Pop psychology is fun, if not that useful. Pop analytics can be dangerous! What IS pop analytics? It’s a term coined (as far as we can tell) by analytics legend Kevin Hillstrom, and we managed to get him on the show to chat about it! The fact that it turned into a therapy session for Tim was just an added bonus. NOTE: We hit a glitch with Kevin’s audio 45 minutes into the episode and have done our best to work around it. It was especially painful, in that he had some very nice things to say about the show, but, alas, the choppy audio means we won’t be able to repurpose the clip for marketing purposes! We apologize for the glitch. It was something we didn’t recognize for what it was when it happened, but now we know!
00:00 Michael Helbling: Hi, everyone. Before we start the show, we have a couple of quick announcement. The first is that we’ll be doing a live recording of the podcast at SUPERWEEK in Hungary at the end of January. We would love to meet you there, so head on over to superweek.hu, and give that a look. Next, we’ve got a couple of big announcements as we kick off our third year of the podcast, we’ve got some exciting updates. We now have a website. You can find us at analyticshour.io. Tim informs me that the “.io” means that we’re cool. And on that site, we now have an official way for you to submit a topic you would like to have us cover on the show. Second, starting with Episode 51, the one with Mo Kiss. We’re now posting transcriptions of each episode on our site. We don’t think you wanna read the show, but maybe it will help if there’s something you heard that you wanna go back and find. Anyways, we’re excited about what 2017 will bring. Check out superweek.hu, analyticshour.io. And now on with the show.
01:19 Announcer: Welcome to the Digital Analytics Power Hour, Tim, Michael, and the occasional guest discussing digital analytics issues of the day. Find them on Facebook at facebook.com/analyticshour. And their website, analyticshour.io. And now, the Digital Analytics Power Hour.
01:43 MH: Hi, everyone. Welcome to the Digital Analytics Power Hour. This is Episode 54.
01:49 MH: Sometimes the best way to look within is to get a perspective from the outside. So get ready for what might be a sharp nose of the proverbial analytics smelling salts or at least we hope so. As thought leaders, Tim Wilson and I [chuckle], Michael Helbling, are frequently asked our opinions on best practices, what’s happening in the industry, and other digital analytics miscellany. And Tim’s advice is even highly valued, but who is watching the watchers? In this case, the watchers being thought leaders in our fine digital analytics community. Sometimes it’s a bit too easy to tweet that self-serving industry link not knowing that it’s eroding the trust and not building the skills of the real heroes of our industry, the fine men and women in jobs that aren’t allowed to speak publicly about what they’re doing, but are doing some of the best work in analytics.
02:47 MH: Well, to diagnose our industry’s problems with his aptly named Pop Analytics, we have a guest. His Twitter is legendary for making social media “experts” confused, and provoking deep thought in the rest of us. That’s right. Our guest is Kevin Hillstrom. Kevin is President of MineThatData, a consultancy that helps CEOs understand the complex relationship between customers, advertising, products, brands, and channels. Prior to founding MineThatData, Kevin held various roles at leading multichannel brands including Vice President of database marketing at Nordstrom’s, Director of Circulation at Eddie Bauer, and Manager of Analytic Services at Lands’ End. You can also find his blog post on his blog at blog.minethatdata.com, and he’s a member of our podcast community. You can find his podcast at soundcloud.com/minethatdata. Kevin, it is a delight and an honor to have you on the show. Welcome.
03:50 Kevin Hillstrom: Thank you. It’s great to be here.
03:52 MH: Well, let’s start with the little question. Where did this term ‘pop analytics’ first come up for you, and what was the genesis of that?
04:00 KH: Honestly, the genesis had a lot to do with what I would call… In the last month or so, we’ve read a lot about fake news, and you’ll read how people were being misled by certain information, either politically or however you wanna define it. There was just a lot of talk about how you could basically publish your own ideas. The ideas didn’t have to be truthful. And then people would embrace those ideas, and run with them. And as I was comparing the concept of politics and what we were seeing there with analytics, I just started to see some parallels. And more importantly, I saw parallels where… I had a person on Twitter from the UK who told me that I was focusing on dogma. And I had to stop and think about, was I reciting or saying things that were basically true or was I saying things that just fit the narrative that I wanted to talk about? And I started to think maybe I’m a little bit biased and that there’s certain ways that I wanna think of the world. And maybe when I evaluate how other people view the world, they also have biases.
05:10 KH: So I came up with the term ‘pop analytics’ because it’s a parallel between classical music, let’s say, and pop music. So in pop music, there’s a structure in place that if you wanna get a song on Spotify, and you want it to be popular, you want your song on the radio, there’s a process you go through to make that happen. And similarly, in analytics, you can do really great work. You could be a classical musician, but your work isn’t necessarily going to get noticed by an ecosystem that basically rewards people who do what the ecosystem wants the ecosystem to do. And so that’s where I started to come up with the idea of pop analytics, that there’s a popular form of analytics that tends to get most of the attention, and may or may not necessarily be the best thing for individual analysts at individual companies to employ. And that’s how I got there.
06:02 Tim Wilson: What’s your top three examples of pop analytics? Like deep learning, is that a, that fall… People who are touting deep learning for optimizing their advertising spend, is that a pop analytics example?
06:16 KH: It sure can be. So deep learning, in of itself, is fine. I don’t have an issue with that, and I would have clients that are doing spectacular work there, so that’s not my concern. I would say that anytime that… I’ll give you an example. When I worked at Nordstrom, I received an email from a vendor, and the vendor was recommending a customer relationship management process that involved email marketing, and wanted to basically personalize how email campaigns were constructed so that response would increase in the email campaigns. And so, I got the email from the vendor at about 10:00 in the morning and at noon, one of my managers came into my office, and the manager sat down and said, “I wanna have a serious discussion about email marketing, I’ve got a great idea. And the great idea is personalization. And here is how I wanna address it.”
