Have you ever seen a one-man show in the theater? It’s awesome. Unless it’s terrible. The same can be said for one-person digital analytics teams. It can be awesome, in that you get to, literally, do EVERY aspect of analytics. It can be terrible because, well, you’ve got to do EVERYTHING, and it’s easy for the fun stuff to get squeezed out of the day. On this episode, we head back Down Under for a chat with Moe Kiss, product (and digital) analyst at THE ICONIC. Whether you pronounce “data” as DAY-tuh or DAH-tuh, Moe’s perspective will almost certainly motivate you find new ways to push yourself and your organization forward.
00:05 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 now, the Digital Analytics Power Hour.
00:25 Michael: Hi, everyone. Welcome to the Digital Analytics Power Hour. This is episode 51. More and more, the function of digital analytics is being tackled by teams, and that’s great. But there’s still a lot of people out there working on a team of one, within their organization. And truth be told, most of you who’ve been around for a while, started out that way too. So on this show, we look up from our hub-and-spoke center of excellence ivory tower, and we talk about the team of one. Since Tim and I are so far removed from these days, we sought out the person furthest away in the world of whom this was true, and we found our guest. Her name is Moe Kiss and today she’s the product analyst at THE ICONIC, where she also heads up the digital analytics function. Prior to that, she’s was a senior analyst at Datalicious. Welcome Moe, it’s great to have you on the show.
01:22 Moe Kiss: Thank you so much for having me, guys.
01:24 Michael: Awesome. Tim, how you doing? You’ve actually met Moe in person…
01:29 Tim: I’m doing well.
01:30 Michael: On all your travels this year, right?
01:33 Tim: This year and the last year as well.
01:34 Michael: Nice.
01:35 Moe: Unfortunately.
01:36 Michael: Thank you, Moe, for pressing past that obvious…
01:39 Tim: Is that your perspective? [laughter] I would say ‘awkward’. She’s working with the government right now. Let’s see if we can stop that from happening ever again.
01:49 Michael: There you go. That’s right. Tim Wilson, added to Australian no-fly list.
01:55 Michael: I don’t know if that’s a thing. Well, Moe, for our listeners, could you give us a brief introduction and maybe just a little bit about your role at THE ICONIC today?
02:03 Moe: Sure. I’m actually pretty new to digital analytics. I actually came to the industry from government. I worked for the Australian government prior to that, and basically, one of the reasons I ended up in the industry is because both my sister and my future brother-in-law work in analytics. And when I looked to move career-wise, I realized what an amazing community it was, and that’s how I ended up here.
02:32 Moe: Yeah. So, I now work at a company in Australia called THE ICONIC. It’s the biggest online fashion retailer in Australia, and we get 10 million visits per month. Having said that, we just celebrated our five-year anniversary, which means we are still really young and our company is actually, in terms of our work force, is very young. So, 80% of our staff are under 35, which I actually love, because it means we have this really enthusiastic and “Let’s try anything” type of work force. So, it’s actually a really contagious energy in office to be in, which makes being an analytics person in the building really good fun.
03:11 Michael: So, what kinds of tools are you using and how are they used by you and others within the organization? ‘Cause I think that’s something that a lot of people try to figure out is what’s the right tool set, when you’re the only person owning and running that function?
03:26 Moe: Well, I actually think a really good analyst, and while I am the only person that looks after digital analytics really at THE ICONIC, we certainly have lots of different types of analysts. I really think they should be tool-agnostic, and everyone certainly has their favorites. I’m a person that uses whatever will make my job faster and easier. So, I use R, I use Tableau, SQL, Excel, Google Analytics. We also have Snowplow and Adjust. Actually, one thing that I do wanna do is clean up the tools that we have and make sure that we’re really getting lots of value, and the maintenance is really up-to-date on all of them. But I really think as an analyst, you need to be able to learn anything. And that’s one of the reasons I love my job is that I get to learn new tools all the time.
04:12 Tim: I think there are lots of one-person shops that are actually using Excel and Google Analytics or Excel and Adobe Analytics. And one of the things where it seems like you’re not necessarily the typical, although be happy to get feedback from people who say, “No. Absolutely, I’m a one-person digital analytics team as well,” is that you are going to the level two tools. And we’ve had the discussions before about, is the US ahead or behind? Every country seems like it feels they’re behind. Every time I’m traveling to Europe or Australia, I start to feel like maybe the US, from the tools adoption, is behind. But when you came to the… So you guys have a BI team that’s got… How many people are in the BI group?
05:01 Moe: About six.
05:02 Tim: Outside of the web analytics, you guys are a GA, dual GA, Google Analytics and Snowplow basically are your main web analytics stuff, is that right?
05:12 Moe: Yeah. So, we use GA primarily for the marketers and the product team to answer those really easy questions or those questions that they can self-serve. And we implemented Snowplow to really let us customize some of the things that we wanted to track, but also to answer much more complex questions. And that’s where really needing to know SQL and being able to connect to our data warehouse, which our amazing data team built, is super important as well.
05:43 Tim: So, when it comes to things like Tableau and R in the data warehouse, were those all platforms that were in existence at THE ICONIC when you joined, or did you have to… Were some of those ones where you said, “There’s a gap based on what I’m being asked to do, I need to go grab another tool.”? Were your glomming on to other tools already being used or were you adding ones yourself?
