Keegan O’Shea is a smarter data person. Here is a transcript of my interview with him.
In this podcast, my very first recorded, you’ll hear about how the best data insights are sentences, how deep learning is like making beer, and the three types of things you say no to.
Keegan O’Shea is cool and interesting. He’s the epitome of a Smarter Data Person. Here’s a medium article he wrote called Data Analysts need to turn up the heat, drawing from jazz music to illustrate his point.
You’ll find the actual podcast and the links to all of the things Keegan refers to here.
Cindy Tonkin: 00:25 Today’s podcast guest is Keegan O’Shea. He’s cool, he’s interesting, and he’s the epitome of the smarter data person. In this podcast, which is the very first one I ever recorded, you’ll hear about how the best data insights are sentences, how deep learning is like making beer, and the three types of things you say no to. If you want to know more about how to be a smarter data person, or if you just like Keegan, listen.
Cindy Tonkin: 00:45 Do you reckon we should launch?
Keegan O’Shea: 00:54 Let’s do it.
Cindy Tonkin: 00:54 Let’s start it. I’ve got Keegan with me. We are talking today about how to work smarter, faster, and nicer, and Keegan O’Shea is my very first guest. He knows more about tech than I do, and I’ll let him say who he is and what he does. Keegan?
Keegan O’Shea: 01:13 Sure, hi. So my name’s Keegan O’Shea, I am a data person, a data guy as I like to nominate myself. I work for a company called Lexer and I take care of the data team there. I’ve worked in various data roles over the years, data science, data analytics, marketing analytics, head count forecasting.
Keegan O’Shea: 01:35 Started off in call centers when I was younger, dropped out of uni and didn’t know what I wanted to do. I knew I wanted to do something with computers, and I did that. I realize I’m going backwards in time, and it’s clear I’m Benjamin Button in this story. I like making data accessible to people who don’t know what data is and how to use it. That’s what I try to do. I work for a company called Lexer, who try to do that a bit as well. So it’s good fit for me.
Cindy Tonkin: 02:07 Keegan and I met at the recent IAPA Leaders in Analytics Awards.
Keegan O’Shea: 02:15 Yes.
Cindy Tonkin: 02:16 And one of the panel there talked about the importance of Article 22 in GDPR, which was all about, in fact, making the information and the algorithms more accessible, how decisions were made and that kind of stuff. So this whole getting data more accessible is not just the new black but kind of a requisite of getting work done really.
Keegan O’Shea: 02:40 Yeah. That particular article talks about making sure that if a decision is made off your data, probably more for things like credit risk and that sort of stuff, it’s designed to make sure that someone has the right to understand why they made the decision. Because typically in history, if you’re denied a bank loan, and back in the 50’s it’s like you’re denied a loan because you don’t have equity in your bakery or whatever it is. Now it’s, we threw it at a computer and it did something, we don’t understand what it did, but..
Cindy Tonkin: 03:06 Computer said no.
Keegan O’Shea: 03:07 Computer said no, and it just seems like a massive chocolate wheel that someone spun and said no to. That’s sort of the reason behind that. But yeah, that whole regulation protection in GDPR is really designed to make sure people aren’t screwed over by the way data’s being used.
Keegan O’Shea: 03:28 I try to think of every which way but it’s just no, you’re just not getting screwed over. It doesn’t affect your human rights. Your data, while it may not necessarily belong to you, you have control over it. It’s putting that back and it’s … We can sit all day and talk about the pros and cons of GDPR, I’m not going to because-
Cindy Tonkin: 03:46 Oh no, I’m not interested in the pros and cons.
Keegan O’Shea: 03:50 It’s dull as … but it’s an interesting change. I think in terms of accessibility though, it’s, for me it’s not so much about from the person it’s about but more in an organization or in a society of making something complex simple to understand, which is yeah …
Cindy Tonkin: 04:07 Which is what we do.
Keegan O’Shea: 04:09 Which is what we do. Well, what we try to do. I find that people that work in data don’t do it as well as they think they do it. I think there’s a way to go for that to get there.
