#175: Searching to Be a Better Analyst with Wil Reynolds

As analysts, it can be easy to get so focused on the data that we lose sight of the imperative that we answer meaningful questions (aka: validating relevant hypotheses). On this episode, we sat down with Wil Reynolds, co-founder and accidental lead generator for SEER Interactive, for a discussion that turned out to be about curiosity and the power of trying to prove yourself wrong (and being willing to invest the time to do so!). In the end, we concluded that Wil has always been a “data person,” even if he doesn’t necessarily see himself as such. That is… actually kinda’ profound!

Links to Items Mentioned in the Show

Photo by Chris Murray on Unsplash

Episode Transcript

[music]

0:00:05.9 Announcer: Welcome to the Analytics Power Hour. Analytics topics covered conversationally and sometimes with explicit language. Here are your hosts, Moe, Michael and Tim.

0:00:22.9 Michael Helbling: It’s Episode 175 of the Analytics Power Hour. Every superhero needs a good origin story. It just makes it better. Just like every analyst comes into the profession from somewhere, often it’s a mix of where we came from that actually makes for the strongest combinations of talent. Hey, Tim, what was it you originally studied before you ended up being the quintessential analyst?

[chuckle]

0:00:54.3 Tim Wilson: Architecture.

0:00:56.8 MH: Yeah, architecture.

0:00:56.9 TW: Not of the software variety. So, yeah.

0:01:00.4 MH: And Moe, you didn’t start out in data and analytics initially. What was your college major?

0:01:04.5 Moe Kiss: I did criminology and political science.

0:01:08.8 MH: Yeah, see, there you go. And I studied History and accounting. So we won’t come to this from lots of different places, and it’s actually the stories of how we got here that are some of the most interesting things about our industry. Well, we wanted to share another story on this show, someone who’s got their own unique perspective and actually began to get a much deeper focus on data and analytics after being, well, pretty darn successful with the other things he was doing. Wil Reynolds is the founder and VP of Innovation at Seer Interactive. Wil founded Seer out of his living room in 2002. And today they there are over 200 strong, helping clients with marketing and data solutions. Wil’s a frequent speaker, an industry contributor. Over the years, you’ve probably seen him, if you’re in the SEO industry, at the Moz Conference and things like that. And he’s also very active in his local community of Philadelphia, helping to serve the homeless and runaway youth populations. And today, he is our guest. Welcome to the show, Wil.

0:02:07.2 Wil Reynolds: Thank you for having me, it’s good to see your faces this morning.

0:02:10.3 MH: Yeah, so tell us your origin story. How did you get to data, what made you as a successful agency founder start to focus on data and analytics so much more? And obviously, we’ll talk about that for quite a while, so you don’t have to try to blaze it through it all at once. But yeah.

[chuckle]

0:02:30.1 WR: Well, for me, I went to school to be a teacher, and just yesterday, I tweeted something that was like, that was just the best gift, because when I came into an industry, and like you mentioned, I speak at all these conferences, I never thought of speaking at conferences as business development. I thought it was teaching. So I had this freedom of just like, “I’m gonna share everything that I know,” because that’s what you do when you speak. People are like, “No, it’s BD. You’re supposed to come back with leads.”

[chuckle]

0:03:00.7 WR: Oh, my bad. I thought this was about teaching people stuff, shocker. But from a data standpoint, I’m still so early in my journey. I’m humbled to be here because I’m really early in my data journey. I think I started realizing how little data I was analyzing when I was coming up with my search strategies, and I reflected on how am I this well-known speaker and yet, I’ve probably worked on a thousand websites in my career, of which there were billions, and I would go after 200 key words and then you realize that you look at a client’s data and you’re like, “Wait, they’re actually targeting tens of thousands, if not millions of keywords at a time in their paid campaigns,” and I’m like, “Okay, let me bring that over into Excel,” and then you go, “Ugh.”

[vocalization]

0:03:52.1 WR: When you try to bring in 4 million keywords into Excel with 50 rows or 50 columns of unique data in each one, and then you try to do a VLOOKUP to join it, you instantly realize that this is the wrong tool, and that’s what got me into learning how to manipulate the data. It was a need to solve a problem that I saw in our industry that started with me. I’m on stage speaking about what I know, and nobody ever just challenged me in 20 years to say, “Wait a second, you’ve worked on a 1000 websites out of billions. How do you know that what you’re saying isn’t an anomaly?”

0:04:26.4 TW: So that… It was actually… That specific anecdote, was it kind of a… It was the volume of the data, and you were just saying, “I just wanna… I want to do something simple. It’s just the volume is such that I’m in the wrong tool, so now I need to dive into… ” Was it SQL? Was it database? What did you…

0:04:42.5 WR: Oh, that’s a whole other story. So somebody at Seer, had only been at Seer for six months, saw me struggling with this VLOOKUP crap in sheets or whatever I was doing, because I share these videos, I probably share a video every day of something that I’m working on, something I’m trying to figure out to the company. And she goes, “There’s this tool call Tableau, you know. And you can join data sets and visualize things.” And another lesson, so often we in the C-Suite sit off in our corner somewhere, getting all of our kudos, and here I am with somebody only six months out of college just teaching me about what Tableau even is, and I downloaded Tableau, but it’s only a license for like two weeks. So I’m a CEO, I have to run the business, I don’t have time, just let me just play into Tableau for two weeks. So literally, after two weeks when it says, “Oh, your data needs to be in public,” which it was client data, so I couldn’t do that, I went and found Power BI, downloaded it and have had it on my laptop ever since.

0:05:40.9 MK: So what did you think in that moment when, yeah, this young person working at the company, is six months out of university, is like, “Hey, have you heard of… ” What went through your head?

0:05:52.7 WR: Oh God, I went, “I’m so glad she’s here with us.” Because one, I was proud that she felt comfortable to come to the CEO of the company and say, “I think I can help you,” when she’s only six months out of college, because I think a lot of times, for a lot of folks, they’d be like, “What do I know?” And I love when team members feel like they can be like, “I know more than him on this, I’m gonna tell him.” Right?

[chuckle]

0:06:16.9 WR: So that was the first thing I thought of is, “Thank God I got somebody who’s not afraid of the CEO,” ’cause then she would have waited six more months or a year before I would have learned this stuff. The real thing that clicked for me is I went, “How has this become normal in our industry? We have data in paid, and data in SEO. There is no paid-only search engine and no only organic search engine, but yet no one has challenged that status quo by simply saying, “If I can put a unique key search term in this table and a unique key search term in this table with rankings and put them together, I can actually more represent what the customer actually sees,” which is some interplay of snippets and paid ads and all these different things. It’s still to this day… I always say this, find me a tool that combines both. No one.

0:07:05.0 TW: Yeah, both. [chuckle]

0:07:05.2 MK: Yeah, I was like, I think people do it.

0:07:06.7 WR: But our customers see it in one place, but the tools are built to keep us in our silos, it’s crazy.

0:07:16.1 TW: So, maybe this is a little bit of a shift, but it is interesting you’re… And having known you for a while, and knowing that you are a creative and active thinker and learner and teacher, and as you’re talking about running the company and doing these things, I do think that one of the great things about coming in with getting all this domain knowledge and doing all of the things that are search-related, then you’re using the data. When you’re coming at the data, you really have, I assume, had a lot of clarity around what you were trying to do, what problems you were trying to solve, what questions you were trying to answer, which I think is actually where sometimes in the industry, especially analysts that are newer to the industry, can kind of struggle, is that they wanna dive into the data first.

