:27:33.6 MH: Somehow.

0:27:39.1 MK: Oh man. Good work.

0:27:39.9 TW: It’s more of the ones who have run into… I had… Years ago in my career, I had… Back when I would agree to manage a team. So I didn’t know how much that would suck the life out of me. I had a guy say… Basically argue with me and be like, “No, no, no, this is how this… ” And I’m like like, “Wow, like you’re not… You’re not seeing how wrong you are, I’ve gotta get you to understand, and I can try to explain it if you’re not open to it,” which goes back, if you’re not open to, this might be better, and really, you wanna be open to at least trying it a little bit, ’cause it sounds like with you. He’s like, “I won’t be good at it, I’ve read some books, I tried to apply it, it was gonna be clunky, but… Oh, it was a net positive. Let me learn more.” But that’s scary, I think to people.

0:28:29.3 HP: I have lots of thoughts here, which is that also… Okay, my first thought is that this is again… Not to just be a broken record here, but this is again, what I see the design world doing that we don’t where it’s like, you won’t find a designer who’s user testing doesn’t matter, you’re just kinda taught that early on, and that’s seen as an important part of the job, and so if we were doing that with analysis where it’s like, “Hey, when you’re in an Intro to statistics class or an applied statistics class, I should say. You have to get user feedback, here’s the form that you’re gonna use.” Even if it’s just a terrible version of it, people would then not necessarily quit. That would be the default, and they would have to proactively go against the default to be like, “No, this doesn’t actually matter.”

0:29:20.3 HP: But then to your deeper point of, it’s really… That really comes down to empathy, and then this is where I get very San Francisco and just saying, I think a big evolution in my life is that I think people, especially in the western world, empathy is seen as this character trait, it’s like you’re born with empathy or not, or this is a set thing, but I feel strongly and I feel like I have evidence that… I believe there’s ample evidence in the world that this is actually a capacity that you can expand through contemplative practice, or other means. For me, it’s been a lot of meditation, and so people are… It’s so sad to me, people pigeon hole themselves and they cut off this expansion that’s possible.

0:30:10.3 HP: And so when I give these talks, I usually end with this super hippy thing of, if you wanna… A way to get better at your job is to meditate or to figure out ways to expand your empathy, that is possible, that’s not just a character trait. So many people in tech think they’re robots, or not empathetic. It’s like, “No, this is just like everything else. It’s a skill, you can build these neural networks, so you can like… This can be something that you cultivate and then you will see your work improve.” But talk about hard… You cannot just tell some of that, and they’re like, “Okay, I’ll totally change my mindset, and my ways of my life,” but I share your frustration. Having someone say, “Well, I’m just not that way,” is like just… It’s frustrating and sad to me.

0:30:55.4 TW: But you can feel where they’re coming from though, right?

0:31:00.8 HP: Yeah, 100%.

0:31:01.4 TW: But the empathy. No, that was my… Tie it back around.

0:31:05.4 MH: Yeah. Tim loves when a conversation steers to empathy, don’t you Tim?

0:31:10.9 MK: Sorry, Hilary, how has this epiphany changed your own work practice?

0:31:19.7 HP: Well, so again, I feel like I’ve spent the last few years trying to… So I read another really influential book called Creative Curve that was essentially saying, practically speaking, here is the way that creative people work, like they consume media, some amount of it. When you interview creative geniuses, they’re usually consuming a ton of media. They usually have these creative communities where they’re spitballing ideas back and forth like, “Here are some other practical steps,” and so I think that if you think about what people are open to is like a bell curve where you have at the very two center deviations above, you have people who are like the trendsetters, and they’re always looking for new stuff and they want to… They wanna be said… They’re the street fashion of the world, whatever application, but then most people are kind of in the middle of the bell curve and they just want like tried and true, they just wanna wear the stuff from Stitch Fix.

[laughter]

0:32:22.1 HP: And then you have the laggers, whatever. But you can think about analytics that way, where it’s like, “Oh.” We all probably are the street fashion people who really wanna know the latest methods and the latest whatever, but the people who are consumers are usually in the middle of the curve where they’re used to seeing Excel sheets. They might be open to a little bit of fancy-ness, but you really just can’t push people far from where they are, and so it’s… I think one of the hardest things as an analyst is that you get these ideas for really cutting-edge beautiful methods and visualizations and like, “Oh, I wanna make a shiny dashboard for this or… ” You kinda go crazy with the cutting-edge stuff, and so you have to wrangle yourself back to like, “Wait, but my consumer is not there, almost by definition, otherwise they wouldn’t be hiring me to do this work.” And so one way that this has really tangibly helped me is just recognizing when I’m… Like at that end of the belt, like at two standard deviations above, and then I have to wrangle myself back to the middle even though I don’t necessarily want to. So that’s… I mean, that’s… Again, that’s kind of abstract, but…

0:33:33.6 MK: So why do you have to wrangle yourself back to the middle? Because you feel like you’re losing the audience? Is that…

0:33:39.2 HP: Yeah, exactly. Yeah. So in general… And this is… Yeah, I’m sort of filling in a bunch of… But in that Creative Curve book, for example, you can kind of study how far people can get pushed past their comfort zone of like, “Hey.” Like if you took someone from the 1800s and you played modern music for them, they would be like, “I can’t comprehend. This is incomprehensible to me.” And so you can only kind of push people so far from their current context. And so within analytics, it’s like, “Oh.” If you work with people who are in financial analytics, they’ve got these super massive Excel sheets with 50 different tabs and macro is going all over the place.

0:34:25.3 MK: Yes.

0:34:26.0 HP: And so you couldn’t necessarily give them like a D3 interactive visualization, like the… Like a bunch of very visual. Kinda what Tim was mentioning. Like you couldn’t necessarily get those people to just adopt this radically new thing, but you could put a little bit of visualization in there. Or you need…

0:34:44.4 MK: Oh, my God. I’ve just had an epiphany about why working with finance is so hard.