07:07 KH: And within five minutes, I could tell that this person had read the email, and the person didn’t have his own idea. This person was doing something because a vendor thought it was a good idea, and the vendor convinced my employee that I should hear this message. Anything that has to do with, be it A/B testing, personalization, optimization, anything that requires you to engage your audience, anything that requires you to go deep into an AI rabbit hole, if you will, those kinds of things, where somebody is requiring you to do something because it’s ultimately going to benefit them more than you, that’s what I call pop analytics.
07:49 TW: I think vendors are definitely largely to blame for that, but as you were talking about that example, is there a piece of that that’s also the crawl-run? We’re struggling to get enough content together for our monthly email newsletter that goes to our entire database, and somebody comes in and says, “We should be doing personalized email.” And you’re like, “Well, that’s great, but that means you’re gonna have to drastically increase your content production because you’re gonna now… ” So it’s trying to skip a step. We wanna do multi-channel attribution and we don’t even have last click attribution nailed down. We can’t get tagging. And maybe that’s more of… Is that a symptom of pop analytics? Somebody’s chasing a shiny object and they’re just… Frankly, I think, sometimes it’s hard and non-sexy to do this stuff, the walk-step and the crawl-walk-run, and run is what everybody is out in the trade journals and talking about at conferences, because it’s sexy and interesting, and sells. But recognizing that, “Hey, we’re not ready for that.” That’s great to have that as a vision, but we have to recognize what our path is to get there. Is that fair?
08:58 KH: I think that’s completely fair. And, to me, it’s a reflection of a business knowledge on behalf of the analyst. When you have a really good analyst, a really good analyst understands what the business needs and what the business is capable of doing. And so, the analyst is only going to implement ideas and solutions that are basically, where the analyst knows that he or she can push the business in the right direction, and the idea is going to be embraced. The analyst has become an ambassador for what’s best for the business the analyst works for. And what I tend to see instead, are outside forces who are pushing analysts towards solutions that, to your point, are maybe four steps further than where the analyst knows that the organization can go. I have a client that I met with about three weeks ago, and their mobile analytics are not good. And in fact, some of their basic, what you would call CRM-type stuff from 20 years ago, is horrible. And so, when I’m trying to outline a project for them, I’m not gonna push them down the mobile path because they need to go down 21 steps before they even get there. Yet, when I’m in the meeting, the Vice President of Analytics is talking about things that are 15 steps down the path. And the Vice President was talking about specific vendors and specific solutions from specific vendors.
10:20 TW: He would happen to have gone or been taken to lunch with, at a conference two weeks earlier or taken… He got a nice dinner from that vendor, just so happens? Not necessarily because they bought them dinner, but it’s a shiny object. The vendors have a nice story and that’s… It’s like the Scott Brinker marketing technology landscape. There’re just so many out there and you can pick anyone of them and they have a good story, and if you’re not appropriately attuned to your own situation and the realities of the reality, and the fact that some of these companies have been around for six months. Do you have a take on the ChiefMartec marketing technology landscape?
10:56 S?: Hah.
11:00 TW: You don’t have to.
11:01 KH: I’ve been lucky enough to work for other vendors in my career, and I’ve been on the other side of the table where I have to hit a sales number during the course of a period of time, and it’s gonna require me to behave differently than I would like to behave. And so, I understand how the marketing technology folks are going to push a certain agenda. And to be really fair to analysts, you brought the dinner concept, that someone took somebody out to dinner, a lot of analysts are isolated at companies. They don’t have a lot of people who love them or understand them. So when you do have somebody from the outside who’s willing to take you to dinner and treat you well, and see the world the way you see it, that does have an influence. And I continuously see that across my client base. I am constantly made aware of a vendor taking somebody out to dinner, and then at the dinner, promoting certain ideas and those ideas then get spread across the organization. There is a love that vendors can give some of my analysts that I work with, that they don’t get with their co-workers. And that love that they get from a vendor then influences how the analyst is gonna interact with the rest of the business.
12:05 MH: You’re not wrong, Kevin. And actually, vendors do that not just with dinners, but hey, put your card in the fishbowl, maybe you’ll win, and oh wow, you won, congratulations and let’s get a conversation about our product and you, and how can you help us? And all these things happen are just part of that process. It’s difficult, I don’t think anybody out there is trying to be evil. But like you said, there’s a motivation on the other side that is, “Hey, I’ve got numbers to hit.” I do the same thing, I take analysts out to dinner and stuff, but I’m actually trying to hire them, so that’s my secret. No one is showing you any love and I’ll, “Come work with Search.” No, I’m just kidding.
12:50 TW: Searchdiscovery.com/careers, jobs, I don’t know. I still don’t know what the URL is.
12:57 MH: Yeah. We will have a new one by the time this show launches probably. [chuckle] So yeah, and on the other side of that, it’s easy to get excited about a cool technology. Thinking about my time as an analyst, you see something it’s like, “Wow, that solution could do so many things for us and you aren’t necessarily ready to think about every piece of that, ownership, maintenance, all the… What data needs to go into it? How does it… And there’s so much more to any kind of platform, or whiz-bang thing that, of course, is never discussed in the sale cycle ’cause those are just details. And that’s where we all get hung up. And actually, from a prior podcast where we had Matt Gershoff on talking about personalization has an upper bound. There’s this Faustian bargain of, well, how much personalization can you afford to do? ‘Cause it’s like, well, every vendors telling you, “Well, you should have a one-to-one conversation with every customer.” It’s like, “Well, I’ve got 25 million customers, I don’t know how to scale that.” So, yeah. No, it’s good.