0:08:13.0 TW: I would much rather have a stakeholder who comes and says, “I don’t know the data, I’m not a data person, I just know I need to do X.” That’s actually an easier conversation to have. A lot of times, to me, it feels like there is a, “Oh, look at all the data I have, what can I do with it?” And that winds up being kind of a recipe, I don’t know if you feel that way. By the time you were digging into that, your time was so constrained, it sounds like you had things you were trying to answer or do that were really grounded in a deep knowledge of the business problems you were trying to solve. Is that fair?

0:08:49.4 WR: Because they’re grounded in hypotheses, that I can’t see anyone answering well. So, for instance, you’re like, if Amazon is paying for ads and I rank number one for a keyword, that’s gotta probably affect that in some way, and it has to. You would think there’s gotta be some effect there because it’s a major brand, I know I can get two-day shipping. People skip over my result to go to Amazon because they understand what the return policy is gonna be like and all that. But yet, SEO’s are just like, “Yo, if you rank number one, you’re gonna get 10% of the clicks,” and you’re like, “That can’t be.” Wouldn’t it be helpful if I could see how many ads were above me, if Amazon is there or a major competitor is there?” And it just… You know what it is, it became an addiction after that though, Tim, because then it became the speed that I was knocking out hypotheses invalidating things that a year ago I set on stage as anomalies and not as something everyone should do, just became this like, “Oh my God, it’s great.” These ideas that I’ve had have been either validated or invalidated, which I also realized was my mission. It wasn’t to be right. It was to either validate or invalidate things, longly held beliefs.

0:10:02.8 MK: Could I just draw an observation here? Which is that you’re not the standard VP or CEO or lead, like all the things you’ve…

0:10:17.6 WR: I keep hearing that.

0:10:17.7 MK: All the things you’re saying, I’m like, “Oh my gosh, like how cool. Can I have you being my boss?” Because I’m just thinking of me marching into some of my ex-CEO’s office and being like, “There’s a tool called Tableau, I think you should download it,” and I’m not sure they would have been as receptive as you. And it sounds like… I think Tim is gonna cringe when I say this, but I can’t think of another way to describe it. It sounds like what drives you is, wait for it, Tim, a real growth mindset.

0:10:47.1 TW: Oh. Killing me.

[laughter]

0:10:52.3 MK: But it sounds like you’re so passionate about what you do, and that’s what’s spurring you to learn things.

0:11:00.0 WR: I’m passionate about knowing when I’m wrong. You know what’s so interesting is, another one of my co-workers, he had more of a background in Python and coding, and one day… The first thing that led to me even trying to join all this data is, he asked me to go head-to-head on his script that he built to do keyword research versus me. And what was great about that is I was like, “Let’s do it,” because it has a chance to invalidate things that I talk about on stage all the time, and the script he built was so good at beating out things that people want me to come all over the world to speak about that it was another one of those moments where instantly, you’re like, “Holy smokes, I can’t believe people have me come and speak at these conferences all over the world just to know how great I do keyword research when I just got my ass beat in 30 minutes by an automated script.”

0:11:49.5 WR: And that was another aha moment, where I think so many of us run from that because it invalidated our experience. So, you’re like, “I’ve been doing this for 20 years,” so like, I don’t even wanna see something that could be automated beat me, so then you fight it for five years along with everybody else, or 10 years, or however long it takes. But then for me, that’s an exciting moment. So, I’ll tell you, Moe, from my standpoint, I guess I’m different in that regard to come back full circle, but I don’t… I started Seer when I was 25, 26. So I didn’t have enough experience out there to have other managers, so this is the only way I know to be. So for me, I’m like, “This is the way it is,” but there’s a downside now. We’re not as profitable as a lot of our other competitors, because you don’t have a CEO who wakes up every day looking at margins, right? I have a whole team that manages the business, so I can focus on my part, but you get me a dataset, and an Amex with no limit, and I’m dangerous.

[laughter]

0:12:50.8 TW: To be fair, Moe, Wil will swap CEO roles with Rand Fishkin for a week. Wil’s a little out on the tail of the normal distribution of executives, so I’m pretty sure I’ve never…

0:13:08.6 MK: Can I even just zoom back to even when you started talking about your experience speaking, and you mention that the point of speaking is to share your knowledge, I think… Yeah, I feel like I’m gonna end up with another professional crush by the end of this bloody episode, because to me, that is the point of speaking, it’s to help share what you know. And I’m like, “Really, someone asks you to come back with leads, is that a thing?” But then I’m like, “Oh, I don’t work in agencies, so maybe that’s why I don’t… ” I don’t think of going to a speaking event as going to get leads. I think of it as going to learn from other people and hopefully share some stuff, but…

0:13:50.7 WR: No, so often, it’s considered BD. ‘Cause here’s the thing, so people would look at me like I had eight heads when I would say… ‘Cause early on at Seer everything was all referral all the time, so I was like, “We don’t do BD, we just get referrals.” And people would be like, “But you speak at conferences,” and I’m like, “What do you mean? That’s got nothing to do with BD,” and people would look at me like I was a liar. They’d be like, “You do BD. What do you mean you don’t do BD?” Because they didn’t process the fact that just ’cause we get all referrals, that I considered that to not be doing business development and just doing good work for your clients and hoping that the word spreads.

0:14:20.0 TW: That’s right.

0:14:22.3 WR: But people literally kinda… They discredited me. They’re like, “That guy is full of it. He’s lying. He’s doing BD, but he doesn’t… ” So then I was like, I guess this is considered business development most places, ’cause I’m getting this visceral reaction back when I tell people I don’t do BD, at the time.

0:14:38.8 TW: I feel like if you go to… And now, when I’m back to going to conferences, I almost wanna do a scoring on a two-part rubric of who is on stage clearly because they are selling there, and I’m thinking specifically through certainly in the CRO world…

0:14:57.1 MH: You could already tell that though. It’s who you like versus who you don’t.

[laughter]

0:15:03.9 TW: Well, that’s the thing. I think that’s what it becomes. You’re like, “Oh, they’re talking through their process,” every time when it’s… And then we do this and then we do that, and then, oh, guess what, at the end, you can come to this link and download. All you have to do is put in your email address, and you can download our toolkit, where it’s such a polished and baked… And this is the marketing and sale. And not that the person isn’t knowledgeable, but they do have that mindset versus the ones who stand up and occasionally say, “Well, here’s where I screwed this up, and that didn’t work at all,” or, “I tried this other thing.” Or “You know what, four years ago I stood on stage and said the complete opposite, and I was dead wrong, and I feel bad about it.” And those are the ones that are actually more interesting to listen to. They’re more memorable, and you tend to feel like you’ve taken more away from them.

0:15:53.2 MH: Yeah, this is why I already, Moe, have a professional crush on Wil, ’cause this is where I feel like he and I are like, how did we become like cousins or something?

0:16:03.8 WR: How did we not earlier? How did it take so long?

0:16:08.0 MH: Yeah, yeah, exactly, ’cause I just introduced myself to Wil randomly a few months ago, and I was like, “Hey, I wanna talk to you.” And he’s like, “I’m really busy,” and I was like, “Please.”

[laughter]

0:16:21.1 MH: But the very first presentation I ever did at an analytics conference, Wil, was titled 10 mistakes I’ve made analyzing the web, and it was the number one rated talk at that conference, because all I was trying to do was help people learn from the things I did wrong. There’s never not gonna be somebody willing to stand up and take everybody’s praise and adulation of like, “Oh, you’re the smartest guy I know.” Thank you, please. I’ll be signing autographs later, but there’s always a shortage of people who are like, “Hey, I’ve been in this work, I really love it, and I’m learning all kinds of things, and I wanna share everything I’ve learned with you so that you can avoid my mistakes or you can be even better than you are today.” Heck, this podcast, frankly, is built on some of that same thinking. Only later did we realize that we were huge deals and so awesome. No, I’m just kidding. [laughter] I can’t even say it, I can’t even say it with a straight face, I’m sorry. But yeah…

0:17:24.5 TW: If you go to analyticshour.io/givemeyourleadcontactinfo, you can fill out a form.