0:34:50.0 HP: Yeah. [chuckle]

0:34:50.8 MK: Oh, my God.

0:34:55.8 HP: It’s hard for me, too, ’cause it’s so far from our training. It’s like… It’s not even like, “Oh.” You’re a part of the bell curve that I was never at. Like I’d never learn these methods so… Yeah.

0:35:04.5 MH: Guys are laughing so hard.

0:35:06.7 HP: I’m happy. This is like…

0:35:09.6 MK: Oh, my God. Seriously, like the last week, I’ve been like, “Why is finance so hard to work with?” And… [chuckle] Wow. Wow.

0:35:16.4 HP: I don’t why, but this sort of reminds me that a friend of mine, back in the early 2000s, like the first tech boom, I guess, I don’t know, late ’90s, early 2000, he needed to create like a… People were doing their expenses, email in Excel sheets, whatever it was at the time, and so he created this program that was just taking an Excel sheet and making it all white and putting one cell in the middle that was like a button, and it was like click here. And it was all like macros to make it happen. But it made it like a submit-able form, even though it was opening the same program and everything looked this same. So, I don’t know. It’s just like creative solutions like that where it’s like, “Okay, this is mostly the same, a little different but like… ” But it’s incredibly… You have to… People aren’t gonna think that’s fun analytics work because it’s not their playful space. It’s like, “Oh, I wanna be out there doing new D3 stuff, and it’s like, “Well, that’s not the job.”

[laughter]

0:36:19.9 TW: Well, I mean… But you have the stakeholders, and then you have the analyst. And I do get frustrated with… It’s interesting ’cause I think finance, because it is… You’re dealing with FASB and regulations and GAAP and all this stuff, like it is a structured, how am I recognizing revenue, and do I have one type of traceability? And so there is a degree in maybe insulting finance that you get a process and you need your repeatable way of presenting, and then you’ve got financiers trying to do some level of forecasting. I feel like in our world, any analyst that says, “No, I’m… ” If the analyst says, “I’m good here,” they need… If they were saying… If we’re having a debate as to whether they’re comfortable taking a small step forward or a huge leap forward, I’m okay. If they say no… And I can think of some analysts who I knew ’em 10 years ago, and they were building dashboards in Excel, and I was building dashboards in Excel, and I was coming up with clever ways to do better layouts and automation and robustness. And 10 years later, they’re building dashboards in Excel. There are three jobs removed, and they haven’t… They haven’t moved in their value that they’re delivering ’cause to me, the analyst needs to be… You need to be kinda pushing the company forward at the rate that they will take it ’cause if you stay still, you’re gonna move back. That sounds like a… That is a cliche. If you stand still, you’re going backwards.

[laughter]

0:37:58.0 HP: Yeah.

0:38:00.8 TW: I mean, if I look at where I was five years ago saying, “I’m gonna dive into this R thing maybe longer than five years ago,” and… I’m in a completely different… I’ve been in this space for 20 years. I was the guy saying, “I can do what you can do in Tableau. I can do that just as well in Excel.” And I may still, maybe outside some of the connections, so… I don’t know, I think I still lack the patience for people who if they say, “I wanna stay still.” Then that’s problematic, and I feel like that’s the same for stakeholders too, ’cause… Well, quick little side rant, I feel like a challenge in our world is that if you’re at a company that is standing still, people are still gonna leave that company, but move up in their careers and go to the next company and think that that is the norm and the standard, that’s how things…

0:39:00.1 TW: How many times have I had a client who says, “Well, we got the weekly report in an Excel file at my last company, and now I’m at the new company, I’m in a more senior title, but I just wanna do what we did at my last company, get the weekly report in an Excel file.” And then you’ve got junior analysts who say, “That’s what analytics is. I produce the weekly report in an Excel file.” And then they move up and they go to another company and say, “Man, I know how to make the week,” and that’s a problem, that…

0:39:28.2 MK: But again, Tim, you could inverse this and say the exact opposite. So that’s companies moving in the wrong direction, but if you had people, I guess, adapting and learning a little bit and then they take that to their next company is those companies are gonna get pushed forward. Does that make sense?

0:39:44.1 TW: But I think the balance is way… I think it is a lot easier to… It takes a lot more energy to move. When I come into a new company and they say… They may say, “We’re bringing you in with your experience to help us move forward.” But I’ve heard that, I’ve been told by a massive pharma company, I’m sorry if they’re listening right now, but the data science work we were doing, they’re like, “We don’t know anything about data science, teach us,” the experimentation work we were doing, they’re like, “We don’t know about experimentation, teach us.” The analytics work, which was kind of my world, they’re like, “Oh yeah, do the bi-weekly PowerPoint, please. But do grander insights.” And I’m like, “Maybe we need to change the way you’re thinking.” And it takes… Because in their mind, that’s what they’re conditioned to do, it is getting people to think that what they’ve done, there might be a better way is a lot harder, there’s much more of a… Yeah. Moe, I feel like this is hard discussion…

0:40:44.5 MK: That was a massive rant.

0:40:49.1 TW: Yeah, it’s a very… But Moe, you and I had this discussion years ago, because in Australia, we were having a discussion where I was like… Your take was that in many ways, you feel like Australia leapfrogged the US, maybe this was years ago, we had this discussion, the US drove into a big fat Lazy Boy recliner of complacency, like the size of that, the mean, there are lots of companies that have settled at a spot that it is…

0:41:21.6 MK: I do think though, Tim, in the hindsight, that the companies I’ve worked at are probably pretty similar to Hilary in that start-upy, tech companies on the bell curve, we probably do lean towards more advanced data analytics capabilities and mindsets versus maybe if I’d worked at some of the big companies here, maybe I would have had a more similar experience.