14:11 KH: If you look at a career path for an analyst, if I go back to my days when I was at a Lands’ End or at Eddie Bauer and I want my career to progress on a certain path, it’s gonna be very hard for me to hit my career goals within the company I’m working with. It’s gonna be a lot easier to achieve my goals if I get the skills that the vendor is recommending I obtain, and then those skills can translate to another organization. There’s a clear incentive for me to learn what different technology providers are offering or what solution providers are offering, because that makes me more attractive to other companies. And so a lot of the analysts I work with then had this incentive that doesn’t necessarily align with the company that’s paying them. And so that’s just a challenge that… We’re talking about pop analytics, that’s all part of that process. The analyst is being pulled into a world that isn’t necessarily aligned with their organization and it’s neither good nor bad, it just is.
15:07 TW: So if I recognized that technology X, which just looks like to be like they’re phenomenal at marketing themselves, they do guerilla marketing like crazy, they show up and crash other technologies, events, whatever. I think they’re gonna have a large footprint and it may be a little poppy, and lots of companies are gonna buy it. And if I can get my company to buy it, and I can learn it well enough, then actually, there’s a good chance that a lot of companies will be struggling to use it. And now they potentially are gonna say, “We need to get somebody who knows this technology.” That’s terrible. That’s like… Maybe that’s how IBM, half of their platforms, have actually managed to work their ways throughout into the entire universe.
15:53 TW: But I think the other thing about chasing, specifically around the technology aspect of pop analytics is to me, it’s this… And again, it’s not an intentional or evil thing that somebody inside a company is doing, but if I hear the story of this technology is going to solve, make my life better. Then as soon as I convince somebody to get that technology, immediately I have bought myself time. I have bought myself three months, six months or a year because we have to actually get through procurement, acquire the technology, go through the discovery, gather the requirements, get it implemented. And so we automatically said, “Oh, you know what? We have the solution. The cavalry is on the way, but it’s not gonna be here for six months,” which immediately gives us some breathing room. And I don’t think it’s that conscious of a process. I think we’ve probably all been down the path of, by the time, six months or nine months or a year hits, expectations are insanely high, because now that new technology is gonna solve… It’s gonna solve why the M&Ms in the snack room are always running out on Tuesdays, and don’t get restocked ’til Thursdays. It’s gonna solve every possible problem. I’ve never worked anywhere that’s stocked free M&Ms, I wish I would. If there’s anyone out there that has that deal…
17:12 MH: Hey, Tim, let me tell you about this exciting company Search Discovery. Just kidding…
17:20 TW: So what besides technology-oriented pop analytics… I think that’s like the… That’s the easy one to point out and I could rail on that all night long. What are some non-buy new technology-oriented examples? As long as it’s not, thinking you should use open source software like R to try to pursue data science. Other than that, any other example?
17:44 MH: Also, that’s the bridge that’s too far, is that one. [laughter]
17:47 TW: Hey. No, we can go after that one too. I question that myself.
17:53 KH: Before I answer that, tell me why you’ve questioned it, or tell me your thought process around evangelizing that as an idea.
18:00 TW: So my thought process has been… I will take a simple example that since I can probably go back 10 years, it has bothered me that watching a report come out that everybody wants and there’s value in it because it shows the volume and traffic of the site, or it’s the volume of revenue. And watching that we look at the week, the report, and we say, “This week is gone up or down from last week.” And lots of people wanna go and chase up or down as being good or bad, like literally up or down, it doesn’t matter how much. And I have done some probably inexcusable things to try to add to that recurring reporting some indication of whether it went up or down enough for us to think that it might not just be random noise.
18:48 TW: And I’ve done some of that stuff in Excel but I realize I don’t really have the right answer for statistically how to say this moved enough that I should care. That’s like one… I’ll take one example and say multiply that by a thousand. I feel myself looking in Ad Hoc Analysis in Adobe or Google Analytics interface and it has just aggregated line charts and I just deeply, deeply believed that there is more to that and it’s more to that in two ways: One, anytime I have to go and build 27 segments, export the data one thing at a time, I think, “Wow, if I could script this and do it faster.” So there’s no statistics in that. That’s just get the data more efficiently and not say, “No, I can’t get that for you because it’s unreasonable for me to get it.”
19:36 TW: And then the second part I think not just doing a scatter plot and putting an R-squared on it and saying, “Yeah, looks like those are related.” Now if I could do that a thousand more ways, I might find something interesting, and actually having the skills to be a little more knowledgeable about, what are the variables… I’m learning terms like factorization. I’ve got 85 eVars populated. I can’t look at every single combination. Which one of these… Can I do some factor analysis? I don’t know how to do that yet. So that’s the short explanation broadly of why I am pursuing… It’s not necessarily R. It could be R, Python or any one of those tools. Have I defended myself or have I lost all credibility?
20:20 KH: You have defended yourself. You gave exactly the answer I’m looking for. ‘Cause the answer you gave is not a pop analytics answer. Your answer is you want to do something that is helpful to somebody else and you’re trying to find the best way possible to come up with a good outcome, and that outcome then is pushing towards a package or a solution so that you can be more effective. So that’s exactly what I wanna see from an analyst, and so that makes me happy. Okay?
20:47 KH: Let me give you another example. I’m on the phone with a client a couple of weeks ago and I’m explaining how I would address dealing with an issue. And the CEO is sitting with his analyst on the call and I can hear whispering going on. And the CEO says, “Kevin, I would hire you but you’ve got understand, we have an analyst here who knows R and knows Python, and so therefore the $20,000 we would pay you would be useless, we don’t need your help.” And so I said, “Okay, that’s fine. I can deal with that.”