Cindy Tonkin: 04:21 What have you done, because you work in data, you have worked in data for a long time, what have you done to get better at explaining that?
Keegan O’Shea: 04:31 I try to put it in terms that my mom will understand. People always step back and go, “Oh, isn’t that insulting your mom?” And it’s not. I want to put it in terms than an intelligent person, who isn’t necessarily familiar with what I do, would understand it.
Keegan O’Shea: 04:44 Because you work with businesses and you work with people who are peripheral to data, but they’re not in it. They’re you’re person number 12 of 30 that they need to talk to to do the thing they’re doing. If you stop and say “Well hang on, let me explain to you what hierarchical clustering is.” They chicken out, they don’t want it.
Keegan O’Shea: 05:05 If someone’s coming to you, they’re coming to you because they want your help. So you want to be as helpful as possible and there’s no point getting all fancy and smart-assy about it, because you want to be able to assist in a meaningful way.
Cindy Tonkin: 05:19 So have you practiced explaining things to your mom?
Keegan O’Shea: 05:22 I haven’t actually. I don’t think she’d care, to be quite honest. I’m sure she’d get it quickly, she’s a sharp one. But I think the way I like to explain it is, I try to explain it to someone who’s not familiar with it at all in a business context, and they’ll go “Hang on, what are we talking about again? Where did you start with this?”
Keegan O’Shea: 05:42 We all are guilty of getting it wrong, but I try to step it back and put it in, I actually kind of simplify the language of, we’ve got a thing, we’ve got to put it next to another thing but if two things don’t talk to each other. That exact language, but they don’t talk to each other, you need this other thing to make them talk.
Cindy Tonkin: 05:57 So specific though, Keegan, are you sure people will get it if you use such specific terms?
Keegan O’Shea: 06:02 Well then it’s a little hard when people don’t get it at that point. If you know that I need to connect data A to data C but the only way to do that is if they both have information B, for example. That is just drawing in two files but it’s about how algorithms work. I mean people don’t really care. I don’t feel the need to explain how they work.
Cindy Tonkin: 06:22 No, I don’t. Yeah absolutely. But that’s not how you get people to get on board and understand it, by explaining the minutiae of it.
Keegan O’Shea: 06:33 Yeah.
Cindy Tonkin: 06:33 What do you do? How do you get people listening, even, rather than switching off?
Keegan O’Shea: 06:40 Well there’s two sides of data as far as I can tell. There’s the automation and optimization, which is all the software, the big data stuff that’s just spinning away in software companies and trying models and predicting models and all that trading.
Keegan O’Shea: 06:54 That’s one but you don’t really need to explain that to people. It’s just, hey there’s some stuff we’re doing at work, some would have saved you a bit of money or made you a bit of money, but it’s an optimization thing. There’s the other side which is guidance, or insight, people get those confused, often.
Cindy Tonkin: 07:10 Okay, guidance and insight.
Keegan O’Shea: 07:11 Guidance and insight versus automation.
Keegan O’Shea: 07:16 People in the data world, where I said we need to get a bit better at doing data and explaining data is that guidance is a whole field, and I feel that it’s really difficult to get that right, because it’s essentially communicating complex concepts simply, and that’s the thing that needs to happen to be better.
Keegan O’Shea: 07:39 I speak with people in my field and other data scientists and leaders and things and they complain about, “Well, if people only understood the lost genius that I am, and one day they’ll figure out what we’re up to and they’ll catch up.” It’s not up to them, it’s up to us.
Keegan O’Shea: 07:57 You look all through history, the early Apples, the Apple II and all those machines are really for nerds. And people say, “Trust me, this computer is really good!” But then … And it took the iMac and Windows 95 to get people on board. The awful mobile phones from the early 90s up to the iPhone.
Keegan O’Shea: 08:17 There’s all this history of technology, and this is, and data is just another technology of this is complicated, we’re going to make it easy. Until there’s a more structured way of making it more accessible, it’s going to be really hard.