0:17:31.5 MH: So, I’m trying to get this around to a question, but one of the things that I’ve found is that I’ve definitely felt like I’m not normal, and I think Wil kind of in a similar way that you said it, but over time, I’ve found this way to align, sort of like… Yeah, when you speak at a conference, I don’t think of it as business development, but I don’t not let it be business development. It’s just, I’m gonna do it the way I do it, and if people wanna be clients after that, well, cool, that’s pretty awesome, and let’s explore that. But I don’t know, as you were going along, Wil, you obviously had a lot of data that you were looking at as you were looking at search rankings and things like that, so what was the switch, ’cause you’re doing data analysis on a certain level, but that was, I think you would say that that’s not you being a data person. I don’t know. How would you describe the difference between those two things? Because it sounds like you were already doing data work pretty consistently throughout.

0:18:35.0 WR: But it was so basic. It was like addition versus… I’ll stop at Algebra ’cause I didn’t do anything after that. But here’s what’s interesting, I just love… I love simple. I love simple, I love when simple is just in front of you, somebody shows you something so simple and you’re like, “How have I gone through the world not realizing this?” So, there’s a guy named Stephan Bajaio from Conductor, and he and I once were just riffing, but he really was the one leading this, where I’m like, we have to… To find out where you rank on Google, unless you use one of their tools to do it, but don’t do that. When you find out where you rank on Google, a scraper is basically going through the entire search results and looking for your domain and saying, “Oh, you rank number four.” “Oh, you rank number 10.” So it’s gotta scrape the entire set of results. And one day Stephan was like, “Nobody considers this data as intelligence. They consider it as a singular data point in a place and time,” so we start riffing.

0:19:36.3 WR: And I’m like, “Holy smokes. Google has invested billions of dollars in research to make sure that that search result understands what users want better than anyone else at scale, instantaneously, yada, yada yada.” So every time you scrape it, you are gaining… You are taking pennies on the dollar for what Google had to invest billions of dollars to learn about what they’ve learned, that is the right answer for this person at this time. So last night, very late, I was up, and we took a million domains, and by “we,” somebody else in my company who knows Python, took a million domains and looked for ad.txt at the end of all of them, and then I applied that million rows to this 400 million rows set of ranked data with domains, and now I can, across my entire client database, say, “Oh, which keyword has Google determined that publishers with the ads.txt are the right answer at scale?” So now when I go back to a client, I can go, “Do you know that 80% of all of this group of search terms that you spent join to your paid data? So you spend this much on these words and got this many conversions, is that good?” They go, “Yeah.” I go, “But what’s interesting is 82% of all the search results are publishers, so that tells us something about how we might need to write the content to compete. And then over time as Google’s machines learn, you start watching them be like, “Publishers aren’t the right answer.” “Oh, this is the right answer.”

0:21:08.6 WR: So when you keep scraping that data month in, month out, you start to be able to track how Google is learning about what answers this question for this person better than anyone else does for literal pennies on the dollar. Once you start seeing a Google scrape as intel, you start doing things like this, “Show me all of my SaaS clients where a map is showing up in the top four results. Why would I do that?” Because a map is a local result, they’re a software business, somehow Google is targeting incorrectly in spending money, right? So my client is spending money on a word that the odds of a map being the right answer are completely false. That’s the kind of thing that you do when you start looking at all those related searches, all those people also ask… You start saying to yourself, “Wait,” Google has seen your search, your click, how long you stayed on that site, if you came back or not, what you clicked on next, how long you stay on that site, and all that’s being learned on over 20 years of click patterns to determine this site never answers the question well, so don’t follow it. So to me, that was the aha moment. I was like, “Oh my God, this is intelligence that I can use in so many different places that it just blows my mind.”

0:22:21.5 TW: Yeah, go back to, it’s the curiosity and the creative thinking and the what if, or all the… I think that’s where there’s the Wil Reynolds beautiful mind part of always asking and learning, then say…

0:22:37.5 WR: Teaching. It’s teaching.

0:22:39.2 TW: Yeah. [chuckle]

0:22:40.2 WR: You’re always trying to learn. You’re trying to learn and teach at the same time. I’ll give you another example just so that people can understand the power of what you can do with this data. This is basics, but I can’t do any of this, I have to get much smarter people around me to do it, but I have the ideas, and then I go, “Can somebody help me, please?” COVID hits. First time COVID hits, so clients are freaking out, freaking out. We took, in about a week, every search term for every client over a 12-month… No, six weeks before COVID hit before, and six weeks after we had gotten used to this new reality. ‘Cause you have to think, COVID changed a lot of stuff in our lives, and our clients were freaking out and they didn’t have good answers because nobody had been through this before. So what do we all do? We base it off our gut and bullshit, and I’m like, “No, we’re gonna go get some data, where do I have data that can make me answer this question better than how every other person with 20 years of experience is going to go at it?” And I said, I got every freaking search term for every client in this data warehouse.”

0:23:45.6 WR: So I said, “Can somebody n-gram every search term for every client, then rejoin back all the paid data at the individual n-gram level?” So each piece of each word. So if the word is “Samsung mobile phone,” each n-gram, Samsung is an n-gram, mobile, and phone. Then we stitched all the data back together and we said, “Which search terms increased the most in impressions after this six-week change?” And you started seeing little pieces of words added to other words that showed potential change in how customers searched. Our biggest finding was, for about three weeks after COVID hit, when we got used to this, “Oh my God, this is a pandemic, it’s a real thing,” the word that people most used, associated it with… Not most. Sorry, not most. A word that had a big lift was FDIC associated with our banking clients words, ’cause for three weeks people were more worried about the safety of their money than they were worried about the rates that you were offering.

0:24:48.0 WR: So then because I’m scraping the results, I get to ingest all the ads and find out where is no one speaking the language of the customer. The customer changed, and we can go, “Oh, we’re scraping all those ads in, so now we can see none of our competitors are speaking their language,” but the customers changed their language, so now can we make better ads right now and get that information back out to our clients? Because that’s running the data warehouse, instead of trying to do that bullshit and sheets and downloading CSVs and whatnot. We did it for every client and recorded a video for each client in our company within a week, five of us. We’re just like, we’re sitting in a room, we’re gonna record it, and like, “These words are jumping, these words are falling off, this is happening and you’re da da da da da,” and it’s like, that’s the power of this data thing that I’m on now. So obviously then I got hooked. And then I would start to look at Google Trends to see if it validated the spikes, and you’re like, “Oh shit, it did spike.” You’re like, “This is right, I think this is right, I’ve now validated it to the best that I can.” And for me, that’s another example of the data that we have, and we have a lot of it, it’s just so easy to reshape it and then turn it into instant client value, especially at this time. Now that COVID’s popping back up again in the United States and pretty much across the world, we already have all that done.