0:41:47.5 TW: And that’s having… I’ve worked with big pharma, old insurance companies, financial services. So I think…

0:41:54.8 MH: There’s a lot of different ways to look at it. Certainly the tech angle makes it a little bit specific because you can go very deep in your specific stack and do very complex things, but I work for a retail apparel company and they use Excel for everything, and we were still doing some pretty amazing analytics, so it’s possible to do, but it goes back and forth, and…

0:42:18.9 MH: Well, that’s what you think, but see that’s just where…

0:42:24.5 MH: Yeah, this is what I think, Tim, and I got the numbers to back it up. Thank you very much. Anyway, but I think…

0:42:31.3 TW: That was 10 years ago. It was an amazing thing.

0:42:33.4 MH: Well, I keep doing the same thing with the same companies that I work with today. I just don’t bring them up all the time on the podcast, Tim.

[laughter]

0:42:43.1 MH: It is interesting, I don’t know where I was going with that, except to be like I think there’s a lot of different ways to look at it, Moe, and that it’s tricky to… The industry is too big in America, because there’s actually probably three different industries in the US, they’re sort of like the tech layer, then they’re sort of fast growth and follower, and then there’s sort of big enterprise, and they all kind of do slightly different things, they do move back and forth between each other a lot, but what I’ve noticed is like there’s people in this layer, that I’ve never talked to, ’cause I spent a lot of my time consulting in this layer. And so as I’m learning new spaces. I’m finding out whole communities and groups I’ve never heard of before.

0:43:24.8 TW: So Michael, I’d like to remind you, this is an audio format. So you’re hands gestures are not explaining the layer you’re on.

[laughter]

0:43:32.4 MH: Yeah, up here, down here. I do that a lot. That’s right.

0:43:35.9 HP: I get that.

0:43:36.7 TW: We’ll leave it up to the listeners imagination as to which layers Michael was referring to.

0:43:43.5 MH: Yeah, just picture in your head, the layers can be either way. See, that’s the beauty of it. Sorry.

0:43:48.3 HP: My guess, pure speculation here, but I feel like probably the places where you’ve seen analysis stagnate is highly correlated with how much regulation there is for the application, both in terms of the company size and financial, there’s so many specifics for finances.

0:44:12.5 MH: If you can’t tie your data back to what your job is or what it means, like I did analytics for McDonald’s for a while. McDonald’s does not care about web analytics very much at that point in time. They do care a little, but it’s sort of like, “How does the analytics we’re doing on this website contribute to same store sales year over year?” ‘Cause that’s actually the problem I’d like to solve. And so my work was not very meaningful in that context, ’cause I couldn’t tell people how many burgers we sold.

0:44:42.7 TW: Golden opportunity for randomized controlled trials.

0:44:47.2 MH: And they’ve continued to invest and do all kinds of crazy things now.

0:44:53.1 TW: But I don’t know about the regulation. I think there’s an age of the company, that’s been my observation, financial services is highly regulated but fintech startups, high growth I suspect, even though they will trip and Robin Hood will trip over themselves multiple times, but I suspect that their infrastructure and the things that they’re doing are probably pretty good, same thing with health care for the health care that’s trying to be disruptors, even though they had a deal with HIPAA. So they, yeah, they figure out a way to hash stuff. So that they can do stuff with it.

0:45:32.3 HP: Like Theranos.

[laughter]

0:45:34.2 TW: Well, yeah, okay. Another good example.

0:45:37.3 HP: Well no, but it’s, I guess it’s funny ’cause even though… I feel the need to defend the people you’re talking about who stays stagnant, because I feel like it’s just all like…

0:45:51.8 TW: Michael can defend himself quite well.

[laughter]

0:45:58.2 HP: I’m trying to think of a quintessential, he’s not a quintessential analyst from what I understand.

0:46:04.5 MH: Oh no I’m not, I’m absolutely not, but I help people understand Tim better.

0:46:13.9 HP: Yes, I got it. I guess my thought is, is just there’s gotta be reasons why someone doesn’t feel the need to innovate on that specific thing. And my guess is that… How many… I know, I personally have been like, “Oh, I wanna try this new thing… ” and even in grad school, I remember [chuckle] this is such a specific product, but do you guys remember Google Wave at all?

0:46:43.5 TW: Oh yeah.

0:46:45.4 HP: Yeah, this is like, I guess in 2010, ’11, maybe 2012, it was Google’s answer to what email should be. It’s essentially, it was essentially a version of Google Docs. The big innovation was the interactive… Like if you were typing the other person can see what you’re typing in real time and… So I was like this big cutting edge person, I always wanted to do the latest tech, and I remember trying to get my advisor Jeff Leek to use Google Wave, and he finally had to just be like, “I can’t do this again. You’re asking too much of me to keep trying these niche products that don’t go… ” Google Wave was to sunset within two years or something. It was a flash in the pan. And so it’s like, I guess what I’m saying is that there are environments you find yourself in where the cost of innovating is so high, where it’s like, “Oh man, I’m gonna have to figure out this way to do… Our compliance docs are only acceptable in xsl or whatever that… The new Excel format is. And so even If I was to do a fancy R program about it, it still would have to spit out this xsl, whatever, and my boss is just gonna be confused, they’re not even gonna reward this behavior.” I feel like there’s probably some complex set of incentives.

0:48:10.8 HP: And also just like that person’s personal motivation or what they find passion in might not be being that person who’s two standard deviations above the curve, and that’s a choice to who you’d hire. And what type of behavior to incentivize. I think I’m feeling defensive of like… I totally get why people just stick with the status quo, especially when you’re dealing with arcane regulations. The other… Not to just keep rambling, but the other example I can think of is in, I think in China, the e-payments or using your cellphone to pay for things is like… Everyone does it. It’s super, super common. Whereas here, we’re still on this, it was such a big deal for us to even get to get to chips on our credit cards, and so there’s like that innovators dilemma thing too, where it’s just like, “Oh, right,” ’cause they were starting from scratch, getting everyone ramped up on mobile payments versus here it’s like this big change to this regulated industry.