Keegan O’Shea: 08:28 The example there of explaining what Lexer does, that’s sort of top line about a thing. If you’re using a metaphor to explain data, data is typically a detailed thing, and a metaphor I found is useful when you’re speaking in abstract terms, “Let me explain how a car works,” that sort of things. But it’s not talking about the RPM or it’s not talking about the specifics of speed and all that.
Cindy Tonkin: 08:53 Absolutely.
Keegan O’Shea: 08:53 Even the miles per gallon, that kind thing. That gets harder, when you go down to a little technical detail, which is what data storytelling is. You can’t, there’s no real emotive parallel to that. There’s no emotive parallel to the conclusion.
Keegan O’Shea: 09:06 If you say, your car gets 700km per gallon, that’s like a horse that never sleeps. It would be the conclusion that that is very good, it gets you far and you don’t have to stop.
Cindy Tonkin: 09:19 Give me your favorite data concept at the moment and let’s see if we can find a metaphor for it.
Keegan O’Shea: 09:24 I think my favorite concept at the moment that I’m still trying to wrap my head around is this concept of deep learning. It’s not really that new, but the way it’s being applied is quite new. It’s taking something that’s been around for ages, and putting it in new interesting applications, like image recognition and things like that.
Keegan O’Shea: 09:48 I feel as someone who’s had a background in building models and data I’m pretty well equipped to understand it. It took me a while to really get how it worked.
Keegan O’Shea: 09:59 I think the way I’ve landed on it is I would just think of it as a set of pipes, and then you’ve got water coming in at different levels and you keep tweaking the pressure of each of the valves until you get the right mix of water or beer or whatever, I’m sure there’s a great …
Cindy Tonkin: 10:16 It’s a great beer metaphor here.
Keegan O’Shea: 10:17 It’s great, yeah. So we’ve got the hops, barley, yeast, everything else, water coming through, you’re calibrating how much comes through and then you taste it and you go, “It’s not quite, let’s turn this one up,” and you randomize how much is coming through. At the end of it you get to a point, I actually think this is quite good but you’ve got several valves and different layers.
Cindy Tonkin: 10:36 And you constantly have to constantly keep checking, it’s not a-
Keegan O’Shea: 10:36 Well let’s get the hops and barleys coming together, let’s tweak down the hops and barley but the straight hops will turn up.
Cindy Tonkin: 10:44 Let’s put a little ginger in just for spice.
Keegan O’Shea: 10:47 Yeah, and it optimizes based on really quick feedback loops of getting that working. I haven’t thought about it like that …
Cindy Tonkin: 10:56 I love it when you think about something new on my podcast. You heard it here first, ladies and gentlemen.
Keegan O’Shea: 11:01 I think the interesting thing there though is, okay well what does that mean for me? I think that’s the thing, you could explain how models work and everything, but it’s like, okay well why would I use this over gut feel?
Cindy Tonkin: 11:12 Right, yeah, yeah.
Keegan O’Shea: 11:13 It’s, well it’ll be cheaper. That’s often what people want to hear, it’s either going to be cheaper or it’s going to make you more money, or whatever it might be.
Cindy Tonkin: 11:22 Yeah, because the decision makers have to make a decision based on some kind of criteria, and yes we need this kind of deep learning to run our thing. Thing is a technical term again.
Keegan O’Shea: 11:32 It’s a thing, yeah.
Cindy Tonkin: 11:35 It’s about how much, do I need 17 people thinking about it or can I just have one person who regulates the hops for the beer. It also depends who you’re explaining it to, I imagine.
Keegan O’Shea: 11:46 Yeah, it’s really good to assume the person you’re speaking to knows more than they’re letting on. When I was younger, I used to work in call centers as I said, and I was on the phone with someone and I was about 19 or something like that.
Cindy Tonkin: 12:06 So you were very wise.
Keegan O’Shea: 12:07 Very, very wise, I knew everything. I was talking to this guy and he was having a very particular issue with his computer. And I put the guy on hold and being an angry 19 year old, I hang up and I was like, “Oh this guy is a bit of an idiot.” Had to think about it and then I looked at his name and he was a doctor, he was doctor something had his own practice.