0:25:57.3 WR: So now we can just refresh that, change the date ranges, and now we’ll start to see how people are changing their language again to see if it matched like I did last time. But that’s customer intelligence. It’s instant customer intelligence, and people put things in Google, they would never put on Facebook and they would never answer in your surveys. So we get to understand things, like the word transmission. I have a client in the auto space, the word transmission changed when COVID hit. So all of a sudden you started being like, Why is the CDC showing up for transmission words, and you’re like, “Oh God, that’s an outlier.” That’s Google’s intelligence. So that when you scrape the results, you’re literally saying, “Why is a medical site showing up? It hasn’t showed up in three years, four years, or any point in my dataset. Why did it show up this week? Oh, my God, Google is actually looking at transmission word so broadly because people are worried about transmission of COVID that that is causing my client to maybe get matched to literally the word transmission at one point ranked at number 10 on Google for the CDC’s website, not for things related to your automotive transmission.”

0:26:58.0 MK: Okay, but you gotta explain to me, Wil, it sounds like there was this real journey, and I’m not gonna… The work you’re doing sounds so fascinating. I like hanging off every single row of every column. How did you go through this as a business? You talked about, at the start of COVID, there was like five of you that got in a room. You talk about importing this data from the data warehouse, so at some point the Google Shade or the whatever wasn’t working, and so now you’ve got a team at work, there’s a data warehouse. What was the evolution? And I feel like there might be some part to SQL for you as well on this trip.

0:27:35.7 WR: It was quick. It was really quick because once you see it, that’s the beauty of owning the business. When I see good shit, I don’t have to go ask somebody if we can go drop a million dollars on it, right? So I saw it, and I was like, “This is just a better way.” It was just a better way, and once you instantly go, “This is a better way,” you can’t have… So Seer, because we’re a high referral business, we gotta be like Starbucks, we gotta taste the same everywhere, because your clients come back in over time. And then if you have big discrepancies between how your teams execute, it’s like, “I thought I was getting that. That’s why I came back.” So then another thing… And I realize that my team was all over the place before. Our whole history has been like a bunch of hacker types, so what happens is each person kinda does their thing their own little way, but our client base was like, “I don’t want that. Why can’t you guys just do it the same way every time?”

0:28:20.5 WR: So also by warehousing this data, it enabled us to be able to take a hypothesis for one client, see if it’s true across the entire client data set, and that was all some people smarter than me. It was just me leaning on the shoulders of team members that have been with me for eight, nine, 10 years, who have pivoted their careers into more data engineering, Python, etcetera. And they go, “Can you set this up? Can you set that up? Can you set this up?” But eventually, I did learn SQL because I hypothesise all the time, and I hated being on someone else’s schedule. So I was that CEO that was like, “Hey, I’m looking at this table and I’m wondering if this.”

0:29:00.1 WR: But the data lakes are so big that I couldn’t bring them all in the Power BI, so I had to cut them down to actually hypothesise or my files would just take forever to run where it would crash my laptop. So then I was like, “Oh well, I don’t wanna be on your schedule Moe, I don’t wanna be on your schedule. What if I have an idea on a Saturday at 8:00 AM that I wanna work on, I don’t wanna have to send it as an idea and wait till Monday. Because by Monday, I might find out that it was completely invalid. So then I went to community college and just started taking SQL classes, not to get good at SQL, but to at least do select where, like, and a couple of basic things to build some basic tables to validate my hypothesis, ’cause I’m addicted to hypothesis validation and invalidation. It’s just so much fun to be like, “Oh my God, look how wrong I was.” Right? Or, “Damn, I knew it.” Right, and anything that gets in the way of that is frustrating for me.

0:29:50.0 MK: And was that the case before you, I guess, had this data pivot, if you’d allow the pun, or was the hypothesis validation always there, and it’s just become… Your methods have changed about how you validate or invalidate now?

0:30:06.3 WR: It was always there. It was always there. And you know what’s interesting is, I think Mike was mentioning earlier about speaking at a conference or something and getting great ratings. So it’s so interesting because… Also somebody asked me a question yesterday about speaker order, and I’ll tell you that’s people that speak. It pulled something out of me I never thought of, which is if you speak last or towards the end of the day, people have somebody to compare you against. So when you get your ratings if it’s a conference that rates you in kind of real time, if you’re the first, second, third person to go, you can’t really use those ratings as well to get better, because they haven’t had the chance to compare you to anybody. So anyway, that’s a little speaking hack that I was not expecting to bring up, so I’ve always been the kind of person that just enjoys validating and invalidating my hypothesis. It’s fun to see how wrong you are on something, to discover the world in a way where you’re, “Oh my freaking God, how am I going through the world seeing this?” And to be on stage speaking, and have all these great speaker ratings.

0:31:05.8 WR: So the data that I have says that I’m pretty good at it. And then to go back into your own data and be like, “I was so wrong. I was so wrong. How can I be rated that well and be so wrong on so many things,” for me, was like it was a joy, ’cause I live on being a rookie. I live on being a person who is in something new and learning it for the first time, versus being… I think we have to be careful, ’cause sometimes your ego just gets filled off all these speaker ratings and shit,” so it feels great. But when everybody’s telling you you crushed it, nobody’s telling you a better way to do something. And then that makes you feel better, so then you don’t go out and seek the people who are telling you you’re doing it all wrong, and to be able to do that for myself was insanely valuable.

[music]

0:31:53.5 MH: Let’s step aside for a minute, something that gets in the way of every great analysis, potentially poor quality data… Seeing your analysis get put aside because it doesn’t have credibility in the data, it’s just a continuing problem for data analysts, data scientists everywhere. Moe and Tim, what have you done when your career is about this, or what could be done to resolve that?

0:32:22.1 TW: Well, using a platform like our sponsor, ObservePoint, to put the machines to use to actually check the integrity of the data across your entire website is definitely one way to go.

0:32:35.9 MK: And I’m not gonna lie, the best bit is that you can get alerts when something goes wrong.

0:32:39.1 TW: That means checking across your most important pages, userpass, making sure the functionality, the data collection that you are expecting is actually getting collected.

0:32:49.3 MK: And let’s be real, I get pretty excited about being able to track data quality and QI progress over time. So, yeah, it’s for the win.

0:32:58.1 MH: I love it, and you can love it too, you can get a demonstration of all ObservePoint’s data governance capabilities at observepoint.com/analyticspowerhour. Go check them out, learn more about ObservePoint. Alright, let’s get back to the show.

[music]

0:33:16.9 TW: It also seems like as you’re chasing where you’re wrong, or that cycle of, “Gee, I’ve stood up on stage,” and there is a degree of being clear and concise, and an effective communicator and telling a good story, there could be lots of things that get told. There are political pundits who say things, and it’s actually, if you’re not really paying careful attention and catching the logical fallacies, you can understand why people say, wow, they just started at A…

0:33:45.3 WR: That’s true.

0:33:46.1 TW: And ended up at 14. They didn’t even follow the alphabet, but it was a good narrative.” And so, I think there are lots of things in the industry where that has come about. My current one is the… We continue to obsess about attribu… We took a wrong turn 20 years ago on attribution, thinking that the tracking user level data, we could just crunch the data and get attribution. To me, that’s one of my biggest ahas in the last 12 months, was, “Oh my God, we took a wrong turn and have been chasing that for 20 years.” And even though I have been anti-multi-touch attribution or the way that it gets pursued, it’s only been more recently that I’ve realised exactly how fundamentally problematic that was, but to get back around to say, I’ve stood on stage within the last three years, and there’s one out there that, thankfully it was not recorded, I’m embarrassed with what I stood up.