21:19 KH: A week goes by, the CEO sends me an email. The CEO says, “We’re having problems getting the answer we want with R and Python. Can we talk about how you would help us?” That is to me what pop analytics is about. Because pop analytics is saying there’s software and the software will provide the solution and it de-emphasizes knowledge and curiosity among analysts and business leaders. And so I rail against anybody who says that some sort of software is the answer. You will see… I get so frustrated because I will constantly see how organizations say, “You need to push… You need to come up with a data driven solution and push your analytics down into an organization and make it easy for everybody to do analytics.” And you’ll read that Tableau could be the solution for you or something else.
22:08 KH: And all of those things to me fit into the pop analytics family of things that I don’t like and don’t appreciate. ‘Cause to me, it devalues what you just talked about Tim, in the essay that you just gave us about how you were having a problem and you’re looking to find a solution to that problem. And so I am on your side. I want analysts to be doing what you’re doing, what you just described. Where you have curiosity for something, you care about it and you’re looking for a solution and you are not married to what the right solution is. You’re just married to coming up with a way to do a better job for the people who are hiring you. So I’m very glad you gave that answer ’cause that’s what I’m trying to encourage.
22:45 TW: Whew! Big sigh of relief.
22:47 MH: And you are okay in my book too, Tim. Good job.
22:50 TW: That I don’t give a shit about.
22:52 MH: I would even say, Kevin, that even if you look at it from the other side and we’re all consultants, right? We all have… We do work with clients. Customers come to us saying, “Hey, help me do this tool.” And so, if I wanna grow my business I need to be able to say, “Hey, I’m an expert with this tool and you’ll find a lot of consulting out there to help do tools. And it’s perverse, because you can do a lot of bad with great tools because you’re just doing the tool and at the same time, customers don’t know how to ask you for what it is that they actually need sometimes. And so you have to have like our whole consultancy is built on this model of, “How do we help you start to derive value out of this set of tools, even though probably you engaged us in the first place a lot of times to help you do a very specific scope of putting the tool in place?” Help me put this tool on my website. Okay, that’s fine. But now how are we gonna pull this into what you’re doing each day.
24:00 MH: How is this metric mean anything to your business and how will you use that to determine if something is more or less likely to happen as a result of that? Ideally with an idea going forward from there into a deeper understanding of actual people who come to your websites. So, it’s tricky, ’cause it’s like you’re always walking this tight rope and being, I’m the now the guy on the outside, I’ve been, have the chance to be on both sides, I think actually all of us. All of us who have been both practitioners and consultants over our careers but it’s difficult ’cause you… I think it’s helpful though to be able to say, “Hey, you don’t need this product from us. You don’t need this tool. What you actually need to do is go hire a really good analyst right now. You need to go hire an analyst and then come talk to me in six months when they’ve had a chance to ramp up and… ” You know I don’t tell people to do that very often but I’m proud of those moments where you do say to people, “The best thing you could do is not spend money on me. The best thing you could do is go spend money on… “
25:01 TW: With Demystified.
25:02 MH: Well. Yeah. No, of course. Slug.
25:07 TW: Well, how often… I feel like, Michael, you, and I are definitely, because we’ve got this web analytics like Google and Adobe and absolutely, we need somebody to help… Come in and help us do stuff with these tools. Kevin, do you get my perception and of course, the grass is always greener on the other side, do the people come to you and say, “I have my database. I feel like I’m throwing money away with a lot of my catalogues I’m sending.” Or are they asking you the right kinda questions or do you get clients who are asking you a question where they presupposed an answer that may be pop analytics type answer, and do you try to steer them to say, “Are you gonna be receptive to a non-pop analytics solution and if you’re not then I’m not the person for you?” How does that work with clients are coming to you based on kinda like your reputation in the space?
26:00 KH: I have branded myself on purpose to not answer the questions that a pop analytics person would answer. So, what I’m typically answering are a CEO or a chief marketing officer, or a chief merchandising officer or a chief financial officer are asking me, why is my business failing? That’s literally the question I get, is why is my business failing, and that allows me a playground that is enormous. So, my scope is not narrow then. I’ve purposely branded myself so that I get to have a large scope. And now because I have a large scope, the answer might be, it’s a digital issue. It might be that it’s an offline issue. There might be tools that can help my client perform better. I get to have a large playground with that. What I end up running into is there is usually an analyst or a team of analysts at the company that’s hiring me. And those people are either not being listened to or those people are just fundamentally wrong.
27:00 KH: And that’s where… Where, for me things get interesting because it’s almost never about the tools that they have in-house and it’s never about how to use the tools. It’s either management doesn’t wanna hear a word that the analysts are saying or the analysts are just completely going down the wrong path and management knows it. And so I end up spending an awful a lot of time on my project than is being just like a people person, trying to get people to talk and trying to tell a story with the information that is going to cause people to talk. And so I get to craft really interesting outcomes then. I don’t have to have a specific outcome that is tied to a tool and tied to a methodology. I get to have a pretty broad range of things that I can do and the way I look at the business then gets to be a lot more interesting. But I had to do that on purpose and build my business that way over 10 years, so that I could have that kind of fun. So that I didn’t get married into having to answer small questions that would not benefit the business.
27:55 TW: Got it.
27:55 MH: And I think, just talking consulting shop for a second, you do have to really specifically guide your consultancy where you want it to go ’cause you know STI’s story was, “Hey, we have a tech management tool that we sold to Adobe and so you guys are the Adobe DTM guys.” Their first-year post that sale, we were the DTM guys. And it’s like, “Okay. I didn’t start a career in analytics so I could do tags for a living.” So, that’s where we’re gonna start with people. But then what are we gonna… And you have to really have an idea of, “Okay, well, that’s not how I wanna… ” And to a certain extent, Tim, I think Demystified is the same way. There’s a philosophy behind the business that is guiding how things are set up and how you’re doing things.