Keegan O’Shea: 12:26 I thought, well, this guy is clearly more educated than I am, he has his own field of study, he was a cardiologist. I thought, he’s not an idiot, he just doesn’t know about this thing. That’s one of those lessons that I’ve carried with me, it’s like …
Cindy Tonkin: 12:38 Yeah, and you’re now recognizing it. Thank God you learned it at 19.
Keegan O’Shea: 12:42 Thank God I learned it at 19. There’s a lot of people I’ve encountered over the years who never learn that lesson. I think that that was important, to never assume something about the audience.
Cindy Tonkin: 12:51 Well exactly. When I go to my GP and she refers me to a specialist, she makes decisions on whether the person’s a nice person or not, but I then come back and give her granularity on that, where it’s like “Yeah, nice person but treated me like an idiot. I’m not going back.” versus, “Yes, they assumed I had my own level of intelligence and expertise, it just wasn’t in dermatology, so they at least treated me like a human.”. I’m sure you have this experience of, I’m just going to say it, at school I was always the smartest kid in the room And you probably had a similar experience of being one of the smart kids in the room.
Keegan O’Shea: 13:41 Probably not as smart as I thought I was, but yeah.
Cindy Tonkin: 13:43 Well no, absolutely, because you’re not 19 anymore.
Keegan O’Shea: 13:45 Yeah, exactly.
Cindy Tonkin: 13:46 When you were 19 that was your peak smartness.
Keegan O’Shea: 13:49 Yeah exactly.
Cindy Tonkin: 13:50 That there’s a certain impatience that I have when I’m being treated like I have no education, no understanding. Someone tries a metaphor, I get that the heart is a pump, give me more than that, cause I can handle it.
Keegan O’Shea: 14:08 I don’t think anyone likes being condescended to, and it just depends on what level of condescension kicks in.
Cindy Tonkin: 14:16 Condescension, that’s back to the beer metaphor
Keegan O’Shea: 14:18 Yeah, well it’s all dripping into the water. I think people don’t like being spoken down to, but I think there’s a qualifying point of, you’ve got to get a read of the person if you’re explaining something to them and you can tell they’re going “Yes I know all of this.”
Keegan O’Shea: 14:36 I’ve seen it happen I’ve been presenting a scatter plot to a senior executive at a meeting once, and the manager at the time says, “Do you think the audience can understand this?”
Keegan O’Shea: 14:51 My boss at the time and I thought about it and thought, well, this executive did study quantitative theory at an Ivy League School in the U.S. so I think they can understand a scatter plot with a trend line through it. But there are other people who may not who might have a background in law.
Cindy Tonkin: 15:06 So how did you approach it?
Keegan O’Shea: 15:08 I just presented it and went with it.
Cindy Tonkin: 15:11 You assumed that that person might have-
Keegan O’Shea: 15:11 I assumed that person-
Cindy Tonkin: 15:13 It wasn’t five people in a room it was just one person?
Keegan O’Shea: 15:17 No, and I think if it were five people, for example, I’d say “Look, here’s a scatter plot.” If you’re in a position where you’re presenting say a PowerPoint presentation, you present the numbers where the numbers are relevant. I think the best data insights are just sentences, if you can’t-
Cindy Tonkin: 15:33 That’s nice, that’s some Tweet-able thing, say that. Say more.
Keegan O’Shea: 15:38 If you could say it in a text message, explain something in text message, that’s insightful. Because if you want something to be insightful, it needs to be easily understood and impact-ful.
Cindy Tonkin: 15:50 That’s a really good insight on insight. I’m excited.
Keegan O’Shea: 15:54 Numbers are good but… the company I used to work for, if I presented an insight that I thought was reasonably interesting, it was, “X percent of people buy these products”, a version of that. I heard that fed back to me for years. But for me it was just a couple of lines of code and I sped it through.
Cindy Tonkin: 16:16 Yeah, and you spent weeks, months, years, honing that one sentence.
Keegan O’Shea: 16:21 Yeah, that sentence just led of its way but the 20 pages of cogent argument about why we should do blah, blah, blah fell on deaf ears. If you can crystallize something that, because we hear that and go “That means we need to do more of this instead of this.”