[chuckle]

0:34:41.7 TW: And it was well-regarded, well-rated… I think they rated it. People were excited. I felt pretty clever, ’cause it was something that was a new idea, but it was dead wrong. But I think that’s actually when you say, “Oh, I found something that was wrong, but you know what, I’m a sharp person. I was trying to do the right thing, I was thinking creatively, I maybe even talked about it, I organised my thoughts well enough to teach others this. Oops, I taught them that the world was flat.” Now, it’s actually more powerful because I’m like, “Wait a minute, other people think this, and I can understand why they thought this, and it’s wrong.” But now it’s actually more fun and powerful and useful to say, “Yeah, I was there. I thought the world was flat. Now, let me help navigate you through why the world is not flat, because I’ve found something else.” So I think there is a point to that. As you keep saying you love finding out that you’re wrong. Well, when you were wrong, it wasn’t because you were being lazy or stupid, it was because you were following a logical path to something, and then it’s only been the other missing piece that you plugged in, which means lots of people haven’t found that missing piece. Now, I’m talking in abstractions to, I guess, protect the innocent, except for multi-touch attribution. Anyone who’s out there…

[laughter]

0:36:07.3 TW: Anybody who’s out there that’s expounding on time decay attribution is just… Should be flogged.

0:36:14.0 WR: Well, one of the things… I don’t know if you’re touching on this there, but the promise of data and how trackable things were back when we all started, this idea of every single person you can track them and they can do this, and you can do this and that. It’s funny to realise how much of a farce that was. We just thought that the data would get more robust over time, and then it just got more fragmented. It got like, there was mobile devices, and it’s like, “Oh, there’s a way around that,” and then it just kept being like, “Wait, maybe we need to stop believing in this promise we were told 20 years ago that you’re gonna be able to track Tim Wilson through your website, and know it was him, and know that he clicked on these five pages and blah blah blah blah pa, pa pa pa… ” And that was always the promise, that this data… That digital would provide all this data that would help us make better decisions, and it does, but I think the way it was sold was like, “You will know this person at this time bought this thing from you because of these reasons,” and it’s never ended up being true.

0:37:10.1 WR: The one thing I love about the whole cookie deletion issue and ad blocking and all that crap, is it just shone a light on something. And I think a lot of us were like, “Oh, wait a second. That’s what I was promised when I was 20 years old and started in this industry, that someday it’s all gonna be trackable.” And it’s like, “Thank God… ” And it never happened in 20 years. It got better, but it never happened, but the promise was always the same, that if you had all this data, you would just get answers. And you’re like, “Wait, actually, getting that data from people and finding insights is really fucking hard.”

[chuckle]

0:37:39.2 WR: Really hard.

0:37:39.9 TW: But that has made it now, because of the whirlwind with the cookies and regulations, there are people… There’s a whole cottage industry around, “But let’s come up with technological workarounds for that,” because it’s still chasing that fallacy that there’s a way to get to all that. There are two parts. You’re never gonna get there, and it’s getting worse. And also, even if you had all of it, there’s still some big gaps. There’s a reason that A/B testing actually gets you to causation. So that’s this other monster fallacy, but I get that the allure of, “If I can just come up with… I can teach people about how to work around ITP,” and it’s like, well, one, that’s definitely gonna be a short-term… Whatever fix you come up with, if you are explicitly or implicitly saying is a workaround…

0:38:25.8 WR: 100%.

0:38:27.6 TW: It may feel technically like you’re doing things, but you’re just rearranging deckchairs on a barge that never made it into the water, much less the Titanic, so… Sorry, that just headed me down. Rant of a…

0:38:41.9 WR: No, preach. I’m still here, like… Preach.

0:38:45.8 TW: I actually have slides that I’m using in a few places that are around that very rant that you just had. Some of the exact same phrases around a farce. I’ve said that. But I was there. Same thing, I was there five years, 10 years ago, I was chasing that too. Let’s just connect all this stuff together, that’s the promised land. And only now as people are wringing their hands and gnashing their teeth about, “We’re losing that.” You’re losing visibility you never had at a great spot, anyway. You had your bounce rate to three decimal points, but that… Forget about the validity of a bounce rate, you actually were missing things, anyway, so.

0:39:27.9 WR: You know what, one of those things for me is paid search performance, and we’re starting to look at how our competitor campaigns perform when they get bad news. If your company gets panned as being a bad company, it typically hurts you for some amount of time. It’s in the news, people are thinking about it. And it’s like there’s all these models being run, but most models that I see on like, “Hey, this is how we attribute paid,” never go, “We brought in the news results to see whether or not you had some bad news at that time, and maybe that’s why your freaking shit didn’t convert as well, ’cause you ripped off a bunch of students, Wells Fargo, and gave them student loans that they didn’t know they had. That’s why your… ”

0:40:06.8 WR: Right? But this is happening across the internet, and we have this… And that’s what happens, then is, we sit in a room, we have all this data, and we’re like, “Yeah!” And this is another thing about data. It’s like… You’re like,”Yeah!” and then someone exactly will be like, “Wasn’t that the same week that the website went down and had problems?” And you’re like, “Oh, I didn’t have that anywhere.” And then you go and run off and get that data, and then you show it to another person and they go, “Wasn’t that the week that we did TV ads?” You’re like, “I don’t have that in my… Ugh. Okay, then where is that data? What’s the schema like? Can I get it?” You know what I’m saying? It’s like, “Oh, we were doing billboards at that time in that state.” “Well, do you have the latitude and longitude of all the billboards?” “Well, no.” And it’s like it just becomes this chase of the data and you realise you’re never gonna get there, so the earlier you can give up on it, the better.

0:40:52.7 MK: But so, Wil, how will you make decisions then? When you’re in the driver’s seat, how do you not end up in that spiral?

0:41:03.7 WR: You know what, I’m very comfortable with saying, “This is the best information I have at the time to make a decision on that I have to right now.” And then I also say, “What’s the fatality of this decision if I’m wrong?” I analyse everything through upside and downside, literally everything. So as a result, I’m like, “What’s the downside if I’m wrong, what’s the upside if I’m right?” And once I’m honest about those things with myself, it makes it very easy for me to make decisions in a world with a lot of imperfect answers. It’s very easy for me to make decisions, ’cause I’m not a doctor. [chuckle] You don’t want me to doing surgery ’cause I’m like, “I’ve always wondered what happens if you slice that part. Oh.”

0:41:43.0 MH: See, in that statement, Wil, I think you’ve actually given us a peek into something, which is, I think you’ve been a data person all along, ’cause the framework for making those decisions, that’s using data. I don’t know. I think you’ve been a data person that just learned SQL over COVID, I think you’ve been a data person, so… I don’t know. I hear you say that, and it’s like, that’s brilliant. People should take that one thing, if you’re a decision maker and just sort of pick two vectors. Is this the best information I have? And what is the fatality, I love the way you said that, of this decision? What’s the worst thing that could happen with this decision? And use that as a framework. I’ve studied decision-making a little bit, and my old boss, Mike Gustafson, he actually maintains a decision journal where he writes down, here’s the decisions I made under what circumstances.

0:42:41.7 MH: And then he goes back and looks and is like, “Was that a really good decision or was that a not so good decision.” so he actually… So I aspire to that someday, but that’s who taught me about it, but this is exactly how that gets done, and I think to be effective at any level of leadership or as you grow in your career, learning how to try to make decisions, just like you’ve described, is probably one of the most important things you can learn to do. And it’s interesting ’cause I just… I sort of feel like I’m like, “Wait a second, you’ve been a data person all along.”

