Three Letters at the Root of Analytical Angst"> Three Letters at the Root of Analytical Angst">

#187: K.P.I. -> Three Letters at the Root of Analytical Angst

How do we measure the performance of this podcast? With well-formulated KPIs, of course! With targets set for them. Since Tim is the taskmaster who insists we revisit our KPIs every year, we decided he would be our guest for this show, and Michael and Moe would take turns trying to stump him with impromptu role playing as difficult stakeholders in challenging scenarios.

Articles & Analysts Mentioned in the Show

Photo by Afif Kusuma 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.

[music]

0:00:21.9 Michael Helbling: Hey, everyone, it’s the Analytics Power Hour. And this is Episode 187. Hey Moe, 187 episodes. You know what that means in America?

0:00:34.7 Moe Kiss: I don’t know. That it’s a law?

[chuckle]

0:00:36.4 MH: It’s a police code for a murder, 187.

0:00:39.8 MK: Really?

0:00:40.6 MH: Yeah.

0:00:41.0 MK: Ooh, that makes today’s topic interesting.

[laughter]

0:00:45.8 MH: Well, let’s jump into that, [laughter] ’cause we’re gonna murder this podcast episode.

[chuckle]

0:00:52.7 MH: That’s right, we’ve decided to do a hard pivot to true crime, here on the Analytics Power Hour. [laughter] We’ve heard that’s a popular format, so…

0:01:02.9 Tim Wilson: We’ve also been acquired by Spotify and…

0:01:04.8 MH: Yeah. And here’s why we’re doing this hard pivot, because we’ve been looking at our numbers and we’ve gotta get those KPIs up, key performance indicators, that’s what I’m talking about. Alright, no, but seriously, it is one of the most important parts of analytics, you take all this data that we pull together and collect it and put it into the hands of users in meaningful ways, and since the beginning of time, those three little words have sort of bedeviled us, key performance indicator, yes, KPIs. Here’s the deal, it seems on the surface to be really simple and yet let’s face it, most of the time most companies or organizations just process and the resulting KPIs are just complete garbage. “Why can’t we get this right?” Well, we, and by we I mean Moe and I, are kind of interviewing a special guest on this ultimate KPI episode. That’s right. Tim Wilson is the Senior Director of Analytics at Search Discovery. Prior to that, he was a senior partner at Analytics Demystified and has held numerous senior analytics roles in his long and noble career. He’s an internationally-known speaker and podcaster…

[laughter]

0:02:16.0 MH: And a previous winner of the illustrious DAA President’s Award. When he’s not solving the most complex analytics problems he is an avid photographer, and he makes a mean fajita. Most importantly, today he is our guest, welcome to the show, Tim.

0:02:31.9 TW: Oh. Thank you for having me. I’m so excited to be on the Analytics Power Hour.

[laughter]

0:02:40.5 MH: That entire… Well, it occurs to me that I don’t think we’ve ever turned the mic on you, I think we’ve done it to everyone else here. I think we…

0:02:47.6 MK: No, we’ve done it to me, multiple times.

0:02:50.4 MH: Yeah. Well, Moe, you started as a guest, so then you were so good we’re like, “Well, let’s… ”

[chuckle]

0:02:55.9 MH: So now, on this…

0:02:57.3 TW: We may go in the other direction on this one, “And also, it’s Tim’s last episode.”

[laughter]

0:03:01.9 MH: That’s right. Great to have you. That’s why it’s 187, we’re killing Tim off. No, I’m just kidding. No, but I got through that whole intro, Tim, without calling you the “Quintessential Analyst.” Aren’t you proud of me?

0:03:13.7 TW: Well, you just ruined it, so…

0:03:14.0 MH: But unfortunately… Well, it’s one of my KPIs for the shows to make sure to mention it in each episode, so…

0:03:20.5 TW: That was not agreed to in annual planning.

0:03:23.1 MH: Okay, alright. So Moe, I think… Do we give you credit for this one? I think this idea kind of started with you, so why don’t you kick us off?

0:03:32.5 MK: Well, I guess one of the things that’s coming up a lot as we grow our team is we have lots of junior analysts, and there seems to be a skill, it’s like coaching stakeholders through how to create, formulate the right KPIs for whatever it is they’re trying to measure, and I was like, “Oh, I should do some training on this,” and I was like, “Oh, Tim’s done some training on this, innit?” One thing led to another, and so basically, I’m gonna steal some ideas from him, and then…

0:04:01.1 TW: When I think about it, I actually did a training at one of your former employers, just ’cause I was in town, you’re like, “Hey… ”

0:04:07.8 MK: Yes.

0:04:08.6 TW: “Why don’t you come be the least fashionable person in this entire building and talk data?”

0:04:13.5 MK: It was a fashion company, to be fair.

0:04:15.3 TW: But you did say, “Talk data.”

0:04:16.4 MK: Okay, so let’s get started. So, you have a meeting with your stakeholders and they’re like, “Okay, Tim, we’re going to run the most amazing campaign. It is cutting-edge, it’s exciting, we’re on 10 different social media platforms, and here are the five metrics we care about.” How do you handle it, Tim?

0:04:40.2 TW: Well, it depends on what those metrics are. [chuckle] I rarely start at that level, it’s usually they don’t volunteer the metrics, they say, “We’re gonna run it… ” Or, we all run into the like, “We ran this campaign, can you tell us how it did?” So, I feel like the… Well, even in that exact scenario, I’d be like, “Okay, can we pause for a minute? Can you just tell me in plain English what it is you’re trying to achieve with this campaign?” And if they say, “Reach or impressions or exceeding the click-through rate of”, then I’ll kind of try to gently stop them and say, “No, no, no, no, let’s not talk about the data, like you’re about to put 50k or half a million K or two million… ” Let’s just say half a million K, that’s good.

0:05:27.7 MK: Yep, yep. I saw that. [laughter]

0:05:28.7 TW: That would be somewhere in the trillions of dollars.

0:05:32.4 MH: Half a million K…

0:05:33.5 TW: So I start off by undermining my own credibility by not being able to speak numbers. So I try to get them to just very, very tightly articulate like, why are they doing this? ‘Cause a lot of times I feel like, it’s like, “Oh, we ran this last year, we’re gonna run it again,” or somebody somewhere decided it was worth spending money to do this thing, whether it’s a campaign or a website re-design, and a lot of times, no one’s really captured and articulated that. So…

0:06:05.2 MK: So are you talking about an objective?

0:06:08.9 TW: Yes, but I usually don’t use the word “objective” because if you asked… I literally watched a strategist and an account manager argue about whether a sentence was an objective or a strategy. So, I kind of… And I’ll see that… You’ll see that in creative briefs where it’ll say “objective”. And to me, objective means 10 different things to 10 different people, so I use… And this is sort of the… Anybody who’s ever heard me speak has probably heard me talk about the two magic questions, which I branded, I totally lifted them from a guy named Matt Cohen. If we just say, “What is it you’re trying to achieve that… And give an answer that the business two or three levels up in the organization would nod their head and say, ‘Yeah, that matters?” ‘Cause if you say objective or you’ll see it in a creative brief and there’ll just be words filled in, it’s like, “Oh, I need to fill in the objective box.” Whereas if you say, “No, what is it you’re really trying to achieve?” which some people would say that is an objective, but I think others would say, “Oh, that’s your strategy.” So that’s kind of why I just feel like the word gets mucked around like that.