0:49:13.7 TW: So one more counterpoint on these people…

0:49:17.1 HP: Yes, these people.

[laughter]

0:49:18.1 TW: Well, one, there are people who are still using Excel, but they’re using it horribly inefficiently, and they’re therefore just like pushing and they’re like… That not even driving towards, if you’re going to do this, at least find ways to be… To do that well. So that’s one part, but I think the really big one that is, I think hugely problematic. And I would say it goes to thinking about the system is paid media. And we’ve had various episodes of this show that have come out in different ways, where if you think through the incentive structure for display advertising and social media advertising, and think through the different players in ad tech, in martech, and advertisers and publishers, it is so obviously a system with stacked incentives to not actually deliver value and yet report that you’re delivering value. And that’s another one that you roll into a company and say, “Can we just logically even if you just think through what you think your display media is doing.” If you describe how you think it’s working, it is such a weird ass bank shot to actually get to value. It logically makes no sense.

0:50:37.1 TW: We can talk through the causality of how you’re actually measuring it, it’s all completely… And they’re like, “Nope, we don’t wanna rock the boat because the dam at the pixel fired and the affiliates are getting credit for this purchase.” Like that drives me, absolutely berserk ’cause that is a critical thinking that is a… Wouldn’t you want to spend your company’s money way more wisely? It’s like, “Well, maybe not, ’cause then I could do just what I’m doing with half the budget and my budget would have shrunk and my prestige will have shrunk.”

0:51:13.3 MH: Okay, it’s time for The Conductrics Quiz, the quizzical query that makes us feel really intelligent and we’re representing our listeners today. Let’s talk a little bit about Conductrics. Do you know how A/B testing vendors, a lot of times promise sort of the silver bullet to make every experiment easy, and running an effective program is actually really hard work? You need a technology partner that is not only innovative, but also honest with you about the challenges of building and maintaining successful experimentation programs, and for over a decade, Conductrics has partnered with some of the world’s largest companies to help discover and deliver effective customer experiences. They offer best-in-class technology for A/B testing, contextual bandits, and predictive targeting. And they’re always providing honest feedback and going above and beyond to help clients achieve their testing and experimentation goals. Go check ’em out at conductrics.com, and let’s get the quiz underway. Okay, Moe, do you want to know which of our listeners you are playing for?

0:52:19.0 MK: Sure.

0:52:20.1 MH: Alright, and you have a special celebrity guest.

0:52:23.2 MK: Yes! I finally…

0:52:25.4 MH: Hilary Parker!

0:52:26.8 MK: I’m pleased, the phone a friend finally worked.

0:52:30.5 MH: And so, Hilary, you’re joining Moe’s team, so… Listen, there is no… But if you don’t… If you two don’t win… I don’t know. Anyways, here’s to hoping…

[laughter]

0:52:46.1 MH: I hope you do win, I think it’ll be amazing. Alright. Who you’re representing is Christopher Patterson. Christopher is a listener, and he hails, I think… It doesn’t matter where he comes from. Anyways, Christopher is being represented by Moe and Hilary, very exciting. So Tim, are you ready?

0:53:04.4 TW: I’m ready.

0:53:05.1 MH: This could be your greatest victory ever, Tim. Tim, you are representing James Murphy, also a listener to the show. Alright, here is the question. It’s a fairly long one.

0:53:14.8 TW: I feel like the stakes are very low for me here, so…

0:53:17.0 MH: Feel free to jot down notes or things like that, it goes for a little bit. Okay, picture this scene.

0:53:24.4 TW: Wait a minute. Matt Gershoff provided a long stem-winder of a setup question for this?

0:53:28.8 MH: Yeah.

0:53:29.8 TW: Had to happen eventually. It’s very on-brand.

0:53:33.4 MH: Yeah, exactly. Alright, picture this scene. Moe and Tim are drinking tea and quietly chatting amongst themselves. When without warning, Michael barges into the room with a panicked look on his face. With his phone in one hand and his other hand covering the phone’s receiver, he whispers a bit too loudly, “I’m on the phone with a client who’s asking how to calculate something called the standard error of the estimate. Can either of you help me out?” And you reply, “Sure. The standard error is just… ” This is where we get into multiple choice. A, the standard deviation divided by N, where N is the sample size. B, the mean value divided by the square root of the variance. C, the mean value divided by the state of deviation squared. D, the standard deviation divided by the square root of N. Or E, the square root of the variance divided by N.

0:54:32.8 MK: I stopped taking notes, because I was like, “Clearly, I am not going to be helpful in answering this.” [chuckle]

0:54:41.7 MH: It’s just standard error, Moe!

[laughter]

0:54:48.4 MH: Alright.

0:54:49.1 TW: This is one that…

0:54:49.6 HP: Is there a time component, or should I…

0:54:52.4 MH: There’s really not, and there’s a lotta leeway here, Hilary, so feel free to sort of explore the space. It’s really not really gonna be…

0:55:00.9 HP: I think I know it.

0:55:01.0 MH: We’re not doing the whole Jeopardy song. Oh, she thinks she knows it.

0:55:02.5 HP: I know, I’m like, I know… Now I’m doubting myself, but…

0:55:05.8 MH: Don’t doubt yourself.

0:55:07.4 HP: I believe it’s… I don’t remember what letter it is, but standard deviation divided by square root of N.

0:55:13.1 MH: That would be D, standard deviation divided by square root of N, okay. Moe, do you concur?

0:55:19.7 MK: I wholeheartedly concur.

[laughter]

0:55:24.5 MH: Tim, what would your answer be?

0:55:28.2 TW: I mean, I have read these so often, and it’s such a simple thing, and then…

0:55:33.0 MH: If these aren’t super complex, it makes me feel like I shouldn’t have been so panicked on the phone, honestly.