[laughter]

0:43:19.9 WR: But honestly though, you know what though… And I know that you’ll be like, “No, Wil, you shouldn’t look at it that way,” but when I think of data people, there’s been so many times I wanted to support something that Tim was doing, and then I would watch it and be like, “Oh shit, R. Oh man.” You literally are showing up for a friend of yours, somebody who you respect and have learned so much from, ’cause you wanna show love, you wanna show support, and you’re like, “Oh man, I can’t get any of this. It’s so far over my head.” So when I think of data people, I look at some of the work that Tim does sometimes, and what he talks about, and I’m like, “I ain’t that. I ain’t never gonna be that.” So I think the people I compare myself to… To me, Tim’s a data person, I’m a guy with data.

[laughter]

0:44:06.8 MH: I know exactly how you feel. It’s exactly right.

0:44:10.3 WR: I am a guy with data, I am not a data person.

0:44:13.3 TW: Except, I’d been a data person for years, the first time we met, and I wasn’t talking about any of the R type stuff. You and I wound up…

0:44:24.2 WR: That’s why I loved you.

0:44:24.3 TW: Having beers and talking about hypotheses and measuring performance, and other things.

0:44:29.9 WR: That’s why I fell in love with you. Literally, I was like, “I hope this man’s not married. Damn!”

[laughter]

0:44:36.3 WR: Because what you broke down in that moment for me was this amazing simplicity of hypothesising, and I kid you not, Tim, it gets referenced in my company once a month, ’cause I have it on auto-repeat to come back into my inbox every month to look at that presentation and ground myself again in how to structure hypotheses. That’s how good it is. And you gave that presentation six or seven years ago. I had it come in my inbox every month to remind me to ground myself and like, this is how you think about… I have the document, I have it linked, and you know it, Tim, because every…

0:45:12.8 TW: You tweet it. [laughter]

0:45:13.0 WR: Four or five months, I tweet it. I’m like, this is still the best way to get the people thinking about hypotheses.

[laughter]

0:45:17.3 MH: Yeah.

0:45:17.8 WR: But then sometimes you were off on some R shit, and I was like, “Oh man, I can’t hang with that. I can’t… ” I’m trying to understand it. I know there’s smarts in there, but it’s brain… I’m trying to still make sure that Select Star, where from like works 90% of the time, so I’m just like… But that’s the beauty, right? It’s the swirling of these different perspectives that help us to make our own, which I think is the beauty. I might not be able to take that part from you…

0:45:45.3 TW: Let me tell you how many leads I got out of that presentation.

[laughter]

0:45:50.1 TW: I was not fishing for that, [chuckle] I was actually trying to agree with Michael, but I think, Wil, you’ve been a data person all along.

0:46:00.4 WR: But that’s real. You know me, Tim, by this point. I don’t sit up here and tell people shit that I don’t believe because I’m gonna forget it all, and then I’m gonna have to like… You’re gonna say something, I’ll be like, “I said that? Oh, I must have said that to be nice.” That was not to be nice. It is a tool that I use to try to ground myself first and then also my team and hypothesise, and we should link to it in the show notes or whatever. It’s so good.

0:46:21.0 MH: Yeah, it is good. You know what Wil, I reference it all the time, too. So you’re not alone. Exactly. Alright, it’s that new segment. It’s time for the Conductrics quiz. Moe, are you ready to represent our listener in a battle to the quizzical death?

0:46:36.5 MK: Prize? To the prize? Yeah.

0:46:38.1 MH: To the prize. Tim, are you ready?

0:46:43.6 TW: Never. I will maintain.

0:46:45.7 MH: Well, Conductrics is always ready to help you elevate your A/B testing programs and experiments and run effective experimentation programs. They partner with you in ways that typical vendors seem to not, and that for over a decade, they’re the partner in A/B testing and personalisation the largest companies have teamed up with to work on A/B testing, contextual bandits, predictive targeting, always providing honest feedback and going beyond expectations to achieve testing experimentation goals. They are our sponsor for this segment. Let’s talk about who you’re representing. So, Moe, you’re representing someone who is a near and dear person to the podcast, and honestly the analytics industry here in the United States, Jared Riley Smith.

0:47:32.3 MK: It ain’t Jared?

0:47:35.8 MH: Yeah. And Tim, you’re representing someone who I think you know, and I’m gonna mispronounce his last name ’cause I always do. Jim Jinoleo.

0:47:43.0 TW: Jim Jinoleo?

0:47:44.4 MH: Jinoleo. See, there’s a lot of extra vowels, and I just… Sorry, Jim. Anyways, so let’s get into it. Here is the question. It’s probably fair to say that a lot of the people who call themselves data scientists these days favor Bayesian methods. They’re easily spotted in the wild, conversing about updating their priors after learning something new, rather than just saying, I learned something new. Anyway, interestingly, well, Thomas Bayes is credited with the idea of Bayes Theorem. Much of the early formalisation of what we think of as Bayesian statistics was done by Pierre-Simon Laplace in the late 1700s. That’s not even the question.

0:48:28.4 MH: Alright, so here’s the question. Moe and I are at some analytics conference maybe in Budapest, and we overhear someone talking knowingly in a group of people near us. It is Tim. We hear him say in a voice loud enough so that everyone in earshot can hear. “Well, I guess my throbbing headache is from that red wine we had last night, turns out my posterior probability of a headache after red wine isn’t close to what I figured it was last night,” I guess, and at this point, he raised his voice just a bit louder, to ensure everyone can hear. I don’t write the questions, folks. I need to update my priors going forward. What did Tim mean by posterior probability? Is it, [A], the conditional probability; [B], the total probability; [C], the marginal probability; [D], the joint probability; or [E], the naive probability?

0:49:32.5 TW: Oh, good Lord.

0:49:34.6 MH: Let me know if you need those again.

0:49:39.7 MK: I think the thing is, I know what they all mean separately. But in that language…

0:49:47.9 MH: It’s out of whack.

0:49:50.3 MK: It’s confusing somehow.

0:49:54.3 TW: Okay, right off with a marginal…

0:49:56.7 MH: Alright, so we wanna know what Tim meant by posterior probability. So, is it conditional probability, [A]?

0:50:02.5 TW: I said that loudly, I needed to update my posterior probability?

0:50:06.8 MH: Yeah, need to update my priors going forward, but what you said before was, “My throbbing headache was from the red wine. Turns out my posterior probability of a headache after red wine isn’t close to what I figured it was last night.”

0:50:19.4 TW: Okay.

0:50:20.6 MH: So, conditional probability, total probability, marginal probability, joint probability and naive probability.

0:50:28.5 TW: I’m down to…

0:50:30.0 MH: I think you could probably get rid of a couple of those, I’m guessing.

0:50:33.3 TW: I think we can eliminate naive, ’cause I think that’s a naive Bayes thing. I think naive is not. And I don’t think it’s… You wanna eliminate one Moe? Well, actually, can we start off with… Can we do that, the elimination? Am I correct that I can eliminate naive?

0:50:50.7 MH: Let’s say that can eliminate [E], the naive probability. It would not be naive to assume that.

0:50:58.0 TW: I think there’s another one I could eliminate, but do you wanna try to…

0:51:02.1 MK: The problem is I only listen to the ones that I thought it was, I didn’t listen to the ones that it wasn’t, so now I don’t know what the third one is. I only what the two is.

0:51:10.8 MH: The third one is the marginal probability. So, conditional…

0:51:12.7 MK: I know marginal and joint. What were the other two answers?

0:51:15.2 MH: Conditional probability, total probability, marginal and joint.

0:51:23.9 MK:: I’ve never heard total. I’m familiar with joint and conditional and marginal, but total I’ve not heard.

0:51:32.9 MH: You’re doing better than I would have done on this, Moe, so you’re cruising.