0:07:17.1 MK: Okay. So, belligerent stakeholder who is kind of difficult to work with, you say, “What is it you’re trying to achieve?” And they come back with like, “Well, we want new active users, we wanna get returning users to come back and we wanna make sure that we get revenue… ”

0:07:34.3 TW: Yeah, which is… I guess that’s the thing, there is the… You could answer that question, and you can use data, it’s not like you can’t put a metric in it, and if you said, “Okay, so new active users… ” If I’m in a company where MAU, I mean monthly active users for… I’ve never worked in-house at a company where that’s a big deal, but there are certainly lots of organizations, and might even be people on this podcast where that’s a big deal.

0:08:00.2 MK: Number one company… Well, one or two.

0:08:01.6 TW: Yeah. And if I understand, if I already understand like, “Yes, everybody knows that’s the thing, that’s how we’re getting funded, or that’s how we make revenue,” then we move forward. If I’m like, “I know we say that, but can you help me understand exactly… Just help me understand.” So I will try to turn it around to say, “I’m trying to understand your business, how does monthly active users… ” And not that there aren’t times where then somebody gets super pedantic and explains like how math works and talks down to me, and I swallow my pride and I let them do that, but then continue to probe until I really understand. I, as an analyst, wanna make sure I understand why whatever their answers are, how that relates to business value, which the flip side is that when they say… So that’s like part one, and if I’m like, “Okay, this is a good metric,” but I will definitely say… If you say monthly… Monthly active users is definitionally not a KPI. It’s not a KPI until you say, “How many monthly active users?” And that’s the other place that organizations fall down, “Revenue, our KPI is revenue.” I was like, “Do you think you’ve fully defined a KPI when you said revenue?” “Absolutely, we’re in the business of making money, it’s revenue.”

0:09:19.4 TW: I’m gonna go ahead and predict you’re gonna make one dollar, there’s your KPI, you’re not measuring performance unless you actually set a target for it, so… I don’t know. It is… You’re forcing me into the actual real-world seeming kind of organic discussion without being formal role-playing. But I will run into someone saying, “Well, we’re trying to drive awareness.” It’s like, “Okay, I get that awareness. Let’s understand why awareness matters. Let’s actually make sure that you’re actually trying to drive awareness,” and if you’re Nike and saying you’re trying to drive awareness, you’re not trying to drive awareness of Nike. Are you trying to drive awareness of a new product line? What are you trying to… ‘Cause awareness… Engagement’s a tougher one, or I guess an easier one to knock down. There’s no inherent business value in engagement, let’s probe on that a little more and say… Which I guess you could go down the same path with monthly active users. There’s a whole thing where you get into parsing, like, how do we define active? But I feel like that’s a whole other episode of the show is, do we have the right definitions?

0:10:33.6 MK: So we were talking… Like I said, I’m trying to think it out, like how do I develop a training for some of the junior analysts around developing some of these skills? And we were… A couple of the leads in the data space were talking about it, and someone came back to the five whys, and it’s actually not a technique I use myself very often but I have used before. But it sounds, Tim, when you’re in these stakeholder meetings, you were just focused on asking a shit ton of questions. Is that…

0:11:06.7 TW: Yeah, I pretty quickly… I feel like more often than not, I find myself in situations where I’m asking questions where they don’t actually have good answers. It is very easy to go down that exercise and very quickly uncover that they don’t have clarity around what they’re doing and why they’re doing it. Conceptually, I like the five whys, it’s just… It’s another one that it’s easy to kind of punt on, but making that logical link, it’s the same concept as saying, “You’re saying you wanna do X, but have you thought about what this campaign is doing that is going to tie to business value?” So, I don’t know. I feel like the five whys gets… I guess for anyone who doesn’t know the five why… It’s W-H-Y, not the letter Y. And you just keep asking why with the hope that you’re getting a link to, “Why does that matter? Well, why does that matter? Well, why does that… ” That’s my interpretation of the five whys, I guess.

0:12:05.6 MH: And I feel like anyone with a toddler is just like, “Nope, sorry, I can’t do that today… ”

[laughter]

0:12:11.9 MK: So true.

0:12:14.2 TW: But I think there are times where it’s… The KPIs are getting listed… If I’m talking to… If I’m trying to teach younger junior analysts, as you kindly pointed out, I’ve been senior now for a long time, I’m super senior, that there’s this, “What is the purpose of a KPI?” And the purpose of a KPI to me, which I don’t know that I can stand up and say, I have any potential blessing that says this is right, it’s just, to me it’s very clear and simple to say, it’s not there to give you insights about the business. It is not there to tell you what you should do. If you focus on a KPI and say, “Really, that’s just about objectively measuring whether you are meeting your expectations.” So, a lot of times I find myself doing training with clients where I talk about it as building a time machine, that you’re setting up your KPIs because if anybody who’s been in marketing for more than six months has seen the case of looking backwards, looking at metrics and then getting the question, “Oh, we drove a 1,000 form fills.” And the question is, “Well, is that good?”

0:13:35.5 TW: And so, the only way you can determine if it’s good or not is if you answer the question, “Well, what did we expect to have?” And the only way you can answer what your expectations were is if back in time, you looked ahead and said, “What do we expect to have happen?” So I have a little simple, little diagram that is probably on a blog post somewhere, where I try to help the business understand that it’s… You wanna blow past this ’cause I’m not gonna give you a report right now with the results of your campaign because you haven’t executed it yet, but if I can help you be… You understand that you are going to ask me whether these results were good or bad at some point in the future, and if we don’t have this conversation now, I will not be able to answer that question, and it will get ugly.

0:14:27.1 MK: Well, this conversation could get ugly because I’m about to… I don’t know. Okay. So there’s this thing that happens when… It’s normally when you’re dealing with a stakeholder that’s like… I’m data-driven, and I still have this internal fight that’s going on where they basically are like, “Okay, we’ll just set a KPI, I need an analyst to do a whole bunch of work to basically forecast what we expect is gonna happen, yada yada yada yada.” And often then, the KPI they come back with is actually just what we forecast, and I’m like, “No, but forecast is what we expect is gonna happen without doing this thing, so the KPI needs to be bigger than that.” But also, the idea that every single KPI that gets set by every business stakeholder is going to need an analyst to do a whole bunch of work, like hours of investigative work before KPI can be set actually drives me insane, even though I am a data analyst who should probably be excited that people wanna use data. And it becomes this chicken and egg situation, where sometimes I say to analysts, I’m like, “Just throw a number out there, and see what they say.” And I don’t know if Tim’s gonna murder me for this.