28:41 TW: I appreciate your generous…
28:42 MH: Well, I just wanna pull everybody into it. Everybody is trying to be you. Don’t… Just take it. Just take the compliment.
28:51 KH: So tell me, Tim, with your business then, do you feel like that you have a vision for where you wanna take things and where you ultimately wanna go and that you are bridging that between the work you’re currently doing and where you want things to be in a couple of years? Do you have a process in place that helps you get where you wanna go?
29:10 TW: Oh my God. You are one perceptive son of a bitch. There’s part of me that is, “Tim can build you a nice dashboard”, like we’ll get your web analytics data good and your web analytics data does need to be good. And you know what, you don’t wanna spend a bunch of time every week pulling a recurring report or if somebody wants to look at something they wanna explore but they don’t wanna learn the interface. So I’ve managed to inadvertently get myself, you can march down the path of, you can automate stuff in Excel that’s using web analytics data or data that’s in the web analytics platform and there’s value there. Right? I have a process we go through, we talk through what’s the right stuff for you to look at, for you to monitor your business and identify where there might be issues.
29:57 TW: But that tends to still wind up as the… You can build a pretty dashboard in Excel. So to me there is a deeper… I look at management consultants which are easy… Take the Deloittes of the world, take the EYs, that are coming in and are tend to be brought in from somebody at a higher level asking a bigger question. And I think in a way, like Gary Angel, this was part of it, when… He has talked about this as well, you’re not starting with, “My conversion rate’s not good, the data’s not good for this thing,” and that’s where you’re starting. It can be difficult to work your way up to say “We wanna tweak and tune these little small things.” So I absolutely… I feel that for the web analytics or the digital analytics world, there’s a risk that, that entire profession could get their clocks cleaned by individuals and organizations that are coming in and starting out asking bigger questions at a higher level, that matter more, that aren’t wedded to this narrow world of digital and the website, and the Facebook page. I think that’s where it’s going, so maybe there’s a little bit of fear. I’m just old enough that I could probably ride it out on doing clever stuff in Excel and doing ad-hoc analysis here and there, make sure we have a good hypothesis and make sure there is action being driven.
31:20 TW: And clients would be perfectly happy, because they would look and say “This is how much money we gave him, and these are the questions he answered, and these were the incremental value or insights that he provided.” I don’t think, 10 years, 15, 20 years from now that that’s gonna carry much weight. I think all this other data that’s happening… All this data’s that’s out there: One, it doesn’t change the fundamentals of, “Do we really know our business? Have we thought about our business? Do we know how the parts move and they interconnect?” And it challenges me when clients don’t wanna think about that; but then the second part is saying “We can’t really think about just one channel. And maybe I will start to sound thought leader-ish, “Ah, it’s all about omnichannel.” Yeah, omnichannel, I’m trying to see if you have that, that makes you fall out of your chair. But so that’s… I probably don’t have as clear of a vision as I should, but I definitely have a vision that what I’m doing right now is pretty solid and narrow, and I’m trying to figure out my way around that.
32:21 KH: See, I…
32:22 TW: Just bill me later. What’s your bill rate for this therapy session?
32:26 KH: Well my homework assignment for you, and it’s not something you’re going to answer today, but I think this goes for most analysts, is that you go down a path where you align with a technology, and you get to be really good at marrying business issues to that technology and coming up with solutions. And then you just brought up something that was interesting, you talked about Deloitte or some big management consultant who will come into a company, and for $750,000, they’re going to provide a 70,000 foot solution that they’re going to be running from in four months, and a bunch of foot soldiers are gonna have to implement and have no passion for.
33:04 KH: And so both of those things to me get a lot of attention from a pop analytics standpoint. And what I hear you saying is, there’s a hole, if you will, and there is a hole where an analyst can go… So to me a really good analyst goes and finds a hole that is not getting as much attention as it should, and that hole, oftentimes is, there is a business issue that business people are missing, that analytics or software or technology is missing, but because you have enough experience now, you know that that hole has an opportunity, and you have experience that tells you that there’s going to be a 70% chance you’re gonna have success if you go and fill that hole. And to me, that is a transition that a generation of what I would call digital analysts are going to go through over the next 10 years. So in other words, this is just my opinion, digital analysts have spent the last 10 years wiring, basically how the internet works, and how commerce works, and how customer engagement works, and how customers interact with each other, they spent 10 years wiring that. And now pop analytics basically wants analysts to go deeper into that.
34:19 KH: And to me, analysts have the opportunity to do the opposite, and analysts have an opportunity then to step into the business a little bit more. And there is just a ton… And Michael, maybe you could talk a little bit about what you experienced at Lands’ End, because you would have a shared experience with me there. There is a hole that is as big as you can possibly imagine between ordinary garden variety analytics and what the business actually needs to be successful. And so I will let you, Michael, talk about that, but I think that’s a gigantic hole that a digital analyst could move into in the next several years.
34:52 MH: No, you’re absolutely right. It was because of that experience that I’ve shifted how I described myself as an analyst, because really what I determined what was needed of me to help was to just be a business person who is using data in meaningful ways. Don’t be a digital analyst, don’t put yourself in such a small box. You need to be able to go up to where the business is working, and it was… That was where I learned, “Hey all these digital metrics that I’ve been working with, they have no English equivalent in how people run businesses.” And the best insights we generated were typically… That were adopted, so in other words we generated some insight from our data, but we got them adopted by the business and we drove value to the business, they were all presented in terms of how the business talked about itself, not in terms of my set of metrics.