Cindy Tonkin: 16:35 Is that why Twitter’s been so cool for people?
Keegan O’Shea: 16:38 Yeah.
Cindy Tonkin: 16:39 I’m not a big Twitter, I’m on Twitter but I don’t spend any time there.
Keegan O’Shea: 16:43 I think Twitter’s really funny because my company has a history of doing social media data, and I’ve spent a lot of time on Twitter. What I find interesting is the 140 to 280 [characters], I don’t think it really matters but if you put in a couple of sentences, it’s really interesting.
Keegan O’Shea: 17:03 I think it’s actually been interesting for comedy, for example, you get a lot of comedy accounts that have Mitch Hedberg’s style of one-liners and things like that. You’ve got a lot of, what I’ve found really interesting about Twitter from a data perspective is something like the gun control debate. A very emotive thing that 20 years ago would’ve just been “How dare you?” This, that and the other, a very emotive thing, but it’s …
Cindy Tonkin: 17:27 Well now it’s, “I’m unfriending you”.
Keegan O’Shea: 17:30 Yeah, but more so it’s like, how can you believe this. There have been 39 deaths in the last 15 days over this things and this was the thing and this is how much money we’ve spent, it’s a very enumerate argument.
Keegan O’Shea: 17:42 There’s a lot of really passionate arguments that happen online now that you see and there’s a lot of numbers in them. It’s not a table of numbers, it’s a sentence including numbers. Those are the things that go viral and get retweeted and everything else like that. If you can present that, I find that great. I love hearing about the stats of the infant mortality rate has gone down three, 30,000% or something in the last whatever period. I don’t want to quote the numbers but you read that, wow. You read that sentence you just pause.
Cindy Tonkin: 18:15 Exactly, so it’s one number-
Keegan O’Shea: 18:17 It’s just one number, in context.
Cindy Tonkin: 18:17 In the context, yeah.
Keegan O’Shea: 18:19 And it’s more useful if it means the audience can do something about it, or if it, unfortunately can be used to satisfy bias and go, “Oh yeah. Cool. He’s going to put this in my back pocket.”
Cindy Tonkin: 18:29 Well I do say a biro can be used to commit murder or a biro can be used to write poetry inspiring millions. So every tool has its ups and downs.
Keegan O’Shea: 18:39 Yeah, exactly.
Cindy Tonkin: 18:40 One of my pet projects is I’ve been developing a series of workshops essentially and resources around how do we say no nicely. My experience in corporate data analytics, with corporate data analysts, people in insights and analytics is that they have to say a no to a lot of things. Sometimes just because this is a stupid idea, this is a waste of our time but also it’s sometimes because we just don’t have the resources to do this.
Keegan O’Shea: 19:12 Yeah.
Cindy Tonkin: 19:13 What experience do you have with saying no to clients or stakeholders or life?
Keegan O’Shea: 19:18 I think there’s, I put it in three categories. There’s saying no because someone is just trying to pan off work, and it’s simply something they can do themselves but they can’t do and that is very simple of, “No, you do it.” And it becomes very clear if someone is just panning off something.
It’s just, “Hey can you …?” I had an example a few years ago when someone said, “Hey, I’ve got an Excel spreadsheet or something. Can you tidy this up so I can put it in a report?”
Cindy Tonkin: 19:56 No.
Keegan O’Shea: 19:56 That’s called Google, they’ll figure it out.
Cindy Tonkin: 19:57 Yeah.
Keegan O’Shea: 19:58 But that’s a separate … I think there’s then the other two issues. There’s the capacity problem and then there’s the importance problem. So if you have I don’t have capacity to do something, that’s just a resourcing thing and then you can go, “Well yes, but when? I can’t do it by that time. Convince me that it’s more important than these other things.”
Cindy Tonkin: 20:21 Yes.
Keegan O’Shea: 20:23 Like the brick is good, it’s a good thing to do but it’s not necessarily the most important thing. I think the last one, which is the most interesting one, is someone who will give you a brief saying, “Give me this, I want to know this number.” Yes and we can do it a better way. That standup comedy is a great routine as you know.