0:15:41.9 TW: No, I… The way… And I was… I get to work with Val Kroll, who is amazing, and so I’m gonna give credit to her for this, ’cause we were in the midst of this with a client that we’ve done a lot of different training and coaching and helping them with their processes, great client, but some kind of stakeholders who were not used to thinking this way, and Val, just like once we were talking to some of the stakeholders and she said, “What result would put a smile on the face of your stakeholders?” ‘Cause these were marketers who were then kind of beholden to… With some people who were in another part of the business, and she said, “What result would put a smile on the face of those people?” And it’s not like those people are gonna… They don’t have the context of all of this, would not have had the context of asking an analyst to go and do all this digging to say, “Should I be excited about this?”

0:16:39.2 TW: The fact is, if I’m gonna spend a million dollars US on a house, do I need somebody to go into a bunch of research as to whether or not that’s a worthwhile expenditure of a million dollars to get a two-room shack in a horrible neighborhood? No, I just have in my head, “I’m gonna spend a million dollars, this is what I expect.” It’s not to say that there can’t be some like looking at that historical data, but you’re nailing… I think this happens to us all the time, it’s really… We’ve been so conditioned to saying the historical data is deterministic, it’s available, and it’s just the facts, and as soon as we ask them to think about the future, of which they are, whether they fully admit it or not, do have some control over the results in the future, they get super uncomfortable. And recognizing that, okay, this is an uncomfortable space, and then they’ll say, “Well, we want the analyst to go and figure out what we should do or what we should expect.”

0:17:44.3 TW: And I think as an analyst that is like, put on your soft skills, your patience, and say, “Okay, I can do that,” but the goal going through life is not to just be average or a little bit above average, that’s not a way to do something. I have another client and he just, he hates benchmarks, and he’s got some funny little quirks, but I absolutely love the fact that he, as a senior marketer, will just rip members of his team if they show up and say that they exceeded a benchmark, he was like, “Really? Is that the… ” Or they hit the benchmark, he’s like, “Really? That’s what we’re aiming, to just meet our own past performance? So, is that what you’re gonna do next year? Do you wanna… So you’re incentivized to do mediocre work so that you’re not raising the bar for yourself, like what the hell?”

[chuckle]

0:18:41.0 MK: I can totally see why Tim likes this guy.

[laughter]

0:18:44.9 MK: While Helbs is cringing over there.

[chuckle]

0:18:48.0 MH: No, I feel the same way, I just don’t say it that way.

0:18:51.7 TW: I mean, that’s… But that’s a lot of times what they’re asking the analyst for, and it’s the same reason that media agencies will… That’s their favorite, it’s what they’ll… They wanna set KPIs that are not volume-based numbers, they wanna do them on rates and ratios so they can say, “Oh, we’ve looked at… It doesn’t matter if you had a million impressions, but the click-through rate was 0.02%, well, if we’re gonna have half a million impressions, then we wanna be better than 0.02%,” or whatever, ridiculous click-through rate, and that’s them trying to punt and not really thinking about what they’re trying to do.

0:19:28.4 MK: Okay. You’re gonna have to delve into that a little bit more about ratios and rates, versus total numbers. Walk us through that a bit…

0:19:37.5 TW: Okay.

0:19:38.2 MK: A bit more slowly.

[laughter]

0:19:40.0 TW: Okay.

0:19:40.2 MH: Yeah, that’s good.

0:19:41.6 TW: So I like to say, we’re gonna see how many soap boxes, tried and true, I can mount. There is never a… Almost never, I don’t think there’s ever inherent business value in a ratio or a rate, because you’re combining two metrics, definitionally you’ve got a numerator and the denominator. And the fact is, if I’m looking to drive quality leads, and I’m looking at my conversion rate of my website and doing leads, let’s go simple, and I wanna really focus on the conversion rate, and say, “Wow, our conversion rate was… Our lead conversion rate was higher,” but it’s like the web traffic dropped by a quarter and you increased… And the sales team is not saying, “Oh, that’s awesome, you got more efficiency on your website.” I’m not saying conversion rate doesn’t matter, I’m just saying at the end of the day, it’s how many leads did you deliver?

0:20:43.4 TW: Now, one way to deliver more leads is to convert a higher percentage of them, but that’s a means to an end. When it comes to I’m doing something, the more I can say I’m spending, ’cause frankly, if I’m spending $50,000, I’ll just keep making up random numbers, and my lead conversion rate goes way down, but my total leads are actually higher because I just drove an enormous amount of traffic very, very cheaply. Am I gonna be happy? Yeah, but not if somebody is saying, “Oh man, I don’t know, we spent all that, we got that shit traffic, the lead conversion rate was down.” I’m like, “Yeah, but I delivered the number of leads you wanted. Are you sweating the server cost for the incremental traffic the site had to deal with?”

0:21:30.6 MK: Okay, annoying stakeholder. But we only wanna drive quality leads.

0:21:35.7 TW: Then, oh, cool, how many… How many quality leads do you wanna drive and how are we measuring… How are we defining a quality lead?

0:21:43.1 MH: Yeah, ’cause again, you can go into like, once you get a lead, then it’s a marketing qualified lead and do a sales qualified lead, sort of a funnel…

0:21:50.6 TW: No. Gotta get sales accepted lead and then a sales qualified lead.

0:21:53.1 MH: Oh, I’m sorry. I don’t do this…

0:21:53.5 MK: Oh, fuck, guys.

[laughter]

0:21:56.9 MH: Sales accepted lead. My bad.

0:22:00.8 MK: Okay. But Tim, just to give you a real-life example. So for example, an iOS user, let’s say it costs us, I’m gonna throw out a number, $3 to acquire, but iOS users are far more likely to upgrade to a trial, convert, blah, blah, blah, whether convert means a trial or become a monthly active user. An Android user costs us 10¢ to acquire, and we acquire a gazillion thousand more of them, but very, very little likelihood of upgrading to a trial, staying on a paid plan or becoming an active user.

0:22:39.1 MH: I think it’s just a mixed question, right?

0:22:41.6 TW: Well, I don’t even know that it’s a mix so much as a math. If that is down to the point where you can say, an Android… An Apple… An iOS user is basically 10 times as valuable as an Android user, then say, “We wanna drive… ” And it’s putting aside that, if there are costs to service the unpaid, the Android, whatever. But if you basically come up with, okay, then, what you really want is to get long-term business or lifetime value or paid subscribers, we’re not gonna wait all the way down there, we’re gonna back into it with some back-of-the-napkin math and say, “We wanna get 10,000 iOS equivalent leads, which is either one iOS or for every 10, and we’re gonna come up with a formula and have that discussion,” ’cause you actually… What’s happened is you’ve gotten to the point… You just did it.