06:06 Moe: So I was really fortunate at Datalicious, the last agency I was at. I actually had the opportunity to learn R and also Tableau, and had a bit of a play with, sort of be querying things like that as well. So when I join THE ICONIC, I already had a pretty good foundation, definitely still learning though. But Tableau was definitely on its way to being implemented with our data warehouse. But as for R, that’s actually something I’m trying to teach the rest of the analysts in the building at the moment, and we just had a session last week. Because, like I said, we are a young company and so we’ve evolved so quickly that sometimes our analytic practice hasn’t moved at the pace that it needs to. And we are sort of replicating efforts, and we need to find a way to automate basically, and R really lets us do that. So at the moment we’re sort of on a path to automate as much as possible, and I see R being really instrumental in that.
07:01 Tim: Okay. I think we’re done. So R is awesome. Let’s just stop right there.
07:05 Moe: Yeah, I’ve definitely drunk the Kool-Aid.
07:09 Michael: R is fine and now that there’s… I’ve got more and more people on my team using it as well and now I’m slowly being sucked into its orbit. So yeah. No, I think R is awesome.
07:22 Tim: I feel like this is atypical. Well, one, from your intro, Michael, saying ‘teams’, and I think we’ve had this discussion in the past, the more distant past, where it was, “Hey, we’re working with huge organizations; they have teams of analysts.” And I always had to say, “Not really, I’ve got a client that’s got an enormous amount of online revenue and they’ve got one full-time digital analyst, and they had none for an 11-month stretch.” So I don’t think it’s that uncommon to have a single analyst even in large organizations, but in those cases those people are really scrambling just to keep the lights on and try to produce some kind of recurring dashboards. But then when I talk to people more locally or at conferences when there’s sole operations, I don’t often get that there’s that much of kind of an awareness and drive to say, “We’ve got to automate all this stuff, we’ve got to bring in other tools, we’ve got to collaborate with other groups in the organization so that we’re all on the same page.” I think there’s enormous value in that. And I don’t know, Michael, what your sense is from the clients and non clients and just other analysts, whether what most describing THE ICONIC, to me almost feels kind of atypical from what I feel like I run into.
08:40 Michael: Yeah. Certainly, there’s aspects of it. Just the way you describe, Moe, the environment and the people, a lot of times, at least, in my experience, and again, we gonna go back a few years before my experience becomes relevant again. But the issue was that it was organizations who weren’t progressing with analytics that weren’t investing in the people, so then you’re kinda stuck with one person in a room with old tools, without a lot of ability to kind of move them forward. So it’s, I like the contrast and paradigm because it sounds really exciting what’s going on where you’re at, and so it’s more like there’s momentum building behind this. And in reality it sounds like it may not be long before there’s other digital analysts that become part of this to help drive these things forward. So I don’t know what your perspective is on that or if you have one, Moe, but it’s interesting.
09:38 Moe: Well, I was actually chatting to Krista from Google about this the other day, because she’s just spent sometime in Asia, working with a lot of Australasian clients. And I do think, and obviously my experience is much more limited and I haven’t spent time working in the US, but I feel like Australian businesses missed an analytic step that the US companies went through. Which was really sort of like GA and Adobe, UI focused, sort of going through that Excel and hitting limits on the number of rows. And we just kind of have skipped that and gone straight from, “Oh, okay, this is what analytics is. Yeah, we should care about it,” to trying to do much more advanced things, and that’s, the business requirements are more advanced. But because of that, our skill set has adapted to those more advanced techniques because we’ve had to, because we kind of didn’t know about the intermediary step, if that makes sense?
10:40 Tim: Yeah, that totally make sense, and I guess that seems like it’s the curse of the first mover or something. I guess this is causing me anxiety with the American analysts who were kinda working in Excel and even the American-based companies that are trying to… The platforms that are saying, “How can we support Excel more?” or “How can we build stuff into our web interface to make it more robust?” That’s leaving people in a comfort zone they probably shouldn’t be in.
11:11 Moe: And I think where that might start to become a real problem is, one of the most powerful things that I’ve noticed with things like R and Tableau and SQL, is the ability to join data sets, and for us that’s super important, to join our digital analytics data with our order and sales data. And you can’t do that in Excel. We have too much traffic, we have to many orders. And I think when businesses realize what’s possible by using some of these newer tools, there might be that appetite to actually upskill their staff.
11:47 Tim: Well, I wonder if that’s where we talk about sometimes the kind of traditional consultancies, that have done business consulting, management consulting, are kind of starting to light up on analytics, if whether they’ll be coming in with a little bit more of a view of the web analytics data is just… See, now I’m saying ‘data’, data, man.
12:07 Moe: [chuckle] Data.
12:07 Tim: So I’ll be dropping a bullet here.
12:09 Michael: Data.
12:10 Moe: Data.
12:11 Tim: I don’t know, that’s probably a big, much broader discussion, is sort of the future of the industry. But the whole ‘skipped a step’ actually makes a whole lot of sense. That’s actually, I hadn’t thought of it that way, and that absolutely seems like it could be the explanation.
12:24 Michael: Well, and it also sounds like you’re going to where you need to go. And in a certain sense, there’s almost even sort of a little bit of evolution that’s happened in terms of the understanding of analytics from a digital perspective over the last 10 years, and so a lot of people wouldn’t. When I was a one-person team, it was web trends and a handy-dandy ODBC connector, which you could export into tools like Excel. And getting data joined or integrated was certainly a good idea, but it was awfully hard. And there are a number of tools now that weren’t there, available then, that I wonder if it just helps elevate that and gives better options kind of, in a way. I know the tools are better today, so I think that’s kinda true. So I might’ve just said something really, sort of, obvious.
13:17 Michael: How was it for you, Moe, 10 years ago, when you could only work with the old tools versus now, when you can only… No, I’m just kidding.