Cindy Tonkin: 20:40 Yes, there is that.
Keegan O’Shea: 20:42 Yes and.
Cindy Tonkin: 20:42 Into the whole improv of “yes and”
Keegan O’Shea: 20:44 You’re not saying yes but or no but, you’re saying yes and we can reach the same outcome. But I like this way instead. Because people are often asking questions because they think they know the outcome. They think that, “Hey, can you give me this number?”
Cindy Tonkin: 20:59 ”Can you prove to me that my hypothesis is correct?” Yes, an entire PhD would do that.
Keegan O’Shea: 21:05 Exactly. But sometimes it’s, “We heard this, we’ve got some researchers and that kind of thing, can you validate?” Sometimes that’s fine. A lot of data people look down their nose at it and go, “Well that’s just a simple query. The technical requirement for me to do that is beneath me or whatever.”
Cindy Tonkin: 21:21 My brain is much bigger than that.
Keegan O’Shea: 21:23 Yeah. I think that more than anything else that does do people a disservice because it means I could help you but I think I’m better than you
Cindy Tonkin: 21:32 Yes.
Keegan O’Shea: 21:33 And I really don’t like that.
Cindy Tonkin: 21:34 It’s because of that condescension thing.
Keegan O’Shea: 21:36 Yeah.
Cindy Tonkin: 21:36 We’ve got a theme of condescension coming up here.
Keegan O’Shea: 21:38 In my experiences in years of working with technical people, working in call centers, IT people, in 15, 20 years, even friends, the people I find that do well for themselves in a technical capacity are the more amenable friendly ones who want to help.
Cindy Tonkin: 21:59 Is that the bottom line? If you want to be a good data scientist-ish person you need to be nice?
Keegan O’Shea: 22:10 Maybe.
Cindy Tonkin: 22:11 Kind?
Keegan O’Shea: 22:11 If you want to be a data scientist building models and things, you just need to be good at maths and coding and stuff. But if you want to be facing people working in data, yeah you should be.
Keegan O’Shea: 22:21 I started out in IT and wanted to be a systems architect and solution architect for all of big enterprise systems. What I found is that in time I was trying to do all of that. IT went from this 90s idea of we can change a lot, we’ve got the internet. Rolling stones are singing the theme song for Windows 95. It’s cool, it’s happening.
Keegan O’Shea: 22:43 Then it went down to the pathway of support tickets and IT as a cost center and they will say no until you fund it, and it became a lot less creative.
Keegan O’Shea: 22:53 I worry that if data people aren’t more collaborative with marketers or strategy of business and whatever it might be, they’ll be forced down the same road of well, it’s a cost, we have to do data, we know we have to do data, it’s the cost of doing business and there will be prioritization discussion and blah, blah, blah. But anyway, it’s no, let’s work with you to solving your problems. I think that works well.
Cindy Tonkin: 23:15 So saying no to someone because the project is important and it needs to be better.
Keegan O’Shea: 23:22 Yeah.
Cindy Tonkin: 23:22 But yes and rather than no that’s really a yes and …
Keegan O’Shea: 23:25 Yeah.
Cindy Tonkin: 23:25 Yes we’d like to do this and … the way you’ve conceived it isn’t going to be as good as if we did it this way.
Keegan O’Shea: 23:32 Yeah. I’ve come … A lot of people in data don’t realize that you don’t have to use data. You can buy it. Businesses around the world, some of the biggest businesses in the world get by without any data.
Cindy Tonkin: 23:50 I think it’s probably that data, come on, keep it.
Keegan O’Shea: 23:51 I don’t want to say that in case they end up coming to try and … There are …
Cindy Tonkin: 23:51 Are there industries that don’t rely too much on data?
Keegan O’Shea: 23:55 I know that there are … I think it’s pockets of organizations. You’ll see some organizations that are light-years ahead in their supply chain. They’ve optimized this, that and the other. But in their HR department for example, they’re just in access data bases and weird stuff.