[laughter]

0:55:40.7 HP: I just wanna know the client that wants this, that’s the suspect part.

0:55:45.2 TW: It’s… I mean, Matt’s been known to give us the homework and find shortcuts for things, and… It’s… When do you do… When do you wind up with an N minus one, though? That’s only… Aren’t there some cases where you do… There were no N minus one, so I don’t have to worry about that.

0:56:03.0 MH: That’s true, none of the answers… Do you need me to read any of those back to you, Tim?

0:56:07.0 TW: Well, I had the wrong… I had the square root of the variance divided by the square root of N for D, so apparently, I got the wrong…

0:56:12.3 MH: Oh, that’s E, Square root of the variance divided by N. That’s E. D is the standard deviation divided by the square root of N. And then there’s also…

0:56:23.0 HP: That is square root of the variance divided by the square root of N, right? This is…

0:56:28.9 TW: Oh, so he’s being…

0:56:29.9 MH: Oh…

0:56:31.0 HP: Yeah. Tim, you did extra. Yeah. [laughter]

0:56:33.3 TW: I think that was… He would definitely be the sorta guy who would give two correct answers. Yeah, ’cause the standard deviation is… Yeah, so, let me… My simple… Minding that standard deviation… You’re basically looking at the… You’re squaring things, so… ‘Cause you can go above and below, and you’re trying to kinda… That gives you the variance, and then you gotta square root it to kinda get it back to where it’s kind of an absolute value sort of thing. I’m gonna go with the square root of the variance over the square root of N, even though that doesn’t… ‘Cause I think that’s the same thing.

0:57:10.1 MH: The square root of the variance divided by the square root of N? That’s not one of the answers, Tim.

0:57:15.5 TW: What’d you say the last… What was E?

0:57:17.6 MH: E is the square root of the variance divided by N.

0:57:21.2 TW: Oh. Okay. Then that would be… I’m gonna pass.

0:57:29.2 MH: He’s gonna pass, alright.

0:57:30.0 MK: Wow!

0:57:32.5 HP: Square of variance…

0:57:32.8 MK: You’ve never passed before.

0:57:34.1 HP: Yeah.

0:57:34.7 MH: Well here’s the thing, here’s the thing, and we’re just gonna just rip the band right off. Moe and Hilary, you win.

0:57:40.5 TW: Right.

0:57:40.9 MH: And Christopher Patterson, you’re a winner. Tim, you had the right answer and I wouldn’t let you pick it ’cause they had already picked me, but you had worked that out.

0:57:51.1 TW: Right.

0:57:52.3 MH: But I feel like…

0:57:53.0 HP: Square root of variance divided by square root of N is standard deviation divided by square of N.

0:57:56.0 MH: Yeah.

0:57:56.7 TW: Yeah.

0:57:57.7 MH: So you basically have worked out the problem a different way, but got to the same answer but…

0:58:02.4 HP: You actually… Yeah. [chuckle]

0:58:02.7 TW: I mean I had it written down. I had written down E. I was trying to… Yeah.

0:58:04.6 MH: But Hilary was like, “I think I know what this is.” So Hilary and Moe took it home. It’s a matter of… It’s a matter of record. So great job.

0:58:14.8 HP: I literally just have this memorized. It’s like one of the few things that I just have memorized.

0:58:19.3 MH: And that’s… Folks, if you’re listening, that’s what you get from a PhD at Johns Hopkins, this kind of knowledge.

[chuckle]

0:58:26.8 HP: I am seriously glad I got it right, because…

[laughter]

0:58:28.7 MK: I was listening to you the other day talk about the different things that you memorized, and how in some cases it’s worth just memorizing these formulas, and I was like, “Oh, I’m never doing that,” And then here is an example.

0:58:41.4 HP: Memorize the P value sentence, you’ll just be happy you’ve done it. Same with like the definitions.

0:58:44.9 TW: But that was… I remember a little bit of an epiphany, that standard deviation is basically your… It’s the overall… Standard error is when you’re getting into looking, trying to figure out how much, like your sampling variability, right, like that little light bulb goes on a little bit, and I think this is amazing and so cool that based on the size of my sample, I can predict the distribution, even if I don’t know the mean, and there’s kind of an inverse.

0:59:18.3 HP: This is the hardest aha moment.

0:59:18.8 TW: That’s what I wanted to get to the intuition around all this stuff.

0:59:21.6 HP: Yeah.

0:59:22.4 MH: Yeah. Well I mean technically Tim you kinda did.

0:59:23.5 HP: It’s the…

0:59:23.5 MK: If feel like you’re getting there Tim. As opposed to me that still needs Matt Gershoff to give me stats homework. I am…

0:59:30.9 MH: Moe, I also just feel like you don’t give yourself enough credit too.

0:59:36.2 TW: I concur with Michael. It’s happened, I agree with Michael.

0:59:38.7 MH: I mean, in this case, you solved the problem a different way by picking a highly qualified teammate.

0:59:47.2 MK: I did.

0:59:48.2 MH: Look, it doesn’t matter how you get to the finish line. You got there, that’s the important thing.

0:59:56.6 HP: That’s funny.

0:59:56.6 MH: Anyway, alright, let’s wrap up the Conductrics Quiz. Big thanks to Conductrics for putting these crazy questions together. Thank you, Moe and Tim, and especially you Hilary for joining in this week. Big ups to everybody. And let’s get back to the show.

1:00:10.5 HP: Yeah, I got really mad about a website. So does anyone know Poshmark.

1:00:17.1 MH: I’ve heard of them.