13:23 Tim: Well, I do wanna have a basic question. So Tableau, you picked up at Datalicious. So when it comes to another sorta challenge that sometimes happens is the asking for stuff, asking for, “Can I get this tool?” Now R is nice ’cause you can get rowing on R with no expense, and it’s only when you need to start getting stuff up on servers that you need to pay. But was there any pushback for you if you came in and said, “Hey, to do this efficiently, I need Tableau.”? Did you have to build a case? Or any other tool? Or is it kind of the mode of recognized, kinda the enlightened, “Hey, that’s relatively low expense. If this person we hired as an expert says they need it, we can get it.”?
14:07 Moe: Sorry, at THE ICONIC, I’m actually both blessed and cursed that we have both a founder and a managing director who is a former consultant and analyst, and also a CEO with a very analytical background. Which means that as a business, if we put forward something that we need to help us with analysis, it’s definitely supportive. What I’m actually working on at the moment though, is that I think we’re actually getting the balance wrong, and that old saying that I probably heard on your podcast at some point, about the 80-20 rule, “You should spend 80% of your analytics budget on people and 20% on tools,” I think is pretty accurate. And I’m not sure at THE ICONIC we have that balance right yet. So I’m about to ask for something which could tip the scales a bit and be a little bit more difficult to put the case forward. Because at the moment, asking for the tools is really easy.
15:02 Tim: You’re gonna ask for a clone of yourself?
15:04 Moe: No. I will ask for someone much better than myself.
15:07 Tim: [chuckle] Okay.
15:10 Michael: No, I think that’s good. And actually, you just touched on something that I think we could expand upon, and other people who are on an individual team, I think could get a lot from, which is that question of, how do you best build your skill set. And you mentioned that you have leadership that’s kind of analytical by nature and has background there. But certainly, digital analytics as a discipline as well as some of the newer things that are happening in that space, you need the ability to kind of learn, find new things. What are the ways you found to stay current? When I started it was, I just read every blog that I could find. But that’s no longer…
15:50 Tim: Both of them? [chuckle]
15:51 Michael: No. Hey, back when there was a Google Reader, okay, I had 120 odd different analytics blogs that I had cobbled together, some of which hadn’t been updated for years. But I would read any blog I could find. That’s not today’s trend. Blogs aren’t kinda the way people get to information. So what are you seeing in terms of what’s been keeping you up to speed and giving you new insights or skills in terms of growing your skill set?
16:21 Moe: So for me, like I said, I’m still learning R and SQL, and Tableau, I definitely need to improve just because all of that data sits in there now and there’s so much more we can do. At this stage, I have capped it at like, “Okay, why don’t I focus on these three things, because they’re all quite different and complex and have their own nuanceAnnouncer” So at this stage, I’m kind of not trying to focus on the new, shiny object and just be like, “These are the three things I really need to focus on, because if I nail them then there’s not a lot of things that I can’t do, essentially.” But ultimately for me, the biggest way that I stay up-to-date with what’s going on is the analytics community. If you just jump in #measure Slack, people are always telling you what’s new and you can read and go, “Okay, I should check out that blog, or maybe this one I can skip,” that sort of thing. But it is something as a one-stop shop I guess, I’m finding difficult is, how do you keep up with the workload and also find the time to upskill yourself and continue to develop? It’s definitely a balance.
17:26 Tim: I think that is the top three challenges for every analyst, because there’s always more work you could do. [chuckle] Some people have checklists or they commit that the first… I’m not that kind of person. “First 30 minutes of every day, I’m gonna just catch up on blogs.” Do you have anything formalized that you’re saying, “Not a day’s gonna go by that I don’t check something out,” or I don’t know?
17:52 Moe: So I used to block out three hours every Friday when I was really starting out with R and I wanted to hit my head against a wall, which I often did. But now I’m at a stage where I just say, “Okay, I know that there’s an easy way to do this,” and that easy way is normally pivoting some data in Excel. My way to keep learning is to be like, “I know that it would be faster to do that, but I’m gonna push back, take a little bit more time to do it, and I’m gonna force myself to use SQL or use R or use Tableau, even though I might be able to answer that question faster if I used Excel.” Because otherwise, you could’ve put to practice what you’re trying to learn, and I think that’s the best way to do it. Well, that’s what’s working for me.
18:32 Tim: That’s good advice, I think, for any analyst, no matter what they’re…
18:36 Michael: Yeah, absolutely.
18:37 Tim: Gotta figure out a way to keep pushing yourself, even if it’s harder or a little slower, if you know there’s gonna be ultimate benefit. So I’m doing that myself, [chuckle] I wish it was only three hours on a Friday afternoon and I almost managed to make headway with R, but [chuckle] you’re gonna learn R this weekend, right, Michael?
18:57 Michael: I do 15 minutes every Tuesday.
19:02 Tim: It’s R next Tuesday, Python the next Tuesday, gonna Tableau the Tuesday after that.
19:07 Michael: No, that’s why I haven’t learned anything in two years.
19:11 Tim: But, just to clarify, you said three things, and I don’t think you were being… Did you list what those, that’s Tableau, R, and one other thing or you were just speaking?
19:19 Moe: SQL.
19:21 Tim: An SQL? Okay.
19:22 Moe: Yeah.