35:49 MH: And so that was the big illuminator that said, “Okay, so what I really need to do is I need to transcend my dataset, or at least understand the translation between the two, so that I can help the business understand what it is that I’m seeing in the web data. ‘Cause nobody at the C level is gonna come down and see my metrics and understand that, “Wow, bounce rate is really important.” Unless they went to a conference and suddenly you heard somebody say, “Bounce rate’s really important.” And then they’re like, “You know what? Bounce rate needs to go on our big time review dashboard.” And it’s like, “But why?” “What does it do?” And then you will say, “Well, it’s bounce rate, you just have to get it lower.”
36:29 MH: You know that half of our traffic doesn’t even come in on the home page. They come via direct… And so it’s this whole other thing. But again, how does it connect? What’s the connection? Bounce rate’s not a thing to a business. That’s not EBITDA, that’s not revenue, that’s not orders, that’s bounce rate. It means nothing unless it’s relatable to the things that actually drives the business. So, that was probably the biggest thing as an analyst I learned from being at Lands’ End, was that, everything we did over there in digital analytics had to come back into the business in business terms. And then we could actually start to have a conversation and move some stuff around. But then the other piece was you had to then step into the role of the guy moving stuff around ’cause you had to do that part too.
37:15 MH: So certainly, you stop being the analyst and you start being the project manager of the new change you wanna see.
37:22 KH: And so to me that’s an important point that you bring up because in a pop analytics world, you’d be told that you have to become a storyteller and you have to tell compelling stories to get people to act. And what you just described was, I just told the story and it’s gonna require me to act. I’m gonna become the project manager. And that’s, an analyst goes through a process where eventually they figure out that it’s going to be hard to get other people to act the way you want them to. And you as an analyst start to basically move much closer to the business as a result. And that is the transition that I see happening among the analysts that I’m working with is they’re beginning to see that that’s happening ’cause not enough people are listening to them. So they’re actually coming closer to the business to make a difference.
38:06 MH: I like that.
38:06 TW: I struggle with that having, and I’ve gotten, I’ve been around long enough that I consider myself 50% a marketer and 50% an analyst, and I’d love to go and talk into the business and talking about and probing a little bit, like, “What are you really trying to do? What’s are the outcomes you’re trying to drive?” I’ve run into a lot of analysts who don’t have that ability. I like the way that you’re putting it. It’s not just tell a better story. At the end of the day, it comes down to, if you’re just about the data then you probably aren’t really equipped to do enough with the data. I’m gonna sound poppy, maybe, potentially. That’s when I work with clients, and I’m often in the room with a lot of non-analysts that I’m dealing with, and they seem to like me and I reflect on why do they seem to like me? I’m like, “Oh, ’cause we’re actually having discussion about their business,” so that’s the, “Hey, Tim’s awesome.”
39:01 TW: I am also at the same time saying, as an outside consultant, “I don’t have access to all your data, I have access to the sub-sets of your data, what can I do to help you address that challenge?” In a lot of cases they don’t have the analyst internally who should have access to any and all of the data, who can also have that discussion. Who can also say, “You know what? Fine, sign me up, I will commit to helping you actually make this thing, make this change, make this a reality.” I feel I ran into plenty of analysts in my clients who are just not… They really struggle to think about the business. Some of them even, they’ve fallen over their tools so much that it’s all they want to do, is “Let me just create another dashboard.”
39:41 MH: And it’s a big space. You could spend, there are people who’ve spent their whole careers just invested in a specific tool and how it works. And that is the summation of their expertise and it’s very deep and it’s profound. That means that you’re part of a good team but you can’t be the team. And that’s the distinction. And I think, Kevin you’re not wrong that over the next 10 years I think that’s what we’re all gonna have to come to grips to because there will come a point in time, and this already happens and I bet all of us hear it from different places where executives are like, “I spent quite a bit of money and I’ve got nothing in return.” I get no value from these things that I was told to “buy”. Or that everybody said, “Well, if you do this that is a common refrain today and tomorrow the money won’t be there to invest in it, we all owe it to ourselves as analysts to start thinking about, how do we bridge that gap a little more?” But also I wanna throw out there that not every single person has to be what we are, but you should definitely recognize…
40:53 TW: That you mean thought leaders, just to be clear?
40:55 MH: Yeah, exactly. That’s exactly… No, in other words, to varying degrees, we moved in the business direction and leveraged softer skills, and things like that. I think there are people who are amazing. And Tim, I feel like you embody this with all the work you’ve done with R this year. People with really amazing quantitative skill sets, in Math and statistics and all those things, they don’t have to become… I can’t think of a good example, but some cool person that none of us are, and smooth and suave with the business people. It’s more to understand what components make up what has to happen with analytics and knowing the personalities. And so as analytics leaders, one of the things that I think about a lot is: What’s the right team? How do you put the right team so that one person can handle those hard technical questions, and somebody else, they’re still minding the store when it comes to thinking strategically and interacting with the people?
42:00 KH: And I think there’s… You talked a little bit there about really what is the role of an analyst as married to a career path? I’ll give you an example. I had an analyst at Nordstrom whose job it was was to forecast sales by the hour every day, that’s what he was supposed to do. And so by him doing that, he got to meet with the Nordstrom family, he got to meet with most of the senior executives on the retail and on the online side of the business. So he has this amazing job, but he was so content with that job that he didn’t want to do anything else. In other words, he didn’t really wanna do anything statistical. He already had what he wanted. And so there’s… I would never push that person to go into new projects or to try something different, because that person was perfectly content.
42:51 KH: And I would have other analysts that were restless, and they wanted to be… You could tell that they eventually wanted to not be an analyst, they wanted to be working in the business. And so the projects that I would align them with were analytical projects that were purposely with a business person who would embrace what they were saying, so that both sides would get something from it. And so, I think there is this career path component to being an analyst that probably doesn’t get talked about enough, but you get to pick where you want to go and I wouldn’t wait for somebody to pick that path for you, I would be very proactive about that. In a pop analytics world, I think there is a… It’s prescribed where it is that you should head. And I think an analyst needs to take more ownership of what they want to do long term and have a vision for where they wanna head.