1:00:18.4 HP: Yeah, so it’s essentially like eBay for clothes, just for clothes. Where it’s a selling and buying platform. And it’s very structured to be selling clothing. And it makes it easy for the seller. But they incentivize the absolute weirdest user behavior where essentially they want you to… In order for your listings to show up, you have to buy that with clicks. So it’s like, Oh, you have to be interacting with other people and essentially retweeting their listings and stuff. So much so that people employ bots or they get bots to do all these behaviors, so that there listings show up. And so now I’m selling stuff on there right now. So now I have bots. So all the clicks on that website are literally bots talking to each other. And it’s just like, “What… ” I literally was like, “Is this a scam?” Was this all part of like, “Oh, let’s get our daily active users up for our earnings or for… ” I was just like, “This is insane.” You’re not even paying the company, you’re paying a third party bot maker to… So it’s… Anyway, I feel you where it’s just like is anyone in that room thinking about this, ’cause this is like not… So I’m acting like, “Oh, you should empathize with the issue as well, but then I get frustrated too.

1:01:36.3 MH: So you gotta go back to meditation…

1:01:40.2 HP: I know, yeah. And just be like, “I’m empathizing with those executives, who want those down now numbers up for their earnings column.

1:01:47.3 MH: Yeah, it is hard because there is a limit to empathy, it’s weird that we started a conversation about systems thinking and data science, and we’re talking about empathy, but I love it. And unfortunately, we do have to wrap up ’cause I wanna talk about this.

1:02:03.4 TW: When?

1:02:04.6 HP: Oh yeah.

1:02:05.3 MH: I actually would love to do another topic with you at some point, but we’ll just see. Because I think there’s something to this whole meditation thing in analytics, for analytics people. So maybe that’s whole other episode of a podcast. It’s a great topic.

1:02:18.0 HP: Can we also decide that everyone needs to go read Hilary’s paper on opinionated analysis development, because I was dying to talk about that and we didn’t get there. And it’s got a lot of topics in there that are near and dear to my heart. So it’s worth a read.

1:02:31.7 TW: Can I read the quote that made me fall out of my chair.

1:02:35.6 MH: Yeah, do it.

1:02:36.2 HP: Yes.

1:02:36.9 TW: Well… The point when you said when an error in an analysis occurs, it is safe to assume aside from nefarious actors that the analyst did not want that error to occur. It could have been that she thought she was producing an analysis free from errors. You must look at the way she developed the analysis to understand why the error occurred and create safeguards, so that the error does not occur again. I was like, “Oh, that’s the case for… ” I was like, Oh yeah, that sort of blew me away.

1:03:09.7 HP: Well, thank you, I’m glad I was able to transfer the way I was blown away from this concept to other people. Because that actually was something that came up. I learned about this sort of framework of error the… I can’t remember, there’s some formal word for this, but how errors… How you frame errors from my work at Etsy, and that was a big thing within the culture there, and there was this framework from this book, the title, I can’t remember, but it was this whole idea of like, “Oh.” I had the same epiphany moment that you did where it was just like, “Oh my gosh,” I understand why it’s easy to blame a person, but if you switch your framework, you can actually get to these productive conversations and to tie it to the empathy thing, I think that in some ways, all these frameworks are just like forcing kids to play nice with each other. It’s like, “Oh, let’s enforce this brainless mindset and like tell you, what to say and how to say it,” and let’s do design sprints, where people are forced to listen to each other, and that’s like going through the motions of, “Hey, let’s make sure you empathy for people, and if you can’t, if you can’t make it you gotta fake it. Right.

[chuckle]
[chuckle]

1:04:24.2 TW: Well, assume good intent.

1:04:26.1 MH: Yeah.

1:04:26.3 TW: Like that… Earlier on in COVID, my daughter’s principal… And she was kinda rolling her eyes. She’s a teenager, so it happens. And it was like the only thing I’ve ever heard him saying, and he was like, “We’re heading into a really challenging… Just assume, good intent.” And I’m like, “Oh, my God, the number of times I have… ” Michael was trying to undermine me at every possible turn, but everybody else, I now assume, good intent.

1:04:49.6 MH: But with all the best intentions, Tim.

[laughter]

1:04:53.0 HP: Yeah.

1:04:53.2 MH: With all the best intentions.

1:04:54.1 TW: No, I don’t think so. Nefarious.

1:04:56.9 MH: Nefarious.

1:04:57.2 HP: I would probably assume good intention as it’s simple, but it’s hard.

[chuckle]

1:05:02.9 HP: I would say that it is not something that… Hearing someone say that does not change my emotional reaction. Like hearing people say I should do that did not make me do it. I would say it was a lot harder work to get to the point where I’m like, “Let me think about it from their perspective.”

1:05:19.3 MH: Yeah.

1:05:19.6 TW: Well, I assume good intent, and then I’m like, “Yeah, but they’re a fucking moron,” so I’m like, “Yeah, no.”

[laughter]

1:05:24.2 MH: So basically, Tim…

1:05:25.4 TW: Okay, I guess I’m not good. Time to start meditating. [chuckle]

1:05:28.8 MH: What we’re… What we’re learning, Tim, is you don’t have good intent.

[laughter]

1:05:31.8 MH: So it’s sort of like…

1:05:35.3 HP: Yeah, yeah. Exactly.

1:05:35.4 MH: So it doesn’t really matter what other people… All right, let’s…

1:05:37.2 TW: I want other people to assume that, though.

1:05:38.7 HP: Yeah.

1:05:39.7 MH: Well, I’ll assume that we gotta start wrapping up the show. Hilary, what a pleasure, frankly. To have you on. Thank you so much.

1:05:47.9 HP: Yeah, this was super fun. And I’m happy to chat again if you want.

1:05:51.3 MK: Oh, my God. I feel like we could have talked for another five hours so…

1:05:55.8 MH: Yeah. Be careful what you wish for. You might just get a couple…

[laughter]

1:05:58.7 HP: But I feel bad that we didn’t talk about the opinionated analysis stuff ’cause I can go on about that, too. [chuckle]

1:06:06.4 MH: Wow, there you go. Well, it sounds like maybe the next time is in store already. But yeah, if you get a couple of one-on-one conversations on your calendar that were, “gonna be for a podcast,” but it’s just Moe or me, then you know that’s probably just us wanting to hang out. [chuckle] Anyway. All right. One thing we do love to do is go around the horn and share our last call. Something that we think might be of interest to our audience. And why don’t we do that now. Hilary, you’re our guest. Do you have a last call you’d like to share?