19:23 Tim: So that’s kind of on the technical tools efficiency and kinda more power side of things. It seems like it could easily be a one-person shop, saying: “I’ve got plenty of technical skills, but while I’m not communicating, I need to really focus on learning the business.” Or “How am I gonna do that? I’m gonna once a week go pick a different product manager, and try to go to lunch once a week with somebody who can teach me about the business, or working on communication or presentation skills.” But it does seem like it’s tough in a… It can be tough if you’re not highly internally motivated. And I struggle with people who aren’t highly internally motivated, so maybe that’s the end of it. The people who look for ‘Give me my path forward’. It doesn’t sound like any of yours has been top down driven? You’re a one-person show in a role that you’re, in many ways, the only role. It’s not like the HR department at THE ICONIC is going to pull out the career pathing document for you and lay out the, “Your goals for this year should be to learn X, Y and Z.” You’ve got to kinda take that on yourself?
20:29 Moe: Yeah, absolutely. My immediate boss kind of, like many people, fell into analytics. But at our company, and this is, I think, actually one of the unfortunate things about skipping that middle digital analytics wave, is that digital analytics doesn’t seem to have the same, what’s the word for it? The same level of importance that it might in the US, in that a lot of Australian companies, and THE ICONIC is one of them, treats it as just another data set that we will use to answer questions. Which is great, but it’s also something that needs to have an owner. And being an owner of digital analytics is a full-time job. It is huge. But at our company, it’s just part of one of many things. So having a boss who doesn’t necessarily play in that space, I really do have to drive, sort of, what my learnings are gonna be, and my KPIs. And the way I think about it is: “Do I wanna have a job in five or 10 yearAnnouncer” Okay, well, you can see what’s trending, where the direction that we’re going in. So, how do I make sure that I’ve got the skills that I need in five or 10 years to have a job?
21:40 Tim: So are you the owner of the… It’s just, I’m thinking of it as a data set. Which is, to me, it’s fair. It’s like, you’ve got your customer data, you’ve got your web analytics data.
21:50 Moe: Yep.
21:50 Tim: Are you the owner of the data set? Are you responsible for the, not necessarily pushing, well, I don’t know. Are you guys using your GA Premium and you’ve got Google Tag Manager, is that right?
22:02 Moe: Yes. Yeah.
22:04 Tim: So where does your responsibility… Are you doing stuff in GTM? Are you speccing out tags and somebody else is doing stuff in GTM? Does that fall…
22:12 Moe: Yes.
22:13 Tim: So are you the data set owner, ’cause I think that’s kinda common?
22:17 Moe: Yes. I do own the data set. However, I’m also really lucky, because most of our developers along the way have had to really upskill. So basically, when we wanna deploy new tags or tracking… We just launched a new Android app. I’ll work really closely with the developers, and I actually sit with engineering quite a bit to sort of enable this, because I’ll map out exactly what things I want tracked and how to do it, and then they just go in and do the actual physical setup for me, which is great. And then I’ll work with QA to test that all those tags are firing correctly.
22:52 Tim: So, when you say, just I wanna be really clear, ’cause I’m curious. When you say you spec out how… So there’s on the one extreme would be: “Hey, I wanna track every screen and every one of these actions,” and vague, and you hand it off.
23:04 Moe: Yeah.
23:06 Tim: Another, more detailed, would be: “These are all the screens. This is the tracking that I want applied,” and you hand it off and they figure it out. The next level would be: “I’ve looked at how the STK works and the documentation for doing this in an app, and I’m providing, not necessarily final code, but directional code stuff.” Where on that spectrum do you… ‘Cause the more detailed you get, the more time and more knowledge you have to have of the implementation. And where are you on that spectrum?
23:37 Moe: So I definitely go through the STK and instructions and whatnot. I wouldn’t say I’m the final for the code. But I definitely do sit and help sort define rules as the devs are writing the code.
23:50 Tim: And then, are you actually doing the work in GTM? They’re putting stuff in a data layer and you’re doing stuff in GTM?
23:56 Moe: It can be either/or. It can be either/or.
23:58 Tim: Okay.
23:58 Michael: So, let’s get this straight, you…
24:01 Tim: Are a bad ass. Oh no, sorry.
24:02 Michael: Yeah. Well, I was working my way up.
24:05 Moe: ThankAnnouncer
24:07 Michael: No, but I think in a certain sense, that’s one of the things that is sometimes challenging around digital analytics. There’s so many different disciplines that are part of it. And a lotta times, people who are on teams of multiple things, end up specializing in one aspect without growing their skill set really into some of these other spaces. I’m really curious for you and as you look at your career and your progress through it, what’s your perspective on that concept, generalization versus specialization in terms of specific skillAnnouncer Do you feel that you’ll go towards a specialty, or do you wanna stay more of a generalist?
24:47 Moe: Oh, so we had some really great discussions recently about the concept of a full-stack analyst at THE ICONIC. And one of the reasons for this is that we have full-stack engineers who can work across all of our platforms. And we’re really toying with the idea of, “Should we have a full-stack analyst?” And I have very strong opinions about this. I definitely believe that analysts should be able to be general enough that they can do most things, but I think, particularly when I look at my role with the product team, they really want to know: How do we improve this feature? Should we pivot this feature and try thiAnnouncer Or even if I’m working with the marketing team on this campaign, and how did it perform? And you can’t do a lot of that stuff if you’re a real generalist.
25:33 Moe: And when I say ‘a real generalist’, I mean in THE ICONIC definition of, you do financial analysis, you do customer service analysis, you do… I’m trying to think of every other data set that we have. But unless you’re really close with the team and you know what’s going on with them, and the features that they’re building or the campaigns they’re launching, the customers they’re targeting, the products they’re trying to sell, it can be really hard to provide good, valuable insight, and that’s why I lean towards: You should be a bit of a generalist so that you can do most stuff or you can hack your way to most stuff, but at the end of the day, to provide really a lot of value to your team, you need to be a bit of a specialist. And that’s where I see myself going. For example, with the tagging space, we’ve realized my role is just doing too many things. So we’re actually trying to hire someone at the moment to help with that, which would be amazing.