Keegan O’Shea: 24:10 I think there’s a cultural thing across organizations but decisions get made based on well it’s worked before, let’s do it again. The mature markets, difficult to disrupt that you don’t have to make informed decisions because they have gotten by so far off the back of their product.
Keegan O’Shea: 24:40 These businesses you’re like, “How could you disrupt that?” So you want to get ahead. But I think even then when those tech giants come at you as advanced as you are, you’re kind of screwed.
Cindy Tonkin: 24:50 Yeah. So you basically are ripe for a take over in the industry that are not already using data?
Keegan O’Shea: 24:55 Yeah.
Cindy Tonkin: 24:55 I have just worked for recently for a public sector organization doing strategic planning and they happened to mention that they were about to develop … and I don’t even remember what it was, but it was something where I was like, “Oh my God, everybody else did that in the 80s, what are you doing in 2015 not having …” It was something really basic like processes and procedures or templates for external communications or something like that.
Cindy Tonkin: 25:27 I think data would be one of those things that is like in dark recesses of some public sector organization, people are still making decisions based on excel or based on what we did last year or … It’s slowly creeping.
Keegan O’Shea: 25:41 I think even in data-rich organizations still it’s happening because this comes out to corporate politics, which is this irony thing. But you have a senior person who makes a decision that is amenable to their business or to their own career.
Cindy Tonkin: 25:55 Yeah, it suits me.
Keegan O’Shea: 25:57 They can justify it however they want and that’s the nature of it, that’s just what happens. Take that example of not using data, there’s a friend of mine who is a relatively senior marketer. She worked in a business for a very short period of time managing the telemarketing business. They didn’t have computers. This was three years ago.
Keegan O’Shea: 26:19 But that business on the surface of it doing really well. I think it comes down to what value, you know a lot of dumb people think well we can change the world and everything is possible with us. It’s well some things, not everything.
Cindy Tonkin: 26:35 Yeah, exactly.
Keegan O’Shea: 26:37 IT didn’t solve everything. We still have to communicate face-to-face. Teleconferencing still doesn’t work. We still use telephony. I think data will slowly chip away.
Keegan O’Shea: 26:52 But I mean, biases still exist. You can get all the information you want. There’s a lot of research out there saying even if you get presented irrefutable facts you’ll just knock it back.
Cindy Tonkin: 27:01 Yeah. You just confirm you’re in the beliefs you have.
Keegan O’Shea: 27:07 Yeah, the confirmation bias will just override and go, “Yeah, I already believe this. Oh there’s …” It’s something science might get down the climate science, right, because it will make me angry. But there’ll be slightly …
Cindy Tonkin: 27:19 And the data science as well. There’s plenty of people in the data science area.
Keegan O’Shea: 27:23 But there’s one paper that’s suggests something else, funded by someone else but I don’t care. That’s fine. We’ll take that position. So I think from that guidance and insight thing, data is a useful tool for those who are open to it.
Keegan O’Shea: 27:39 But it’s up to the data people to be able to communicate it in a way that, you know, if you’ve got advocates, let’s work with those people to figure out how would you like to interpret this information.
Keegan O’Shea: 27:49 I don’t need a 20-page deck or a dashboard of whatever. I just want you to tell me what to do. Oh great. We have that established relationship because you know I understand your business or whatever.
Keegan O’Shea: 27:58 But that only scales so far because that requires humans interpreting data, talking to other humans and I think data and software then has a part to play in sharing information and having a common language.
Cindy Tonkin: 28:11 Bottom line, humans are irrational.
Keegan O’Shea: 28:13 So I try to be in the office near Central at 7:30, 8 o’clock. My most productive time is from then till about midday.
Cindy Tonkin: 28:24 What’s the first thing you do when you hit your desk?
Keegan O’Shea: 28:26 Open my laptop, plug it in, then go make myself a coffee.
Cindy Tonkin: 28:29 Nice.
Keegan O’Shea: 28:31 We have internal workflow tools in our company. One called Asana, which is like a …
Cindy Tonkin: 28:36 Yeah, planning tool?