1:06:35.7 HP: Yeah, so I was thinking this through… As you all know, and maybe listeners, I took this long break after the Biden campaign, so I couldn’t really conjure something work-related for this. So instead, I’ll just talk about hobbies. So I started using a 3D printer. I bought one.

1:06:53.5 MH: Oh, nice.

1:06:53.8 HP: I built it for my kid. Oh, my gosh, it was really hard. I don’t recommend that, but I… If that’s something that you’re curious about, it’s been such a joy to learn that and way easier than I thought. So I was super intimidated by that. And I just… I wanted to use this opportunity to be like, “You should try it. If you’re thinking about it, you can learn the software, you can do it, like growth mindset.

1:07:17.6 MH: But you’ll also… You’ll also throw in pottery, like that seems like…

1:07:20.1 HP: Yeah, yeah.

1:07:20.6 MH: Have you tried to actually print anything that you also… To see whether you can make things that are…

1:07:24.9 HP: This was the grand plan. So it’s funny ’cause I’ve never had a hobby like this where… I’ve had like over a year where I’m trying to do… So my whole 3D printing thing was in service of the pottery, where I wanna be able to 3D print models, that then I can build plaster molds for, which I can then use to make pottery items. So I actually, just today, for the first time, have a pot… Like a clay item that I’ve made, that’s mimicking a 3D printed thing. Anyway, the point is, I never thought this was me, but I want people to… If they’re like, “Oh, I really wanna take a pottery class. I really wanna do 3D printing,” I want to be the person who’s pushing them over the edge. [chuckle] So yes, I’m deep in the physical world realm right now and loving this.

1:08:14.5 MK: Nice.

1:08:15.8 MH: I will… I will say past guest, Matt Policastro, another kind of data science person over COVID completely got into 3D printing as well.

1:08:27.0 HP: Nice. It’s like very self-sufficient. You’re like, “I’ll make my own stuff. I don’t need anything.”

[laughter]

1:08:33.7 MH: That’s awesome. All right, Tim. What about you?

1:08:38.5 TW: So I am going to recommend a book that I just read that is a data work, A Jargon Free Guide to Managing Successful Data Teams by Taylor Rogers. Now, a caveat, it is a self-published book, but not self-published, like get a PDF, like there is a physical book, but earlier we were talking… One of the things he talks about is this evaluating your stakeholders on two axes, one is their low to high IQ, and then the other is low to high direction, which I think the low to high direction was like how clear are they on what they want. But…

1:09:14.7 TW: And it is strange. His very last chapter is best personality types for data work. And I love the list of good for data work, like skeptics, number geeks, tinkerers, consensus builders, creatives, storytellers, and the life learners. And he talks about each one of them, what he means. And then he has the ones like… And these are the ones that are kind of crappy for data work, and it’s know it alls, bossy pants, blamers, reactionaries, incurious, technically averse, and rogues. And I was like… I was thinking of so many people who have been awesome to work with, or terrible to work with, and just that last chapter, and it’s literally the last chapter. It’s a self-published book. I think if he’d gone somewhere else they would have said, “You need to have a conclusion chapter that brings it all together, “Nope,” it’s just like literally ends.

1:10:01.9 MH: I’m done.

1:10:01.9 TW: I think he was like, “And I’m done. It’s my own book.” But it’s a quick read. There’s only like 160 pages but if managing a team, I actually thought there were some really good points. Actually, he also talked about how one of the big parts of the responsibilities is just to have data that can be trusted, which is funny because Hilary you talked about on the Biden campaign, you’re like, “Yeah, basically, our number one thing was to not have a big boo boo that we had to come back with later so… ”

1:10:32.9 HP: Yeah, like not be responsible for Biden losing and Trump winning.

1:10:36.1 TW: Yeah, you did good. Good job.

1:10:36.2 HP: Having to sleep with that every night.

1:10:39.0 MH: Wow, so Tim, you’re thinking of going back into management, huh? Very interesting.

1:10:44.3 TW: Nope. I thought, I hope I might be buying this book for people who are on our…

1:10:50.8 MH: Oh, you’re just gonna be like, “Since you’re my new manager here, I want you to read this, I’ve circled the appropriate parts for you.”

1:10:56.6 TW: Yeah.

1:11:00.4 MH: Moe, what about you? What’s your last call?

1:11:03.2 MK: So, a couple of days ago, myself and Zoe Mears, she is an amazing data analysts that I work with, who used to work for the ABC here in Australia, which is all like public broadcaster, so she was doing like kinda data journalism. We did a session together for Canva on storytelling with data. So it’s all like data viz tips and tricks and how to use Canva to make stuff look good, in data visualization. So that’s recorded, and I’ll share the link from our design school, so you can check that out. But everyone know like… Well, you guys know that I’m obsessed with Adam Grant like I love the man. And I’ve had his book sitting on my shelf for ages that I’ve meant to read, but I just kind of haven’t gotten there. And Adam Grant started tweeting bits of that book. And now it’s giving me, I guess the kick up the butt, I need to go read it, which is Daniel Pink’s book.

1:11:57.5 MK: So one of the things that Adam Grant found in it, that he was, I guess, going on about was like how much people are multi-tasking during virtual meetings. So I don’t know where the number 30% came from, ’cause I’m gonna have to read the book, but basically, in Daniel Pink’s book, he talks about how to have meetings at the right time of day to limit how much multi-tasking people do in virtual meetings. And so Adam Grant is talking about the fact that basically put them in the afternoon, keep them short and small to avoid multitasking, so now I feel like I have to read that book that’s been sitting there for a while, because if Adam Grant’s into it, then I probably will be too. So…

1:12:36.6 HP: It’s funny ’cause I was writing down the title when you said it, and I was like, “Oh, I’m multitasking right now.”