0:23:34.6 TW: This is where you’re really bad at being a crappy stakeholder, ’cause you’re articulating that, really what we want is to drive paid subscribers, and then you’ve already taken the step to say, “Well, backing into that, we know that this quality differs,” but we don’t wanna target a mix either, likely, it’s like targeting new versus… That’s always a great one. We have a KPI of new versus returning, like, “A mix? What the fuck is that?” We can manipulate that all over the place, “Oh, ultimately, I want new leads or new users,” and new versus returning may factor into that, but that’s kind of how you get there, you’re trying to ultimately get to new subscribers back up, and you say, a quality lead… I’ve talked about… I’ve done quality visits or an engaged visit, and it’s like that is… To me, I’m okay with that, if we say, “Let’s… ”

0:24:37.2 TW: What we want is to get human beings who are engaging with our content, because if they consume our story in a meaningful way, we feel like that’s about as much as marketing can do, and we wanna make sure we’re not just sending total crap traffic, so let’s sit down and come up with a criteria we’re gonna agree to, that’s gonna be a much, much smaller, some sort of arbitrarily defined, and you’re not gonna necessarily have the historical data, and they’re gonna say, “Well, can you, the analyst, go pull what we’ve historically gotten?” And there’s a degree of saying, “Sure, we don’t wanna define something that it’s gonna be impossible for you to get, but… ” I don’t know. It winds up going back to, “What is it you’re trying to do with this marketing thing? Let’s measure that,” which can then help them actually frame how they run the campaign. I don’t know. I’m just off on eight tangents at once.

0:25:33.1 MH: So what happens if it’s something that’s brand new to them though?

0:25:36.6 MK: Hmm. Now, I love this.

0:25:38.7 MH: Because for instance, you’re saying like, “Okay… ” But a brand is like, “We’re gonna try to do advertising on TikTok,” which is… a lot of companies were taking the plunge.

0:25:49.5 MK: Yeah. We’ve never done TikTok before, we can’t set a KPI.

0:25:53.3 MH: Yeah, how do we set… How do we start going down the right path from your perspective, Tim, on something like that?

0:26:03.0 TW: The “We’ve never done it before,” I think that, that is to me what’s under the… There’s a… What they believe is that they have no expectations whatsoever when it comes to the results, and I think they truly believe when they say, “We’ve never done this before, so I have no idea what good would look like.” What they’re saying, and this is… As an analyst, I will recognize that if that was absolutely true. If they have no expectations, then no matter what the result are, they’ll just shrug, if they say, “We have no idea what good is.” Now, again, they’re spending time and money on this. So, I can come up with a result that I think they would say is not good. So two ways that I will come at that, and one is harder to get them to buy into, so… Well, one is to say, “I know you think you have nothing, but you know what, we’ve got four people and I am the analyst, I’m one of them that really care about this campaign. On our next status call, we agreed that we wanna measure leads from TikTok. We feel like we can measure that, but we don’t know what to set as a target, so why don’t we all just go back and think about it and just using…

0:27:19.6 TW: None of us can spend more than 15 minutes doing this, but just come up… Pretend that we had a gun to our head and we had to come up with a target and justify it, and let’s all do that, and then we’ll come back.” And everybody said, “This is incredibly uncomfortable, I have no idea.” And you get four people, and all of a sudden three of them have numbers that are somewhere in the same ballpark, and somebody has one that’s way higher or way lower, and already you’ve sort of gotten everyone to realize that, “Oh, wait a minute, I do have some expectations. Because of how much this is costing us, I have an expectation,” or, “We haven’t done TikTok, but we did this… We did an Instagram Reels thing,” or, “I read something in Ad Age.” And so people do… So that’s one way to come at it, the other is to come at it from a… And I call it bracketing. You guys have both heard me talk about this.

0:28:16.9 MK: Oh, I fucking love bracketing. Every time I explain it to someone, they’re like, “That’s genius.” And to be honest, Tim, I don’t even know where you got it from.

0:28:26.1 TW: If somebody knows where I got it from, I may take personal credit for coming up with that…

0:28:30.7 MH: I assume we’re not talking about the naval warfare technique of firing your guns to figure out how to aim at enemy ships. That’s…

0:28:41.1 TW: Maybe not directly.

0:28:43.5 MH: Okay. Maybe historically, that’s…

0:28:43.9 TW: It’s a similar… You’re trying to get… You’re trying to hit your target. So…

0:28:46.8 MH: Yeah.

0:28:48.7 MK: Yeah. Interesting. And explain.

0:28:49.5 TW: Well, if the Navy started off by saying, “I’m gonna fire way off… ” If the Navy didn’t… If it wasn’t expensive to shoot in random directions. So bracketing is… And this is, it’s like there’s the formulaic way to explain it, and then there’s the reality if you can do it in conversation, conversationally…

0:29:07.1 MK: Tim, can you just do it with one of us? Use one of us as your stakeholder with bracketing, ’cause I feel like that will make it practical for someone to use it.

0:29:15.9 TW: We’re gonna do our TikTok campaign.

0:29:16.5 MK: Yeah, TikTok campaign.

0:29:17.9 TW: We’re trying to drive monthly active users. Okay.

0:29:20.7 MK: Yeah, can’t measure it. We have no idea. We’ve never done TikTok before.

0:29:23.8 MH: Mm-hmm.

0:29:24.7 TW: This is where I’m not gonna know whether or not you can easily see if somebody clicked through from TikTok. Let’s assume that you can, that you can somehow get a resource…

0:29:33.8 MH: TikTok just sent an email to all of their advertisers letting them know they’re enhancing their use of cookies, so…

0:29:40.1 TW: That’s good…

[laughter]

0:29:41.2 MH: I know, it’s like, “Wow, TikTok, read the room.” I don’t know. Now is the time. [chuckle]

0:29:46.7 TW: So we have to be clear, we shifted from identifying which metric we wanna use as a KPI to actually setting a target for the metrics.

0:29:53.5 MK: Totally.

0:29:54.5 TW: So we’ll say that we’re trying to look at new sign-ups from TikTok and they just can’t set a number.

0:30:00.3 MK: Nice.

0:30:01.5 TW: And what are we spending on this TikTok campaign? I don’t know, 50 grand.

0:30:05.3 MK: $80,000.

0:30:06.7 TW: $80,000.

0:30:07.1 MH: 80 grand. Wow.

0:30:08.3 TW: Maybe… Maybe Euro dollars, maybe US dollars, maybe Australian, who knows?

0:30:13.3 MH: Maybe Canadian.

0:30:14.5 TW: So we’re spending 80,000 over the course of a month, and you have no idea, it’s like, okay, so you’re spending $80,000 and this is a service that maybe costs 20 bucks… Maybe 60 bucks a year, I’m just trying to get a sense of the business, we’ve got… Okay. If you’re spending $80,000 and you’re trying to drive new sign-ups, would you be comfortable if… Let’s just say that we got 20 new users for $80,000, would that be cool? You okay with that?

0:30:44.0 MK: Hmm. I think that’d be like a fucking disaster.

[laughter]

0:30:48.7 MK: My boss would be pissed if we only got 20 sign-ups.

0:30:50.5 TW: Exactly. Okay, cool. So that’s…

0:30:52.1 MH: Okay.

0:30:52.7 TW: Yeah, I kind of agree with you. So for $80,000, what if we actually got like 100,000 new users, how would your boss… Would that put a smile on your boss’s face, like less than a buck?

0:31:03.4 MK: Shit. I mean…

0:31:04.7 MH: Yeah, that’d be amazing.

0:31:06.4 MK: She’d be really happy, but that’s huge.

0:31:11.5 MH: Just on the fly roleplay.

0:31:11.7 MK: We didn’t get anywhere near that last month on Instagram.