0:51:38.0 TW: Okay. So, we can eliminate… Let’s eliminate total. Let’s say we can eliminate total.

0:51:41.7 MH: Sure, let’s eliminate total.

0:51:45.2 TW: And then I want to eliminate joint, which means if I can’t eliminate it, then Moe wins. But I want to eliminate joint probability.

0:51:50.1 MH: You want to eliminate joint probability.

0:51:51.8 TW: Yeah.

0:51:52.8 MH: Oh my gosh, I’m not sure if I’m allowed to do that, ’cause that leaves you with two options.

0:51:57.0 TW: Right, and then Moe has to actually take the next shot at what she thinks… Which one she thinks it is.

0:52:01.7 MH: Oh, so that Moe will guess the next option? Are you sure you’re fulfilling your fiduciary responsibility to Jim Jinoleo?

0:52:10.7 TW: The fact is he knows what the correct answer is.

0:52:11.6 MH: Yeah, I was gonna say, Jim, totally does know the answer.

0:52:14.9 TW: He absolutely does.

0:52:16.9 MH: Jim’s like, “I wanna represent myself.” Alright, I’ll allow it. We’ll exclude… Which one did you say you wanted to exclude?

0:52:25.9 TW: Joint probability.

0:52:28.8 MH: Joint probability. Yep, so that one is gone. So, we have eliminated all but conditional and marginal probabilities.

0:52:35.6 TW: Moe?

0:52:37.2 MK: And so he’s calculating the posterior that he’s gonna be hungover after drinking red wine.

0:52:41.7 MH: He’s updating… Yeah, what did he mean by posterior probability? At this point, it’s a 50/50, you’ve got two options.

0:52:51.2 MK: I think it’s marginal, but I feel like a moron, because it seems really easy, but there isn’t another condition. It’s like he’s hungover or he’s not.

0:53:03.7 TW: Yeah. I don’t think it’s… Yeah, I don’t think it’s conditional…

0:53:06.8 MH: Well, it’s about the likelihood of being hungover, is what he’s trying to measure, I think.

0:53:13.5 TW: That’s the question, is whether it’s conditioned on the red wine or if it’s supposed to be conditioned on the something else, but…

0:53:18.4 MK: That’s why… Anyway, fine, I’ll go with marginal. Tim can go with conditional.

0:53:24.7 MH: Alright, Moe has selected [C], the marginal probability. Tim, are you comfortable going with [A]?

0:53:31.3 TW: I’m as uncomfortable as I was at the very beginning before I even knew this was coming from the mind of Matt Gershoff.

0:53:37.6 MH: Yeah, it does come from the mind of Matt Gershoff.

0:53:39.4 TW: Which is extremely nerd. So, yes.

0:53:41.1 MH: The correct answer is [A], the conditional probability.

0:53:46.1 MK: You knew it.

0:53:47.9 MH: Yes, one can use the law of total probability and the multiplication role to relate the conditional joint and marginal probabilities to one another.

0:54:00.0 TW: Oh, well, almost, Moe.

0:54:02.0 MK: Tim, can you explain it?

0:54:03.7 TW: So, this is where you’ve got a P and a H and a something else, and you’re actually combining them. It’s how you’re looping back into the beginning that… I’m still not exactly sure what the condition is, but I think that’s how you’re actually closing the loop. And maybe the total probability would cover all conditions, I don’t know. The more I talk… This is bringing up… I actually was reading a book that was the history of Bayes’ Theorem, the most underrepresented…

0:54:41.6 MK: Of course you were.

0:54:43.7 TW: This was a couple of years ago, I learned nothing about Bayes, but I remember Laplace being captured in that, and…

[overlapping conversation]

0:54:47.3 MH: I’ve heard that word as well before, Laplace. And now is the time to wrap up the Conductrics quiz. If you wanna learn more about Conductrics, go check out their website at conductrics.com, and see how they might be able to help you with your A/B testing and experimenting programs. They can definitely help you with some of your statistics. Congratulations, Jim, you’re the winner. Jared, I’m sorry, you didn’t prevail this time, but keep submitting the sticker requests or reach out to us on Twitter if you wanna be entered for a chance to win, and we’ll keep doing the quiz, and we’ll keep having…

0:55:26.7 MH: I have to update my priors, frankly, every time we do a quiz because we keep having to change. Anyway, let’s get back to the show. Alright, we do have to start to wrap up. Although, I feel like we could just start a whole second show at this point of just being like, let’s tell stories and have a fun camp fire. But this is the way the world works. So, we do have to start to wrap up, but really great discussion. It probably is worth noting, if you’ve been listening and been hearing what kind of person Wil is and things like that, you could go check out their website. They are actually hiring for a lot of roles, so not for nothing, but I figured I’d give that little plug, a lot of people listen to this show, and frankly, find a lot of worst places to work than Seer, if you just take my word for it.

0:56:11.8 MH: Alright, one thing we do on the show is a last call, and that is just something we go around the horn and share with our listener audience, something we find might be interesting for them or for their community. So Wil, you’re our guest. Do you wanna share a last call?

0:56:29.4 WR: I got two. So they’re not last calls.

0:56:31.4 MH: Alright.

0:56:31.9 WR: I heard that’s cool, I heard Tim does this often.

0:56:33.0 MH: Yeah, Tim does this all the time, so totally allowed.

0:56:36.9 WR: Most impactful book I’ve read as an analyst is a book called ‘Mistakes Were Made, But Not By Me.’ A really good book.

0:56:46.1 MK: Interesting.

0:56:46.8 WR: What it does is it shows how we trick ourselves into believing data because we’re going into it with a bias, so my favorite example in there is around a doctor called Dr. Semmelweis in 1850, he determined that we should wash our hands when we do procedures. And he found that out by realizing that women were dying during childbirth in the 1850s, and the ones that were, were from students who were working on cadavers and then going and delivering babies and not washing their hands in-between.

0:57:18.6 WR: And he presented that information to the Vienna Medical Society. They fought him for 15 years on that data. And all he was saying was, “If we washed our hands, we would have saved women’s lives.” And what we learned from that is doctors would have had to admit to themselves that they actually had killed women by not washing their hands, so it was easier to fight the data, and this book is full of examples where people are like, “DNA evidence, prosecute,” “This DNA evidence found out that this guy is innocent. Nah, keep him in jail.” And you’re like, “How does that happen? How do you use the exact same thing and exact same science to come up with different outcomes based on what your bias is telling you?”

0:57:56.7 WR: And that book just really makes you look internally, so that’s one thing. And the other thing is in November, around November, I think, 15th or so, I’ll be doing a sleep-out for homeless youth in Philadelphia. And I think we’ve raised probably over the last 10 years I’ve been doing this around $150,000, $160,000, but I’m gonna do it again. Last year, I raised 43, and now that I deleted my Facebook, I’m gonna have less reach, so I’ll use this last call to tell people “If you follow me, please follow me at least through November so that you can donate when I put my links up,” because I’ve literally listened to and worked with kids who’ve been kicked out of their homes in the winter because they’re gay or LGBTQ, or their parents don’t like them. And they have nowhere to go, nowhere to go.

0:58:44.0 WR: And I think that’s a problem that we should all want to be a part of solving. So, I would love support from that, if anybody is willing to support me, ’cause I’ll be tweeting about it a lot when we get closer to November. Thank you.

0:58:53.1 MH: Outstanding. Wow, thank you. Moe, what about you? What’s your last call?