1:12:49.0 MH: Yeah. Oh… Yeah, there’s just a lot of that that happens. Especially in the Zoom world. Alright.

1:12:57.1 TW: What about you, Michael?

1:13:00.2 MH: Well, I’m glad you asked. So I ran across on Twitter the other day, an author who just released a book recently, her name is Renee Teate, and she wrote a book called SQL for Data scientists. So I thought that was fascinating, given her background as a data scientist and sometimes SQL as a language is not necessarily one that data scientists necessarily start with, but it’s becoming much more common to see interplay across those. So I thought for our listeners who kind of live in either of those realms, that might actually be an interesting book for you to pick up and read. So thought I would mention that. And I’ve got one more. It just so happens Tim, I’m cruising through your neighborhood in a week or so on October 12th, it looks like I will be coming on to the Search Discovery Education community to talk a little bit about people and soft skills.

1:13:48.1 TW: Urgh.

1:13:50.6 HP: Yay… Woohoo.

1:13:50.9 MH: So, yeah, super excited, I don’t know exactly what I’m gonna talk about yet as we’re recording this, but I’ve been in contact with Adam and we’re working on a little topic there. So it should be a lot of fun, I’m excited, there’s a lot of good ground to cover and might even use some of this podcast as material, ’cause it sounds like Hilary, you’ve delved in some of these realms as well. So always good to get lots of ideas for this stuff. What makes an analytics person tick?

1:14:20.5 HP: Yeah.

1:14:22.5 MH: Usually anger and wanting to see their work matter, those are the two things…

1:14:26.2 TW: There’s rage coding. Is that one of the…

1:14:29.5 MH: Yeah, rage coding… I don’t know what makes them tick anyways. Okay, those are our last calls, of course, as you’ve been listening to this episode, and we’ve covered a lot of different ground and it’s just been delightful, Hilary, to have you on the show, you’re someone we’ve wanted to chat with on the show for years, and obviously, we all really appreciate Not So Standard Deviations, your podcast with Roger, and if you’re a listener of this show and you haven’t listened to that show, I think you need to start ’cause it’s pretty great. Anyway, we know no show is complete without chatting with you, the listener… Just a little bit, and we’d love to hear from you. So if you wanna reach out to us, you can do that. It’s a really easy. There’s ways to do that on the Measure Slack or on Twitter, or we have a LinkedIn group for you to reach out and ask questions. And Hilary, I know you are active on the social medias… Do you wanna share your Twitter profile? People could follow on Twitter?

1:15:21.2 HP: Yeah. My unpronounceable… H-S-P-T-E-R, So hspter.

1:15:24.4 MH: Yes.

1:15:25.8 HP: To be probably more success… Just Google Hilary Parker.

1:15:28.6 MH: Yeah. Anyway, but Hilary is on Twitter and is a great person to follow. I think I’ve followed you for many years, so that’s a good place to keep up with what you’re up to and doing, so that’s awesome. And of course, no show would be complete without thanking Josh Crowhurst, our… I don’t know. We gotta think of the right acronym…

1:15:48.2 TW: Esteemed? Inimitable?

1:15:50.8 MH: Yeah, it’s sort of like, but he is dignified.

1:15:52.9 TW: Quintessential? Oh…

1:15:55.6 MH: Inimitable is a good word. But he’s our producer. And we love him. Thank you, Josh, for all you do. We think you’re fantastic. Alright, so I know that as you go through life, you may have many challenges and situations where you need to step aside and just meditate for five minutes to regain your center, but no matter what happens, I know I speak for both of my co-hosts, Tim, and Moe, when I say, keep analyzing.

1:16:24.7 Announcer: 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.

1:16:43.2 Charles Barkley: So smart guys want to fit in, so they made up a term called analytics. Analytics don’t work.

1:16:48.8 Thom Hammerschmidt: Analytics, oh my God, what the fuck does that even mean?

1:16:58.5 MH: But Tim is the smart one, he’s also known as the quintessential analyst, I don’t know if he mentioned that and that’ll lead up to the show. But if you could drop back that during any time, any time, that would be really appreciated.

1:17:11.3 HP: Okay. Got it.

1:17:13.8 MH: So, yeah. We should actually get that sponsor, Tim, that’s what I’m gonna do. I’m gonna get you sponsored, and you’re gonna be the first sponsored analytics person, that’s gonna be amazing.

1:17:22.9 TW: Well, I think you’ve taken this far enough.

1:17:25.8 MH: Okay, let’s go. Let’s keep going.

1:17:26.7 TW: We’re now forced to film it, that we have the… People can request. So not only do we have this, I have to, I’m the fulfillment guy.

1:17:39.0 MK: Oh, funny.

1:17:41.9 TW: People have to rec… I have to actually send them out. So every time somebody puts they were like five podcast stickers in zero. I am like in you’re special… Zero quintessential analyst.

1:17:52.4 MH: But it’s good, this is… Hilary does meditation and Tim you do this podcast, it’s sort of how you kinda like process through this stuff.

1:18:04.4 S1: Rock flag and empathy.

2 Responses

  1. […] (Podcast) APH Episode #177: Design Thinking, Empathy, and the Analyst with Hilary Parker […]

  2. […] [Podcast] APH Episode #177: Design Thinking, Empathy, and the Analyst with Hilary Parker […]

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

(Bonus) Marketing Analytics Summit is Nigh!

(Bonus) Marketing Analytics Summit is Nigh!

https://media.blubrry.com/the_digital_analytics_power/traffic.libsyn.com/analyticshour/APH_-_Bonus_Episode_-_Marketing_Analytics_Summit_Is_Nigh.mp3Podcast: Download | EmbedSubscribe: RSSTweetShareShareEmail0 Shares