43:40 TW: At the risk of opening a huge can of worms, what’s your take on… I feel like there’s a challenge… Or I’ve made this point to several clients that say your only analyst is somebody who is content to check these boxes, and punch these buttons and do these little simple trend lines in Excel. Those analysts are training every one of those stakeholders into what analytics is and how it works. So I think it’s… If you got a team and there’s somebody who just wants to… This is their comfort level, this is what they wanna do, I think that’s okay. I think I struggle with anyone, where you’re representing the industry. And if you’re letting yourself define it, if you’re defining it by your little comfort zone, which means you’re not trying to push yourself, and learn and do more, then that would be okay if that was only limiting your career and maybe you could just ride off into the sunset and do that for the rest of your career.
44:36 TW: What I have trouble grappling with is they’re the stakeholders who are getting that. You’ve trained every one of them now that that’s what analytics is and a lot of them are gonna leave and go on to another company, and it is like a virus that is gonna spread. So I… Conceptually, I like the the idea of saying, pick your narrow path, get in your comfort zone and you just wanna be the stable analytics person. I feel like those people can really do harm to the industry, because they’re not pushing themselves and they’re potentially training people in a bad way, they’re overly limiting analytics, and then other analysts who are more motivated not to overcome that barriers at the next company that those people, stakeholders encountered. That’s been on my mind of late, as well. [chuckle]
45:24 KH: So let’s pretend you were managing a team and you’ve got three analysts and two of the analysts have that kind of mindset. Tell me how you would work with the rest of the people at your company to help them understand what the potential of analytics is versus the career path that two of your three analysts wanna take. So tell me how you would consider managing that process.
45:47 TW: So I got into consulting partly, so I wouldn’t have to deal with that sort of situation, but I’ve always struggled with the… I’ve had a team that was almost 20 people in total and then it was just like the two or three people who were in that mode of, “I wanna go heads down and I don’t wanna grow and learn.” We could manage around it. If it was a team where one person was gung-ho and driving and pushing and saying, “How can we move the organization forward through what we’re doing?” And two people saying, “Everything that you introduced scares me, because it means I’m gonna have to think harder or do something different or potentially my job will look different in six months or a year from today”, I’d be having the conversation with them to say, “Is that the right role for you to be in?” I had a lady who worked for me once at a very large insurance company in the customer insights and analytics group.
46:44 TW: And about two months into her working for me, when we had a free, free to us, college level class, Ohio State was gonna be teaching a statistics class, or more of a design of experiments class, and we were all required to take it and she was resisting taking the class. And I was like, “Why are you resisting this?” She’s like, “I’m not really a data person.” I was like, “You’ve been at the company for 15 years and you applied for a job inside the customer insights and analytics organization. What were you thinking?” She’s like, “Well, that’s like the organization that has a lot of buzz and a lot of cache right now. So it seemed like it would be good for my career.” And I’m just like, “I can’t help you.” I literally do not know how to communicate with somebody who is in this space and not passionate about it. So yeah, I guess the answer is, I would leave or I would probably be managing those two people out, and trying to find people who are more motivated, is a very long winded answer to that question. Do you have a better solution? [chuckle]
47:48 KH: I don’t know, I do not have a better solution. I think it’s a challenge to align what analysts are good at, what their career paths are, and where you want them… What the business needs. So that’s just a part. I don’t think there’s a right or wrong answer to that. I do know that over the next several years, analysts are gonna have to come closer to the business and probably a little farther away from the tools, and the vendors who provide those tools. I think that’s going to be very important.
48:19 TW: I agree.
48:20 MH: I agree also. And upon that agreement, unfortunately we are gonna have to start to wrap up. But, before we do, one of the things we do on the show is called “The Last Call” where we go around the horn, talk about something we’re finding interesting, maybe a cool tool or a vendor that we think… No, I’m just kidding. [chuckle] But anyways, Kevin, do you have a last call you wanna share?
48:46 TW: Hi, podcast listeners. This is Tim jumping in. I’m not Kevin. But as you noted, we started to have some audio quality issues there. So what I’m gonna do is paraphrase Kevin’s last call. Basically, he started off by saying that he’s speaking at a conference, and there’s a link in the show notes, the Vermont/New Hampshire Marketing Group Conference, where he’s actually built a little business simulation tool in Excel, where all the attendees get to actually kind of run a mock business over a period of time with three products. He’s got it built in with the cost of goods sold, and they get to make the decisions about which channels they’re investing in, and then this tool simulates the outcomes.
49:26 TW: And he isn’t necessarily pushing that conference, although it does look pretty interesting, if you’re in the New England area, it’s a pretty low cost conference, this is me editorializing. But he went on to say the reason that he was bringing that up is because he sees simulation as being one big thing that will be coming down the pipe for analysts. It’s what’s done with weather forecasting. If you’re a Formula 1 racing team, they’re building simulations, because simulations can really help you take a model, or take a collection of models, and actually really start to understand cause and effect when you start moving around different variables, and really understanding your business. So, it was really a last call pitching simulation. As a matter of fact, we chimed in and thought, “Wow, that would be a great topic for us to have in the future.” So, that’s his last call. I apologize for the audio quality, that you got to listen to me again, rather than hearing Kevin himself, but that’s what he covered, and it was a great last call, even if the audio was very choppy. And now, back to the actual episode with the actual audio.
50:34 MH: Which is totally the wrong time. Way to bring up a brand new idea, Kevin, right at the end of the show.