26:29 Michael: No, that was not disclosed at the beginning of the show. [chuckle] No, I’m just kidding.
26:33 Tim: Yeah, wait a minute.
26:37 Michael: Glad we got you on now before you are no longer relevant for this.
26:41 Moe: Yeah, yeah, yeah.
26:43 Tim: It’s funny ’cause it’s, we’re asking the generalist question. There’s two ways to listen to it, and from where we were just talking about getting into GT… There’s generalists within digital analytics, where everything you’ve talked about, with kinda this is where you’re focusing and trying to grow, to me says, you absolutely are going towards the analysis communication, working with the business. And then it’s, “Oh, yeah. Oh, and by the way, I also own the fucking data set and have to spec out tags and learn the STK.” Which you’re fine with doing. And that’s, to me, that’s a benefit of being a one-person shop. I started out as a one-person shop too and that meant I had to learn how the internet works and learn what server calls are and what goes into a server log. So that provides this awesome foundation and then it seems like whether you do or not, you gravitate towards, “Hey, I wanna get into actually doing stuff with the data. Buy hey, I’ve built this incredibly solid foundation ’cause I had to know what an STK was, I needed to learn how to read it and think through the data collection.” But then you brought the generalist in of… Super generalist of, “Hey, if you’ve got the word ‘analyst’ and if there’s analysis being done, financial… ” That seems borderline insane.
27:55 Tim: Again, having some knowledge of those areas, knowing where they tie in, being able to interact with them and communicate with them. But I’m thinking about some of the teams of analysts that I work with, where clearly, they never went through that one-person team role. They have never learned the basics. And it is continuing to hinder them throughout their career. ‘Cause they’re trying to just live, just be a Tableau person. It’s like, Jesus Christ, you’re not, that’s not gonna… You’ve got to understand how this stuff works. So it seems like it’s a strong case for being the, “I’m responsible. I’m the owner of the data set. I’m the owner of the tag. I’m also the one who’s gonna be expected to use that data.”
28:36 Tim: But in your case, I think, you’re already figuring out where you really want… When that growth happens, you’re not gonna say, “Finally, somebody else is actually gonna do the analysis. I can get back to where I’m full-time working on the tagging.” But there are other people who are one-person shops who really get into that, and they gravitate towards being an implementation specialist, whether that’s in a different role or it’s when they expand the team. I worked with a client a couple years ago where that had happened, where the person who was the sole owner, when they brought in the other person, that guy really just wanted to be doing Adobe Analytics, the implementation side, and was happy as could be to just be doing the tagging. I don’t know. We just hit on six different things. Or I just rambled all over the place. Either way it was interesting.
29:19 Moe: Sounds like a typical conversation with Tim.
29:21 Tim: Yeah. Now we have that recorded. Great.
29:25 Michael: I knew there was a great reason we had you on as a guest.
29:28 Moe: Yeah, yeah.
29:28 Tim: We said that seven times. Now we have it on tape.
29:30 Michael: That’s right.
29:31 Moe: Excellent.
29:33 Michael: Well, when something’s true, Tim, you just gotta keep saying it.
29:37 Tim: I embrace it. Because I haven’t been able to fight it. [chuckle]
29:39 Michael: Here you go. Well, what else, Tim? [laughter]
29:46 Tim: It’s interesting, so you kinda hinted that you guys are likely going to expand. And that’s likely, it sounds like, on the data set ownership, I guess, or it’s the tagging. I’m putting words in your mouth.
29:56 Moe: Yes. So we’re looking to hire what we’re calling a Digital Analytics Specialist who’s more on the implementation side, which will be really fantastic. Because I definitely, I as a person, I’m much more passionate about sort of doing the analysis and trying to communicate it to the business and, yeah, I’m much more sort of the people-focused. I find the implementation stuff interesting and I think it’s super important to know, but I also think one of the reasons that I’m learning all of these different things, is because I’m trying play catch-up. I’m in an industry that most people have been in for number of years, where they’ve had time to learn these things as they go, and I really do need to dive in the deep end and try and get my hands dirty so I can understand as much as fast as possible.
30:38 Michael: No, that’s good.
30:39 Tim: Yeah. ‘Cause you’re experienced. You’ve got experience working, so you’re not getting steam-rolled, you can actually seem to be able to ask ‘why’. But how does that, you’re an organization, you said you’ve got basically the founder, knows you and interacts with you regularly, you’ve got your formal boss, you’ve got dotted line bosses; everybody knows what you do. You’re trying to balance the work-work with your professional development, which will ultimately support the work-work. So how, when everybody knows what you do, what kinda challenges are there with prioritization of requests when you can have… Obviously, the seniority of the person to some extent, but you’ve got product managers, maybe the BI team, you’ve got your manager making requests. You don’t have necessarily a great gate keeper or a big pool of resources to kinda even staff out, “Hey, it’s a 10-person team, so when I’m really hammered, this other person, they maybe a little [31:38] ____, they might be able to help me out with a couple of things.” How much of a challenge is that, of just keeping your head above water on the requests and figuring out which ones to work on and to what extent? How deep to go on them, I guesAnnouncer
31:49 Moe: So THE ICONIC actually uses a Spotify model for our teams, which are cross-functional teams. So this actually works fantastic again for engineering. The place where it gets a little bit more complicated is me, and I’m sort of the odd ball, and that’s because well, my cross-functional team is the product team, which means that I report to head of analytics for my professional development. I report to the director of product for my priorities and tasking. But that does get complicated, because I also manage the digital analytics space, and that means working a lot with marketing. So I am in conflict quite a bit, and it’s one of the things that as a team we’re talking about right now, how do we structure ourselves better so we can take on each other’s overflow? Is there a better way we could be doing thingAnnouncer
32:40 Moe: But in terms of my day-to-day priorities, the one thing that I find really, really tough is that the long-term big strategic pieces of work are the most important. But often one of the questions that I constantly get asked or I’m constantly trying to manage, is something like, “Why did sessions per user increase this month?” And, “Can you look at it this way? And can you look at it that way? And can you apply this segment? And can you do thiAnnouncer Because then we think we’ll have the answer.” And then the product team obviously wants to hear the one answer, which is that they’ve launched some new feature and they were the ones that drove that metric up. And then the marketing team wanna hear that it was their amazing campaign, and ultimately a metric like that has multiple drivers in a business as big as ours. So I feel like I spend time managing that, that probably is wasted, and could be better used on those big strategic priorities. And if anyone has any advice on that, please tweet me @MoeMKiss, because it’s definitely something that I’m struggling with at the moment.