0:31:16.3 TW: Oh. What did you get… Do you remember what the cost per acquisition on Instagram was?

0:31:21.7 MK: Oh…

0:31:22.4 TW: Now I’m shifting out of bracketing.

0:31:23.5 MK: I know. I love that you’re totally going to a different metric.

0:31:26.7 TW: Well, no, but I mean… Well, but… But I guess that’s the… Well, we don’t need to walk through the… Okay, so we’ve said more than 20 or 80, whatever the bottom one was, less than 100,000, so… Okay, let’s just put the 80,000, 20 is way too low. What about a 1,000? And already… You did sort of throw out… I don’t know, I don’t wanna walk through the… It could go on for a bit. They’ll figure out what you’re doing is to saying, “We set one that’s way low, we set one that’s really high.” Sometimes they’ll say, “Oh, oh, okay, well, based on what happened with Instagram, or based on something,” but if not, you say, “Well, if 80 is crazy low,” and you start realizing, “Oh, we’re looking at what the cost per new subscriber is, what’s our gut on that, if 80 was too low? What about 100? What about 500?” So the idea of bracketing is you go crazy low, crazy high, and then start moving it in, and you’re comfortable with winding up with a range. That could be a 15-minute discussion, which I don’t want to subject our listeners to…

0:32:30.6 MK: And to be fair, Tim has written articles on bracketing, and it’s actually a technique, I explain to pretty much every analyst that I coach…

0:32:38.2 MH: Yeah, it’s a great way to approach it.

0:32:39.1 MK: As like, “This is… ” Especially when it comes to top of the funnel or brand or things that have never been done before, it is a way to get a stakeholder to set a target, and I am… Hands down, it is one of the best things I’ve ever learned, Tim.

[chuckle]

0:32:56.9 TW: And if you’re listening, chances are they will start to articulate what’s really going on in their mind, so you can… It’s not like you always get all the way to… You’re like, “This is the plan I’m gonna go down,” that the more you’re listening and kind of adjusting as you go, and being comfortable that it’s a range, and letting them be… Pushing them to a point, and if they say that range is too big, well, then they’re also admitting that they need to help narrow the range, and it is trying to get them to have… Be psychologically invested in the metric. If the analyst tells them what it is, then if they miss it, then they just say, “Well, the analyst must have set the target wrong.”

0:33:38.0 MK: Totally. Totally.

[music]

0:33:42.6 MH: Alright. Let’s step aside for a brief word about our sponsor, ObservePoint. Moe, if you had to sum up what ObservePoint does as concisely as possible, how would you do that?

0:33:54.8 MK: Hmm. Well, I would say the platform brings insights, automation, and compliance to the chaos of your customer experience data. How’s that for concise?

0:34:03.1 MH: So I’d say that’s really good. Now, Tim as the host who is most known for not being concise, [chuckle] how would you describe what ObservePoint does?

0:34:14.1 TW: I would say they bring technology governance by validating the deployment and accuracy of analytics and marketing technologies and data, that they ensure privacy compliance by ensuring adherence to digital standards and government regulations for customer data, and they aid with campaign performance by standardizing an automated campaign tracking, and marketing measurement across the customer journey.

0:34:35.5 MH: Yeah, that’s definitely a little less concise. [chuckle]

0:34:37.0 TW: And I would call out that they now offer essentially a free trial. You can go to ObservePoint’s website, enter your domain and contact info, and they’ll do a free audit that they then deliver to you as a video so you can see the basics of the platform in action on your own site.

0:34:51.6 MH: Okay. Are you done?

0:34:54.6 TW: Am I ever?

[chuckle]

0:34:55.3 MH: Alright, good point. So, maybe I should just declare that you’re done on this topic for now. But you, dear listener, are welcome to learn more about ObservePoint’s many capabilities, including requesting that free audit Tim mentioned by going to observepoint.com/analyticspowerhour. Now, let’s get back to the show.

[music]

0:35:15.0 MH: Okay. So, some organizations seem to love KPIs, Tim. In fact, they’re willing to set just tons of them.

[chuckle]

0:35:28.0 MH: Is there a KPI for how many KPIs one should be considering?

0:35:33.1 TW: I feel like there will be at least one listener who will pull out, which, again, I don’t know that I can take credit for, but I did use it in presentations that the “K” in KPI is not for “1000,” and I’ve been known to point that out to people. I’ve also got visuals where, of… That have also been re-used of people kinda barfing KPIs. The fact is, the rule of thumb is, for anything like two to five is what you should target, but I will sometimes come at it saying, “If you haven’t set a target for it, if you’re not able to set a target for it, then it’s not a KPI,” that is the self-limiting factor. Because going through a bracketing exercise or going through any of these can become pretty onerous, nobody wants to sit through all of them, and you say, “Oh, maybe I don’t need to look at the lead conversion number and the number of leads and the lead to quality lead conversion, and the number of quality leads,” you’ve kind of just defined your flow, what you really care about is the number of quality leads for the campaign. It doesn’t mean you don’t wanna sketch out what you think kinda under the hood ratios, but your KPI is really the number of quality leads, how you’re gonna get to there is a bunch of those other… Those are the levers you can pull.

0:37:01.4 MK: Yeah, and I know that I’ve talked about this before, but I do work in a business where we have two company level, I’m gonna say KPIs, but I’m like…

0:37:11.8 TW: Yeah, yeah, KPIs…

[overlapping conversation]

0:37:12.6 MK: I feel like it’s such… I know, but I feel like it’s such a loaded term now, and I’m like, “Wait, they do have targets, okay, check, check.” But the thing is, we know now from so many experiments, when one goes up, the other goes down. They almost… Not always, but very often, ’cause we have annualized revenue and we have monthly active users, and if you start pushing people towards a paid plan, you will have a drop-off of users, and if you push people towards users, they’re less likely to upgrade. Anyway. And so, when I’m working with teams, pretty much there’ll always be like some variant of those metrics, or like a metric which is a leading indicator for those metrics or the things I care about, and I’ll constantly be like, “You can’t have them both as a primary metric, because we know that if one goes up, the other… You have to choose which is the primary metric.” Is that the right way to handle it?

0:38:15.0 TW: That’s one where… Really what you’d like is to have both of them to go up, so what you really wanna do is figure out ways to make one of them go up that doesn’t… That has a tempered effect on the other one… Having KPIs that are… They’re not in contradiction, they’re just kind of more like, I wanna monitor the whole, “Oh, I wanna lose weight, but I also wanna stay healthy,” that keeps me from saying, “I’m just gonna stop eating.”

0:38:44.2 MK: I would say that it’s the same as “I wanna lose weight, but I wanna build muscle.” Which, to do both at the same time is actually quite tricky because building muscle you put on weight, right? So…

0:38:57.8 TW: Well… Yeah, so that’s a good one. It depends on where you are… Where you’re starting from. If you feel like there’s an opportunity, if they’re…

0:39:04.6 MK: Totally.