0:58:58.9 MK: So, a friend shared with me a blog post from the Farnam Street, which I mean, they have like a bunch of articles and stuff like that that I’ve been kind of trolling through. This is like… I don’t know, it’s a bit weird. But there’s a Nobel Prize-winning biologist, I think he won it in the ’70s, Peter Medawar. And he wrote a book called ‘Advice to a Young Scientist,’ which I don’t recommend reading ’cause it sounds really long, but the Farnam Street blog did a really nice kind of “Here are the key concepts of the book,” which I don’t know, it just… It kinda made me feel good and kinda got me a bit in touch with, I suppose, some of the values and the moral dilemmas that you come up against, and I don’t know. It’s just… A friend shared it for me as a bit of a light read and I just kinda enjoyed it, and yeah, it reminded me of my younger self and my curiosity and… Yeah, anyway. That’s it.

0:59:53.6 TW: I’d like to recommend ‘Infinite Jest’ by David Foster Wallace. No, I’m kidding. Definitely don’t read that, either.

1:00:00.8 MH: Tim, what about you?

1:00:03.1 TW: So I’m gonna do kinda two half last calls. One, Wil, it is always fun to talk to you because of your energy and thinking. So I’m gonna do a last call that literally just Googling and “Wil,” and Wil’s with one L, Wil Reynolds, and then clicking on the videos and then just randomly watching one, there’s stuff that goes back years that are always thought-provoking and fun and entertaining. It is why I could sit for four hours and listen to Wil talk. And then the second, I’ll just be silly. So we have this new branding, and we do have a store, and it’s a Zazzle store, ’cause we put no research into it like five years ago, so it’s not a great user experience, it’s not great pricing, but we do have all the branded merch updated. So just think. We’re probably at least 12 months out before realizing that we always hated the current branding, [laughter] so now is the time to get your analytics and get the maximum life out of the currently branded merch. There’s some tiny little percentage goes to my daughter’s college fund. We have yet to hit the $100 minimum of those to even get that initial payout going back on that, but it’s there. I’m drinking out of my Analytics Power Hour coffee mug, which, as we’re recording…

1:01:29.7 TW: So check that out. Or, for free, there is… We do have stickers. I think we mentioned this on the last show as well. Those are totally free, if you go to bitly/aph-stickers or there’s a link in the FAQ on analyticshour.io. So get your merch, help us spread the word about the podcast, and also, absolutely feel free to tweet at us with miserable aspects of your e-commerce experience, because frankly, we’re not making money with our merch. So ripping an e-commerce experience would be entertaining, and we’d address anything that we can, but it won’t be much ’cause it’s just a Zazzle store. What about you, Michael? Top that. I set the bar low.

1:02:13.2 MH: I will top. Well, I’ll do my best. So I actually wanted to share, just really briefly, the thing that actually made me reach out to you, Wil, in person, a few months back in the first place, because… Obviously, I’ve known about you for a long time. We both know people who… But you and I have never spoken up until a few months ago, and the reason I emailed you in the first place was a couple of articles you wrote about what does it mean to live a life of enough, as an agency owner, as someone who has had a lot of success? And you spell it out in pretty good detail in two blog posts about what you’re doing and why you’re doing it, and that was what made me finally say, “I need to talk to this guy because I’m trying to, myself, build a company, and what wisdom could I get from someone who’s already done most of this journey and could maybe share with me things that… ” And so that was actually what did it, but if you’re someone who’s interested in any aspects of business and entrepreneurship or being a CEO, I highly recommend those two articles to you, so… Sorry, Wil, you’d put up with that, but that’s just… You write this stuff.

1:03:26.0 WR: Thank you.

1:03:26.2 MH: You put it out there so…

1:03:26.8 TW: Unless you really wanna be a Silicon Valley tech bro, and then Wil is not gonna give you… Not gonna be your guiding light on that front.

1:03:34.9 MH: No.

1:03:35.0 WR: Daddy needs a fifth Porsche at his fourth house, so why don’t you work 80 hours this week, buddy?

1:03:42.9 MH: That’s the… Yes, so that was the first one. Second one is, in about a month, we are going to be having the Digital Analytics Association One conference, so it’s our annual conference. Last year, because of COVID, obviously, it was all virtual. This year, who knows, ’cause when we’re recording this versus a month from now when this comes out, what’s gonna be the case, but right now we’re planning to be in person.

1:04:07.5 MH: This is actually more of a personal thing. I’m planning on being there, and I would love to see you there, ’cause I just love hanging out with analytics people, and we’ll do all the right things and wear masks and be vaccinated and all those things, but if you can make it to the OneConference, I think you will love the speaker lineup this year, and I think you’ll have a great time, so…

1:04:31.8 TW: I’ll be there, and I’ll be ranting about the thing that Wil, and I just went back and forth about a little bit.

1:04:35.1 MH: Yeah, there you go.

1:04:38.5 TW: I will absolutely use the word ‘farce’ as part of my talk.

1:04:40.1 MH: There you go. Perfect. Alright. Well, this has been a great discussion. Obviously, we’d love to hear from you. We’re available out on the Measure Slack group, on Twitter, or our LinkedIn group. So please reach out to us. You can actually order stickers for free, so you don’t have to go to our store, so you do it at Bitly, B-I-T.L-Y/aph/stickers, and then you can order some stickers, and we can send those out to you, and you can see on Twitter sometimes people will… And they’re actually getting pretty good times through to Europe, ’cause we’ve had a few people get them in about a week, so that’s pretty amazing. That’s that Tim Wilson, all… Just Tim Wilson, in the US…

1:05:20.8 TW: Working the warehouse.

1:05:22.4 MH: At the Postal Service.

1:05:23.1 TW: My weekend job.

1:05:24.3 MH: The fulfillment center.

1:05:24.8 TW: Fulfillment, yeah.

[laughter]

1:05:26.8 MH: Anyways, we appreciate all that you do on the… Tim, and obviously, we also appreciate our producer, Josh Crowhurst. This is going so smoothly. Moe is like, “Why, why?” Anyway, we’d love to hear from you. And remember, it doesn’t matter if you’re the CEO or you’re the 20-year-old kid who just learned about Tableau, the one thing you can do is you can be a hypothesis assassin like Wil Reynolds, and keep analyzing.

[music]

1:06:00.7 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.

1:06:17.9 Charles Barkley: So smart guys wanted to fit in, so they made up a term called analytics; analytics don’t work.

1:06:23.2 Thom Hammerschmidt: Analytics. Oh my God, what the fuck does that even mean?

1:06:33.3 MH: This is turning… We’re on a record pace here, Tim’s been flipped off twice already. It’s pretty much started…

[laughter]

1:06:43.8 TW: Rock, flag and hypothesis assassins. [laughter] That’s actually a nice Philly…

1:06:53.1 MH: Oh, yeah, that’s a Philly reference.

1:06:53.2 TW: Yeah, I hadn’t thought about that. That’s a whole… “It’s Always Sunny in Philadelphia” thing.

[laughter]

1:07:05.4 MH: You guys are great.

1:07:05.5 MK: I can’t tell if you’re in that “If you don’t laugh, you’ll cry” phase or if you’re genuinely laughing.

[laughter]

1:07:09.9 S6: No, that was horrible, is was what it was. [laughter] That’s what that was. I’m embarrassed for you.

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

#181: Qualitative + Quantitative = A Meet Cute for the Ages with Jenni Bruckman

https://media.blubrry.com/the_digital_analytics_power/p/traffic.libsyn.com/analyticshour/APH_-_Episode_181_-_Qualitative__Quantitative__A_Meet_Cute_for_the_Ages_with_Jenni_Bruckman.mp3Podcast: Download | EmbedSubscribe: Google Podcasts | RSSTweetShareShareEmail0 Shares