33:45 Michael: Well, I guess then the other… Even when being asked… How often are you finding yourself, there’s a question that you could do the basic, “Hey, I’ve looked into it. There’s no obvious answer. Here is the response, I can check it off. Or I could spend literally five times as long, which means instead of two hours, it’s 10 hours. Instead of half my morning, it’s more than a day, and dig way deeper,” do you find yourself kind of battling that as well? Analysis, you could spend as long as you want on anyone [34:16] ____ or you could spend…
34:16 Moe: Definitely, definitely. And I think that is definitely difficult at a company like THE ICONIC, which like I said, I’m blessed and cursed in that our company really cares about listening to our analysts and what the numbers are and what we’re recommending. But I think sometimes we do verge on caring about the numbers too much and that actually cripples us. That there’s an appetite for wanting every piece of analysis to be so perfect that we know what the answer’s gonna be. When at the end of the day you’re talking about the future, you’re making a series of assumptions, those could be right or wrong. And it actually would’ve been more efficient to just turn it on, test it, and if it doesn’t work, turn it off.
34:58 Moe: And that’s definitely, yeah, as a young company, something we’re trying to balance is that what level of analysis do you go to, and I constantly I’m like, “Where do I stop?” One thing that I am trying to use a little bit more, which I’ve written a blog on, is called Analysis of Competing Hypothesis, which for me is a little bit of an easier way to show to the team that, “Okay, this might not be 10 weeks of analysis, but I have given it very wide thought and you can see that by the hypotheses that I proved and disproved in my metrics.”
35:39 Michael: Wait, so how does that work? I had clearly missed this post.
35:42 Moe: Yes, so there’s a really sort of detailed complex version, but a really sort of quick version that I use is, if I’m trying to answer a question, I list all of the hypotheses down and then I have a column that’s evidence that proves, and I have a column that’s evidence that disproves, and obviously I focus on those first. And I’ll share that with the team and then we can say, “Okay, we know that these three hypotheses have probably contributed to this metric going up.” I’ll try and quantify how much that particular hypothesis contributed to it. But they can most importantly see that I haven’t sort of missed anything, because if I have, they’ll be like, “Oh, hey, you haven’t got this hypothesis on your table,” or they can see why I’ve actually disproved one, which is also really useful.
36:30 Michael: Got it, okay, yeah. Yeah, I guess it’s a meta-analysis, and listing stuff that maybe you haven’t done. But rather than trying to go and do all of it, it’s trying to put the options in front and do a little bit of sort of prioritization and feedback in group discussion, and hopefully making the team be a little bit more not… I feel like a lotta times, and I think this can definitely steamroll one-person digital analytics teams, is if the person has a question and they know they aren’t supposed to be able to get the answer to the question, they can kinda throw the question at the analyst, and then they’ve kinda moved that off their plate ’cause they’re waiting for the analyst to get back to ’em. And I don’t think that’s super consciously done, but I feel like it’s kind of, “Oh it’s easy to just kinda spout stuff out, and if we spout it as as we’re expecting you to return it to us, and it’s gonna take you two weeks, well then I’m absolved of having to actually spend any more thought on this business issue ’cause I’m waiting on the analyst.”
37:32 Moe: I completely agree. And I think, for me, I definitely work with a team that wanna see every single angle of every single metric, and what I’ve started saying is, “Yep, I can to that. It’s gonna take this amount of time, but FYI it means I’m not gonna be able to do this other thing.” And then they can make the call about whether that’s a good use of resources.
37:54 Michael: Yeah, it’s that commitment to prioritization, I think, that a lotta people really struggle with, and it puts people in a position where they just feel like, “Ah, everything’s dumped on me and I can’t make any movement at all.”
38:07 Tim: Until they react and they’re like, “Now I’m just saying ‘no’ to everything and nobody wants to deal with me.”
38:10 Michael: Right, and then nobody wants to work with you any more, and it’s good, that’s very good.
38:16 Tim: But what’s the level of… Kinda back to this sort of Tableau and R, and you’re in SQL world, how often are you trying to create things that are self-service that, “Okay, if you wanna slice this thing 27 different ways, here’s a thing where you’ll be able to slice it 27 different ways for this request, the next request.” How much do you find that you’re able to, or it makes sense to, empower the marketing or the product managers to self-serve?