0:39:06.9 TW: And so that, if you’ve got it to where it’s like, “We have got… Everything is clicking along so well that we cannot squeeze more efficiency into the one growth without… ” If it is so tightly linked, but I would think more often… It’s… The Fed use the US… It’s hopefully not a political right, inflation and unemployment are kind of two that they… And which one do they peg it to? And I guess, in that case they’ve said, ultimately they… I don’t know. I am not enough of an economist to know. I know those are two things that tend to… You can’t have them both to the perfect spot, economically speaking, putting aside politics.

0:39:55.0 TW: I think having KPIs where there’s some degree of balance, like if you just had “I wanna drive new subscribers,” and you don’t have any sort of cost control on it, then say, “Great, let’s just spend a billion dollars and we’ll get new subscribers, our incremental cost is gonna be insane,” so having ones that are… That are actually… It’s tough to move one without negatively impacting the other one. But that gets back to setting the target, saying, “Well, I can’t just go run and shout from the rooftops that this one’s going great,” if I’ve tanked the other one, that’s gonna immediately draw my attention to, “What am I gonna do about the fact that I’m tanking this one?” And I’m gonna rely on some business judgment to say, “Well, the answer isn’t just swing in the other direction so that they flip back, that… ” I don’t know.

0:40:47.4 MK: We’ve talked before about, at Canva we have this concept of nanny metrics, which is like, we need to make sure basically, if there’s a particular KPI we’re targeting, nothing is going off a cliff, for example, like customer complaints, or like maybe it’s page load time hasn’t suddenly slowed down to stalling or… There’s metrics we keep an eye on to make sure they’re not tanking.

0:41:09.4 TW: You said… Is there like a threshold? That one, it’s like the sort of one that you say, this is… And we call them guardrail metrics when we were doing our optimization to say, it’s not good, but we’re not trying to drive it, we just wanna make sure we don’t do something that…

0:41:23.6 MK: Yes. Yeah.

0:41:25.1 TW: And it runs the car off the road and another…

0:41:27.3 MK: Totally.

0:41:27.4 TW: That’s kind of interesting, ’cause those aren’t really… They still have… They more have thresholds to be monitored, but they’re not how you’re… Yeah, those aren’t… Yeah, I like that, nanny metrics. I like that.

0:41:36.9 MK: Yeah. So, this is actually really funny ’cause I was chatting to Ton Wesseling the other day over email about a conversation we had years ago at SUPERWEEK, and I remember him saying like, “If you have two metrics that are conflicting, you’re constantly getting your stakeholders to have to consider and weigh up what is the right balance of one dropping for the other one, what is that level?” And we’ve had discussions internally of like, “Do we need to come up with a new metric which makes that tradeoff of balance for a stakeholder? Or do we just call a spade a spade and we stick to what we’ve got?” And… Sorry, I feel like we’re really getting into the nitty-gritty of…

0:42:18.2 TW: Well, except… ‘Cause it’s interesting, when running a… In a testing context, when you say, “What’s our optimization metric? Let’s pick one.” And then you say, “Okay, but I’ve got these other ones that I can’t totally tank them.” So I wonder if that’s another way to look at it, that if you got two that are competing, saying, “This one, I’m gonna set a floor or a ceiling… I’m gonna set something that I don’t want it to go beyond that, I wanna maximize this other thing,” this other one is like… It’s still a KPI. I have not performed against my expectations. One of these I may want to… I wanna hit it or exceed it and exceed it as much as I possibly can, whereas the other one maybe, I just… I don’t wanna drop below it, it doesn’t mean that I wanna stay wildly above it, but there’s one that I’m really trying to drive, the other one is just really important, so I’ve set my floor for it or my ceiling for it.

0:43:12.5 MK: Yeah, got it.

0:43:13.4 MH: Okay. It is time for the quizzical query that confounds the complexities of our statistics in our podcast with Moe and Tim competing head-to-head on behalf of listeners. Moe and Tim, are you ready for the Conductrics quiz?

0:43:34.9 TW: Ready as I’ll ever be.

0:43:36.2 MK: Ditto.

0:43:36.3 MH: Awesome. Let’s talk a little bit about Conductrics. They sponsor the Conductrics quiz and the Analytics Power Hour, and they build industry leading experimentation software for A/B testing, adaptive optimization and predictive targeting. Go to conductrics.com and find out more information on how you can leverage their amazing tools and products. Alright, here we go. Tim, we’ve got some exciting person for you to represent today. Are you ready?

0:44:10.2 TW: I’m ready.

0:44:10.7 MH: Justin Beasley, that’s who you’re representing.

0:44:12.6 TW: Ooh.

0:44:13.4 MH: And Moe, you are representing… No pressure. He’s been called the “Godfather of Analytics.”

0:44:19.1 MK: Oh geez.

0:44:20.5 MH: Jim Sterne.

0:44:21.6 MK: No!

[laughter]

0:44:23.0 MK: Jim. Aww. I love Jim, but the pressure.

0:44:28.0 MH: Alright, no pressure, none at all. Just know that’s who you’re representing. Okay, here we go. The Power Hour gang was enjoying some vacation time together in Amsterdam. I only wish. While at a cafe, they were chatting about how marketing optimization problems can be thought of as dynamic programming, or more accurately, reinforcement learning problems where one has to learn a sequence of paths such that we optimize the overall journey, a joint optimization and attribution problem. Moe noted that underpinning the ideas behind dynamic programming and reinforcement learning are the concepts of the Bellman equation and the fixed-point theorem. A well-dressed Dutchman sitting next to them listening [chuckle] watched Michael stir milk into his coffee. That’s inaccurate, I would never stir milk into my coffee. Thank you. The man leaned over and said, “Did you know a Dutch mathematician got the insight for a famous fixed-point theorem from watching how milks swirls in a cup of coffee?” Michael smiled and said, “I believe it’s… ” The man was shocked, “I can’t believe it.” What did Michael say to this man? I believe it’s A, Brouwer’s fixed-point theorem. B, Hilbert’s fixed-point theorem. C, Banach’s fixed-point theorem. D, time to mind your freaking business.

[laughter]

0:45:56.6 MH: Or E, Jensen’s fixed-point theorem.

[chuckle]

0:46:03.2 TW: Moe, you wanna take the first elimination on? I feel like there’s a… [chuckle]

0:46:04.9 MK: Yeah, I’d love to take the first elimination. I’m going to eliminate number four.

0:46:10.6 MH: Oh. D, time to mind your freaking business. It was sort of a giveaway because I probably would have not said that to a random person on the street.

[chuckle]

0:46:21.1 TW: Just too nice.

0:46:22.6 MH: The milk was the dead giveaway, right? It’s not really Michael, but an alien has taken over my body. Okay. What else, Tim?

0:46:31.7 TW: I am going… Wow. Which of these does not… I feel like I’m gonna eliminate three, Banach’s fixed-point theorem.

0:46:41.7 MH: Banach’s… Okay. Banach’s fixed-point theorem. Okay. Okay. I think we can do that, we can do that.

[chuckle]

0:46:51.4 MH: Alright. Moe, you’ve got A, Brouwer’s fixed-point theorem. B, Hilbert’s fixed-point theorem. Or E, Jensen’s fixed-point theorem.

0:47:01.2 MK: I’m gonna try and eliminate E.

0:47:04.9 TW: Jensen’s fixed-point theorem?