50:40 TW: Well, it’s funny because I’m gonna jump in with my last call, because it is by no means the anti… I am fascinated by simulation. I feel like that may be something I get… My sense is that with simulation, it forces you to do some work to understand, to break down your business into some component, bits and pieces. Like you have to understand your business in order to build a model that you’re gonna use to do simulation, and then run multiple simulations, as I loosely understand it. And literally three days ago, I sat with a statistics professor, who was trying to explain how simulations worked. But my last call is an article, and this might be the first time we’ve had a semi repeat on the last call front, because Michael, you have done a last call about the entire Analytics Made Skeezy blog, Mr. John Foreman.
51:33 KH: Oh, yeah, I’m a big fan.
51:35 TW: I’m gonna last call a post from him from four years ago, called “Projecting Meth Demand Using Exponential Smoothing.” What that is, is it basically explains how Holt-Winters’ forecasting works. And the reason I came to it was, I was looking at some Adobe data, and I was in R, and I was looking at Adobe’s Anomaly Detection and reading up on that. And I’d seen Holt-Winters before, and I think legitimately, wanted to understand what’s behind his forecasting. And the good old Measure Slack, Jules, I’m gonna butcher his last name, Stuifbergen, he’s @zjuul on Twitter and in the Measure Slack. Said, “Hey”. After I got some lengthy definitions of how Holt-Winters works, he pointed me to this “Projecting Meth Demand Using Exponential Smoothing” which is as John Foreman does, a witty scenario that then works through step by step, “Here’s like a shitty way to forecast. Here’s how you can make it a little bit better, here’s how you can make it a little bit better. Now stick with me, and oh look, we’re doing exponential smoothing, and see how that works. We’re decomposing the signal, we’re putting this into place,” so I recommend that specifically, especially if you’re using Adobe. They’re using it out the Yin Yang for how they’re doing their forecasting. And maybe it’s a terrible way to do forecasting, but it seems like it’s a hell of a lot better than fitting a trend line in Excel and stopping there.
53:02 MH: Nice.
53:02 TW: How’s that?
53:03 MH: No, that’s really good. And I’m a huge John Foreman fan. Local Atlanta guy, MailChimp, thank you very much.
53:11 TW: I need to go back and finish reading his book. I’ve got it on my Kindle. What your last call?
53:18 MH: My last call is a long overdue last call. My colleague and coworker, Lea Pica’s Present Beyond Measure podcast. She’s been a guest on the show, but I think she’s got some new episodes recently, especially the one she did recently with Garr Reynolds, which I think is really phenomenal. So, check that out for a very specific take on how visualize and tell stories with data better.
53:47 TW: That’s great. So I think you have another podcaster on, and he took a little bit of a potshot at data storytelling, and you recommended Lea. Way to go. Kevin’s never coming back now.
53:58 MH: I’m not afraid.
53:58 TW: This is not gonna be the part 1 of 17 that we were hoping for.
54:02 MH: Part 1 of 17. Yeah, in our ongoing series.
54:07 MH: Well, If you’ve been listening to this show and you’re thinking to yourself, man, what Kevin just said, or what Tim just said, or even if I, what I just said, resonated or you have questions about it, we would love to hear from you. Also you heard Kevin give a homework assignment. I recommend everybody take that homework assignment and do some thinking on their own about it. It’s a really good thing for us all to consider as analytics people, but as you’re thinking about it, you’ve got comments, questions, please don’t hesitate to hit us up both on our Facebook page, our site analyticshour.io and on the Measure Slack. We’d love to hear from you. You can also reach Kevin on his blog at blog.minethatdata.com, and he’s pretty accessible via the Twitter @minethatdata on Twitter. Kevin, thank you so much for being on the show. We’re so glad we finally picked the lock on what it is that I can actually secure a guest of your stature. We think the show after now these many years, two years in, we finally figured out… No, I’m just kidding.
55:20 MH: Anyways. Thank you so much. It’s been a delight to have you on the show.
55:25 KH: Thank you both and there are ten of thousands of analysts out there who don’t have any connection to anything. They’re working with executives and other coworkers who don’t understand them. And you provide an opportunity for them to hear about some of the challenges that you have and how you deal with those things. Don’t underestimate how important that is. There’re just so many people that are lonely out there as analysts, and you provide a really good way for them to be able to hear different challenges and to do it in a lighthearted way and not in a serious way, which is really good. I appreciate what the two of you do and I hope your listeners do as well.
56:02 MH: Wow. Thank you.
56:03 TW: We made it through 54 episodes before we said, “That’s it, were done, drop the mic, we’re out.” We just peaked.
56:09 MH: That’s right.
56:11 TW: It’s all down hill from here.
56:13 MH: I like to say that we are the number one explicit analytics podcast on iTunes.
56:23 MH: Alright. For the rest of you out there listening, thank you so much and as always, from my co-host Tim Wilson and I, keep analyzing.
56:34 Announcer: Thanks for listening and don’t forget to join the conversation on Facebook, Twitter, or Measure Slack group. We welcome your comments and questions. Visit us on the web at analyticshour.io, facebook.com/analyticshour, or @AnalyticsHour on Twitter.
56:54 Charles Barkley: So smart guys want to fit in, so they’ve made up a term called analytics. Analytics don’t work.
57:03 TW: Using a name with profanity, but in a very complimentary way.
57:07 KH: That sounds good.
57:09 TW: Have you ever thought of doing one, two, three, four, six, seven, eight, just to see if anybody freaks out, if you skip? Just to mess with them?
57:19 KH: No. I haven’t thought of that.
57:23 TW: Questions you’ve always wanted to ask Kevin. If I remember what my point was. Son of a bitch.
57:33 TW: You know what, Michael, why don’t you just take it from there?
57:38 TW: We will see how much of this I am comfortable leaving [laughter] publicly. Rock flag and pop analytics.
This site uses Akismet to reduce spam. Learn how your comment data is processed.