38:46 Moe: I think that’s absolutely huge, and that is my life quest at the moment. The product team are fantastic, they’re really across Google Analytics, and often what they’ll do is they’ll come to me and be like, “Hey Moe, I built this report, I built this segment, can you just check for me that the logic is correct?” And that’s great. With Tableau it’s a little bit of a steeper learning curve, so I’ve created a bunch of dashboards for the relevant teams, so they can go in now to the web view and manipulate the data, apply different filters or date ranges. But I definitely do need to spend a bit more time training them as well, so that they understand the full realm of possibilities that they can do with our Tableau data source, because it is so fantastic with the way that the data warehouse has been set up.
39:34 Tim: Yeah, which I think that one is being a double-edged sword too, in that you both want to help people learn how to use data efficiently but you also want to enable them. I’ve seen cases where somebody gets access to the data and they just disappear for like four days, and they’re slicing and dicing the data, and they come back with some crazy conclusion or something like that. Although I think product is probably less… I’m not gonna call out… Yeah, what the hell, I’ll call it out. As I asked about marketing and product and you’re like, “Yeah, our product guys are awesome.”
40:07 Moe: No, marketing are, too. I don’t work with them as much, and they use a lot of different tools.
40:14 Tim: Yeah, and I think in many cases marketing is kinda looking at where’s the traffic coming from. That’s actually a little easier to be a little more formulaic, unless somebody’s just got a hair up their butt that they’re gonna do multi-channel attribution craziness and want you to punch the magic button on that. But when it comes to traffic source then realize, it seems like it’s a little more straight forward.
40:34 Moe: And I do also just wanna say, what you point just then about making sure one, people don’t get lost in the data, but two, make mistakes in it, is super important. Because our data set that lives in Tableau, there are whole series of flags that you need to apply and things that you need to know, that if you don’t do properly, your numbers won’t be right. And it does get really difficult managing that, because you obviously wanna enable people but you also wanna make sure that when you go to a meeting, six people don’t turn out with different numbers because they’ve applied different flags, and then it leads to more confusion instead of less.
41:10 Tim: But it sounds like if you’ve got them kinda conditioned to, if they’re gonna make some big pronouncement, they’re gonna wanna run it by you to say, “I’ve don’t as much as I can, can you take a look at it and see if it makes sense?”, so they have the stamp of approval?
41:23 Moe: I absolutely hope so. I hope that that is the way for it.
41:30 Tim: [chuckle] We did a thing where we, and this was as a team, but where we were actually basically trying to brand the stuff that we produced. We put a lot of work into our data visualization so that it was really clear to everyone that this came out of our team, because it was just going to look better, ’cause we had templates and standards. And if somebody else was just slicing away with their Qube and Power Play and dropping it into Excel, it would be kinda obvious that they had done it themselves. And that led to people would then question if somebody presented something else, they would say, “Have you run it by Shelly, the analyst on our team?” So that turned out to work…
42:07 Moe: That’s a great suggestion.
42:08 Tim: Work pretty well.
42:09 Michael: Yeah. Alright. Moe, we’re almost getting to the point where we’ve gotta close, but I wanna ask you one more question, which is: From your vantage point, what advice would you give to someone who’s just beginning their career in digital analyticAnnouncer
42:25 Moe: Well, Michael, I’ve actually written a blog on this too, so if anyone’s interested they can go read my blog.
42:29 Michael: Oh, nice. I have a feeling that we might link that in our show notes. Yeah.
42:37 Tim: It’ll be in the show notes, definitely.
42:38 Michael: Yeah. Perfect. Okay, yeah, but please continue.
42:40 Moe: Yeah, it’s a lot of the common advice, so read blogs, but also be careful about not getting inundated. My sister bought me a book and I think it’s because mainly she was sick of me asking her really, really silly questions, to just get the basic terms and that sort of stuff down. But then the place where I actually learned the most was by building a website, because as soon as I had to do that, I was like, “Oh, I need to now build tags,” and “Oh, I have a Google Analytics account I can play around in.” And I actually have an intern starting in a few weeks, and he’s already got a long list of things to start off on with those. So, yeah. Website is one of them.
43:20 Michael: I love it. And you know what? Every person I hire, I have them go tag a site of their own. Or make a website and go tag it. That’s part of the process. So that’s great, and I love that, and not just because it agrees with me, but I think it’s just…
43:36 Tim: I think that’s exactly because it agrees with you.
43:39 Michael: Well, I’m happy. Obviously, that’s great coincidence. Smart people everywhere, Tim. I’m just excited that I can call myself one of their number in that particular category. Well, this has been a really good conversation, and I think it’s really helpful to a lotta people who, frankly, are a lotta times underserved today in our community, because they’re taking on a unique challenge that a lot of us don’t continue to face today and so, thank you very much, Moe, for your perspective and thought on this. And we’re not finished yet, ’cause one of the things we like to do on this show, is what we call our last call. Where we just go around and talk about something we’ve seen interesting. Doesn’t have to be about necessarily analytics, but doesn’t hurt if it is. What have you got for last call, Moe?
44:29 Moe: Okay. So mine, few of your listeners might’ve found and used a lot before, but it’s a recent discovery to me and it’s changing my life, is a lot of the resources that Jeff from Jeffalytics has posted on his webpage. He’s just so great at giving you these tips and tricks to sort of save time. And like I said, I’m all about saving time and doing things faster. So, at the moment I’m kind of reading through everything he’s ever published.
44:56 Michael: Very nice. Alright, well, I’ll go next. Since I’ve been encountering R in our work lately, one of my team was walking me through some of the very cool stuff he was doing, and he was having an opportunity to work with a very great R, I don’t know what it’s called, is it a package? Yeah, package.
45:18 Tim: Package.
45:18 Moe: Yeah.