0:47:07.1 MH: You’re gonna eliminate E, Jensen’s fixed…

0:47:09.2 MK: Jensen’s.

0:47:10.0 TW: E? Okay.

0:47:10.8 MH: Fixed-point theorem. Alright. This is for all the marbles. Jim Sterne. Are you ready? Yeah, you can eliminate E, that’s fine. It’s right. So now, it’s literally between A and B.

0:47:20.8 TW: Yep, I am gonna go… Which I referred to. I tried to go numerical last time, so now I gotta switch back to letters. I’m gonna go with… I think that it is Brouwer’s fixed-point theorem. I have no idea…

0:47:33.6 MH: A, it is Brouwer’s. Okay. Well, let’s consult the computer. Fun fact, this episode or this little anecdote actually did happen to Mr. Matt Gershoff and his partner Heather Greene, while they were in Amsterdam at one point in time, so this is ripped from real life. And…

0:47:55.6 TW: Wow.

0:47:56.6 MH: Brouwer’s fixed-point theorem is exactly what Matt Gershoff said to that man. So Tim, you’re the winner. Justin Beasley, you also win. Jim Sterne, you’re a winner in my book, but not on the quiz, so sorry.

0:48:10.5 TW: And can we also express our condolences to Heather that they are vacationing in Amsterdam, and that was the subject of conversation?

0:48:17.5 MH: Yes, that’s right. [laughter]

0:48:18.7 TW: Okay.

0:48:19.7 MH: Okay. Well…

0:48:21.4 TW: That’s what led her to start drinking whiskey.

0:48:24.3 MH: That’s… [chuckle] That’s right. And look at her today. So…

0:48:27.6 TW: Also brought to you by Milam & Greene.

0:48:29.6 MH: Milam & Greene. Alright. Excellent work. The Conductrics quiz, brought to you by Conductrics, sponsor of the Analytics Power Hour. Check them out at conductrics.com. Let’s get back to the show.

0:48:43.5 MK: Now, not to rush you forward, but there is one thing that you and I have talked a lot about, and actually it’s really interesting, our head of growth marketing now, I love him ’cause he thinks about this stuff in a very similar way, where he’s like, “We need success and failure actions, and we need everyone to agree to them before we do anything.” Talk us through what are success and failure actions?

0:49:08.3 TW: I don’t know.

0:49:08.8 MK: Oh fuck.

0:49:08.9 TW: Why don’t you tell me what success and failure actions are? And then I’ll realize what we’re talking about.

0:49:13.7 MK: Oh. So, I swear, I have taken this from training you’ve done before, where it’s like, “This is the KPI, this is the target, and if we hit this target, this is what we’re gonna do, and if we fail to hit this target, this is what we’re gonna do,” and committing to those actions before you start your campaign, you launch your product… Okay. You’ve got a face. What’s going on?

0:49:36.2 TW: I don’t know. Michael?

0:49:37.6 MH: No, I love this question too because it goes into, basically, some companies are totally willing… It’s way too easy for them to do it because they know they’re gonna ignore what that comes out of it, and then other companies are scared to death to put anything on the line because they know it’ll be their life, ’cause of the way the politics works in their organization.

0:49:58.5 TW: Yeah, that’s a great point. One, I do feel like as a rule, human beings over-index towards the calamity of failure and missing that, and I will have conversations where people are like, “Oh well, I worked at a company, or I knew a guy who worked at a company where wow, if you missed your target you were like… ”

0:50:17.4 MK: Gone.

0:50:18.3 TW: “Doomed.” And I’m like, “Really?” I’ve been working for a long time and actually having… Yeah. But that’s a separate rant. Maybe I think of it a little bit differently, and this is just because of the way that I’ve dealt with… And this is based on my experience with the organizations I’ve worked with, is that there’s so much struggle to even agree on what good or bad is, that trying to get them to recognize the difference between measuring performance and just knowing something is not hitting our expectations, and then figuring out what to do about it. I don’t necessarily think that it’s realistic in all cases to say, “If we miss this KPI, this is what we’re gonna do,” ’cause there may be a hundred different reasons that you’re missing the KPI, some of which may be, a pandemic hits. So, just getting focused on what our expectations are, that’s performance measurement, objective, where are we relative to the expectations that we had? That’s everything we’ve been talking about.

0:51:19.1 TW: Very separate from hypothesis validation, which is, “Oh, we’re under-performing on this,” and then let’s pause and say, “Why might that be? Do we know why it is, that when we set our target, we didn’t realize that a competitor was gonna come out with a much more superior offering?” Okay. Now, ’cause we don’t… We couldn’t have planned for what we would do if we didn’t know that was gonna happen. So the hypothesis validation is around saying, “What do we think… What do you think might have happened? And if that’s what is happening, what might we do about it?” And it’s a whole other episode around trying to think first before just saying, “We missed the KPI. Analyst, go figure out, figure out why.” ‘Cause half the time, if you just ask the right business person they’ll tell you why they missed it. And half of the time of that you say, “What are we gonna do about it?” Well, nothing, we’re gonna know that we… The next time we plan, we know why we missed it, we’re smarter, we’ve learned something, it’s not like we have to explain everything, I’m feeling like I’m in… I’ve got 40 minutes worth of things to say, and I’m forcing myself to summarize in generalities…

0:52:39.8 MH: We definitely don’t have a time for it, so we do have to start to wrap up. ‘Cause usually, if Tim wasn’t the guest at this point, he’d be sending me a little message and be like, “Wrap-up time.” And then Moe would be like, “But I have just one more… ”

0:52:52.6 MK: One more question. One more question.

0:52:52.9 TW: She has one more…

[laughter]

0:52:57.0 MH: Alright. Well…

0:52:57.6 TW: I didn’t really answer that one all that well, but…

0:53:00.1 MK: Tim, lovely, you killed it in the vernacular of our new true crime format. So [laughter] One eight seven! Okay. Let’s do this, let’s pivot to some last calls that we would like to share. And Tim, you’re sort of our guest, would you like to go first? [laughter] You’re cracking me up every time.

0:53:22.0 TW: Sure. I’m gonna go with one that it’s been out for a while, but it was a… It ties into the crime piece, that’s gonna be my link. It’s a cool visualization. It is posted by The Markup and it’s called “Crime prediction software promised to be free of biases. New data shows it perpetuates them.” I don’t feel like this is news, we have talked about biases and the AI biases and the analyst biases before, so I would say in that sense it’s not crazy surprising. But it was something that The Markup and Gizmodo did, it’s actually got a pretty cool kind of visualization, where they were specifically really looking at race and predictions, crime predictions, and looking at neighborhoods based on their racial makeup. And it’s honestly not that surprising, but it’s actually pretty clearly explained as to the challenges of racial bias when you’re looking at these… This system was PredPol, I think was what it was called, predictive policing, but had the great promise of the objective AI is not gonna have bias, and this was a pretty good deep study that showed, yeah, not so much. And that’s my true crime, last call.

0:54:49.0 MH: Okay. Moe, what about you? What’s your last call?