45:18 Michael: For Adobe Analytics, RSiteCatalyst, written by Randy Zwitch, and some others that have put quite a bit of effort into it. If you’re using R and Adobe Analytics, it’s hard to find nicer than that. In fact, it may be one of the best ways to explore the reporting API in Adobe Analytics, is probably what I would say there. And then one bonus item, because this I think, if Randy were to ever listen to this show, he would appreciate the combination of this, is I’ve learned recently about this thing called ‘The Snake Person Whoop’ which is…
45:56 Michael: It’s a thing that happens in music, and I don’t really know why it’s called ‘The Snake Person Whoop’, but I think it’s because it’s in modern pop music, because of modulation in tone. Anyway, it’s kinda silly, but it’s something I learned about and I like. Okay, Tim.
46:14 Tim: So my last call is gonna be the emerging sport of kabaddi. No, it’s not gonna be that. Although I did just learn about it.
46:21 Michael: Great sport.
46:22 Tim: So mine… It sounds like it is. I think there are a lot of sports that would be great if they took off. So, when I was in Australia, a little while back, there was a guy speaking at a conference I was at. And his name was Michael Yates. Works for ABC, which, those in the US, it is not that ABC, but it is media. Is it fair, Moe, to say that ABC is kinda like BBC? It’s the…
46:49 Moe: Yeah, kind of. Yeah. It’s government-funded media.
46:53 Tim: Government media. Media, including TV.
46:54 Moe: Yeah. The whole point is that it’s open and free media without any bias, because it doesn’t have to succumb to advertisers etcetera.
47:04 Tim: There you go. So, it’s kinda public. Government. But they use Snowplow, and he was talking about Snowplow and how they were setting up their event structure and he said, “Yeah, we just came up with our even structure and we were thinking we’d just kinda use the noun-verb connection.” And that got me thinking about how much I hate Google Analytics event category action label, and I especially hate the one example that is, no matter which Google search you do, you always wind up with the same fucking video click example.
47:31 Tim: So I was like, “Hey, that’s kinda cool. This noun-verb construct.” So I tweeted it and the next thing I know I’ve got Nico Miceli pointing to a blog post on Siri interactive that was about a better way of event tracking naming strategies. And then Simo Ahava jumps in and he’s like, “No, no, no. Check out this post by Alex Dean, who is one of the founders of Snowplow.” And it’s fascinating. It’s basically made me realize that the problem, the reason I hate Google Analytics event category action label, is that that’s not enough, and this whole concept is using grammar, the grammar of human language, and basically it’s like diagramming sentence, but dropping those into different buckets.
48:15 Tim: And it was a mental exercise, if you thought about, if you were using GA, and you just have five or seven custom dimensions dedicated to subject, verb, direct object, indirect object, prepositional object, and use that in lieu of event tagging, you could have this incredibly elegant thing, which… It seems like people who’re using Snowplow have to think through that syntax or that ontology, I think, is the fancy word for it. And it just sent me on many, many mental cycles of trying to nail a better way of a tagging action that is more sliceable and universally applicable, as opposed to, “You just have to get experienced, and then you can feel your way into what the best way to tag these specific actions.” So there were multiple people involved in that, but it led to a fascinating blogpost on Towards Universal Event Analytics: Building an Event Grammar.
49:16 Moe: Tim, when you started talking about that, I assumed you were gonna go in the direction of Michael’s other point that I have become obsessed with, which is people in the ABC submitting hypothesis requests instead of a Gyro-Tech request, which was my other thing that I absolutely loved about his presentation.
49:36 Tim: Yeah. And technically, he’s like an IT guy. And he just got up there and politely… He’d never spoken publicly, presented publicly before, and was just like dropping… It was a Google Analytics conference… He was like, “Ah, we don’t really use GA.”
49:47 Moe: Awesomeness. Dropping awesomeness.
49:49 Tim: And then he was just like, “Yeah, yeah. Here’s how we manage our intake for requests.” And I’m like, “Geez, wheeze!” So that was pretty awesome.
49:57 Michael: I’m writing down his name so I can…
50:01 Tim: We’ll link to him. He’s sorta on the Twitter. I think he’s a…
50:04 Michael: It’s great.
50:04 Moe: He’s not super on the Twitter, but I actually have tweeted with him. Yeah, so he does exist.
50:10 Michael: Nice. Alright. Awesome. And as you’ve been listening, you may have been forming some of your own questions or thoughts, and we’d love to hear them. And what’s awesome is you could find all three of us on the #measure Slack, and pretty active there, as well. And you can also reach out to us on our Facebook page or on Twitter, and we’d love to do that. We’re getting down to one of our last episodes of the year. If you’ve listened to the show this year, and you’ve liked it, we’d love if you would rate us on iTunes as well. But it’s not something that we demand of our audience, so it’s completely optional.
50:49 Tim: But do please rewind the podcast before you do turn in…
50:53 Michael: Be kind, rewind.
50:56 Michael: Anyway, Moe, a pleasure and a half to have you on the show. Thank you so much for making the time. And it’s very obvious that you are someone with a great perspective on analytics, and as you grow and keep going, and it sounds like probably in the near future will grow past a one-person team, but your insight into this has been something I think will be a huge benefit to our listeners. So thank you very much.
51:23 Moe: Thank you for having me, guys.
51:24 Michael: And, Tim, you were okay too.
51:27 Tim: One of these days, I’ll bring it.
51:29 Michael: Alright. Well, thanks. And from Tim, my co-host and I, remember, everybody, keep analyzing.
51:41 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, facebook.com/analyticshour or @analyticshour on Twitter.
51:57 Charles Barkley: So smart guys wanted to fit in, so they made up a term called ‘analytics’. Analytics don’t work.