0:54:53.1 MK: So, I have been delving into this topic, basically, about data teams as product teams instead of like data as a service, data as a product, and how I can use this thread, and I ended up in a rabbit hole of articles about it the other day, and one of the ones that I somehow ended jumping to was called the Future of the Modern Data Stack in 2022, which is on towardsdatascience.com, and I feel awful that I’m probably not gonna say this lady’s name right, but it’s Prukalpa, I think, wrote the article, and she actually covers six different points. The first one is the data mesh, which I now actually have a bit of an understanding about, which I didn’t previously, my boss was all over it too.

0:55:41.9 MK: The metrics layer. Three, reverse ETL. Four was active metadata and third-gen data catalogs, which, that one I kind of disregarded. But five was data teams as a product. And six was data’s observability. And it just… I thought she did a really good job of explaining these different, I guess, data trends that are going on, but she also links to some really phenomenal articles on each of those individual six points, and some people in the industry that are really thinking deeply about those topics. So if any of those six topics spark your interest, the link itself is also full of other good further reading, which is really nice, and…

0:56:20.0 TW: You know she is referencing great people and there are multiple past…

0:56:23.8 MK: I know.

0:56:24.5 TW: Analytics Power Hour guests who she…

0:56:26.6 MH: That is true. References.

0:56:27.4 TW: References.

0:56:27.5 MK: References. Yeah.

0:56:28.9 MH: Very nice.

0:56:29.0 MK: And I’m sure you’re gonna hear more about data teams as product teams from me, ’cause it’s like the topic that’s consuming me at the moment, so…

0:56:36.2 MH: It makes a ton of sense. And after you shared it, Moe, I read that article too, and I really agree, it’s a good one. Very nice.

0:56:43.6 MK: Alright, Helbs.

0:56:45.3 MH: Oh, well, thank you so much. Yeah. So my last call, so something that happened back in January was that the Austrian Data Protection Authority said that Google Analytics was a verboten on account of the fact they were moving data from the EU to the US, and that’s created a big stir, but one of the better write-ups about it was from a good friend of the podcast and former guest, Cory Underwood. So his blog, he did a nice write-up on what that means and what are some of the things that might happen as a result of that and things like that. I thought it was actually really well-done, and I don’t totally understand all the ins and outs of it, but that’s because I am not a full-time data privacy professional, but I know plenty. So, that’s how I stay up-to-date on that.

0:57:35.4 TW: Gotta say working… Being a co-worker of Cory Underwood is pretty handy when these sorts of things come out.

0:57:41.6 MH: It is, it is. Luckily, he also blogs about it, and also we’re friends, so I can just ask him as well…

0:57:48.2 TW: cunderwood.dev.

0:57:49.9 MH: That’s right. Anyway. Awesome. Alright. Well, Tim, hey, well, thanks so much for coming on the show…

[laughter]

0:57:58.0 TW: Thanks for having me. I hope you ask me back at some point in the future.

0:58:01.0 MH: Well, you know what, just… It’s an elite group that’s come on multiple times, but you keep up the good work and we may find another topic to discuss with you, let’s say in about two weeks or so. [laughter] So anyway. Yeah. That was fun, and it’s good.

0:58:17.2 TW: Oddly enough, I will not be on the episode that is in a couple of weeks…

0:58:20.0 MH: Oh, that’s right.

[laughter]

0:58:22.4 MH: Or me either. Alright, we’ll pick this back up in a month. Yeah, that’s awesome, ’cause we’ll have our International Women’s Day episode coming up then. Alright. So, yeah, Tim, great conversation. We’d love to hear from you, the listener, so reach out to us, let us know what you think. Come talk to us on the Measure Slack, or our Twitter. You can reach us both in those places, you can DM us if you prefer privacy, or you can just @ us, @AnalyticsHour. Alright. And no show would be complete without thanking Josh, our wonderful executive producer who does so many wonderful things, thank you, Josh. And as always, I know that I speak, in this case for just my one other co-host this time, Moe. [laughter] When I say, no matter whether you’ve bracketed the KPI or thought about your targets, or if you’re having a hard time nailing down that first one, just remember, keep analyzing.

[music]

0:59:26.1 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.

0:59:44.1 Charles Barkley: So smart guys want to fit in, so they made up a term called “analytics.” Analytics don’t work.

0:59:50.8 Thom Hammerschmidt: Analytics. Oh my God, what the fuck does that even mean?

[music]

1:00:00.7 TW: Moe’s giving the side-eye.

[chuckle]

1:00:02.3 MH: You can do it, Moe.

1:00:04.1 MK: Energy, energy, energy, spirit fingers.

[music]

1:00:09.2 TW: I’m pretty sure my parents just showed up and…

1:00:11.3 MK: Do you wanna go say hi?

1:00:12.0 TW: I’ve got the dog… No, ’cause my mother will have less respect for the podcast than the dog does.

1:00:20.5 MK: You mean than my mom who burst in during it in song and dance?

[music]

1:00:25.6 MH: Well, I gotta figure out the last call, but other than that…

1:00:28.2 MK: Ooh, fuck. I’m gonna do that article…

1:00:29.2 TW: Fuck. [chuckle]

1:00:32.0 MH: Wait, which article was that?

1:00:33.4 MK: The one that I thought I’d read before and then Tim pointed out that it had only been published six days ago, so I kind of possibly read it before.

1:00:39.9 MH: Oh. I find that I don’t do any reading or learning at all. And that’s my problem.

1:00:48.5 MK: I’m with you, I don’t… Although at the moment I am. But normally, I’m always like, “What the fuck am I gonna share? I’m not doing anything.”

1:00:55.5 MH: Yeah. I’ve not bettered myself in a long time. [chuckle]
[music]

1:01:03.0 TW: Rock, flag, and bracketing, it’s not just for naval warfare.

9 Responses

  1. VK says:

    Thank you, the conversation was engaging.
    I wonder if you could have a ‘role-play’ session setting KPI’s with a client, live. Would love to witness how Tim wields his charm when it comes to KPIs.
    Thanks.

    • Tim Wilson says:

      LOL! I think that was the idea…and my internal clock kicked in and thought it was going to take too much context-setting (what’s the *actual* situation)–it never feels like it’s unduly lengthy in practice, but it felt like it would be tedious on the show as we started to try it.

  2. Nathan Hennon says:

    Could you point me to where I can review Matt Cohen (sp?) “5 Whys”? I want to properly cite this in an upcoming training session

  3. Lori Brok says:

    Can you link to a blog post/presentation/YouTube video where Tim describes “bracketing”? Google isn’t coming through for me. Thanks!

  4. Cara Ziegel says:

    The web-in-europe article linked in this page (episode 187, KPIs) actually links to Valerie (Lambert) Kroll’s LinkedIN page . . . .. definitely the wrong link!!

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

#260: Once Upon a Data Story with Duncan Clark

#260: Once Upon a Data Story with Duncan Clark

https://media.blubrry.com/the_digital_analytics_power/traffic.libsyn.com/analyticshour/APH_-_Episode_260_-_Once_Upon_a_Data_Story_with_Duncan_Clark.mp3Podcast: Download | EmbedSubscribe: RSSTweetShareShareEmail0 Shares