Keegan O’Shea: 28:37 Planning tool, tick boxes. I look at it and go, “All right, that’s what I’m going to do today.”
Cindy Tonkin: 28:47 Right. So you don’t pre-load anything, any to-do list, it’s basically what you’re going to do.
Keegan O’Shea: 28:47 Well it’s an ongoing tool. Like I did it three days ago and I said I’ll do that on Thursday. It’s now Thursday, oh, that’s what I decided I was going to do.
Cindy Tonkin: 28:52 So you have pre-planned this?
Keegan O’Shea: 28:54 Yeah. I plan out my deadlines where I can and before I go home and the day before I just look at seven things, I’m going to that, that’s Monday.
Cindy Tonkin: 29:02 Right.
Keegan O’Shea: 29:02 I try to have it ready for the day and start to smash through it and do as much as I can without the world interrupting me. I find it’s really effective and then try to, where possible, try to book meetings later in the day and try to head off around 5:00. Get home. I have a family.
Cindy Tonkin: 29:22 Because you got a family and little kids.
Keegan O’Shea: 29:25 I have two young daughters and so I try to spend an hour or two with them and then …
Cindy Tonkin: 29:31 Before dark.
Keegan O’Shea: 29:32 Before dark. Well at the moment, yeah. Put them to sleep and then I watch TV or lately there has been a lot of working until sleep just to do the stuff that I haven’t had the head space to do like coding or stuff that doesn’t require me to be in the office to meet with other people.
Cindy Tonkin: 29:47 Right. What’s your TV, what kind of TV are you watching at the moment? What’s your favorite if you don’t want to say what you’re doing right now?
Keegan O’Shea: 29:55 At the moment actually I’m watching GLOW Season Three.
Cindy Tonkin: 29:57 I just finished season two on Friday.
Keegan O’Shea: 29:59 No, don’t tell me anything. I try to watch a combination of just comedy. I’m really enjoying Netflix’s standup comedy specials, well certain specials. But my favorite show is always Mad Men, Breaking Bad. Those week-to-week cerebral weird shows that are really addictive.
Cindy Tonkin: 30:31 I love The Good Place. I’ve watched The Good Place maybe five times. It’s ridiculous.
Keegan O’Shea: 30:33 Yeah. So I’m re-watching having seen it. I’m re-watching now knowing the twists that occurs at some point in the show.
Cindy Tonkin: 30:41 Yes, it becomes … So if you haven’t seen it, watch it.
Keegan O’Shea: 30:44 Yes. It’s fantastic. That 22 Minutes, I’m going through it.
Cindy Tonkin: 30:50 Yeah.
Keegan O’Shea: 30:51 And yeah, documentaries about a lot of stuff, Dirty Money. It’s all on Netflix. So Dirty Money, Icarus, great documentary about doping. That kind of how the world works documentary.
Keegan O’Shea: 31:07 I love the books I’m reading of how the world works and how it used to work and all that.
Cindy Tonkin: 31:11 So you’re reading books. Tell me about what books you like to read. Or podcasts? I know you listen to podcasts.
Keegan O’Shea: 31:17 I listen to podcasts. I listen to podcasts and books. I try to consume. I’m not that big on Twitter. I like taking in music, movies, books, TV, music, just all of it. Books, I have a history of reading all sorts of stuff.
Keegan O’Shea: 31:39 Lately I’ve been reading books about history of humanity but not from a future-looking perspective. So Jared Diamond’s, Guns, Germs, and Steel, Yuval Harari’s Homo Deus and recently finished Steven Pinker’s, Enlightenment Now.
Cindy Tonkin: 31:56 Right.
Keegan O’Shea: 31:56 It’s all about, this is everything that’s happened in history to where I’m now and this is a result of that.
Cindy Tonkin: 32:01 Right, but he’s not ambitious at all, that guy.
Keegan O’Shea: 32:03 Yeah. But it’s all kind of we have progressed through the history of time for these reasons and this is what’s promising for the future and it’s dispelling a lot of myths around, like Guns, Germs, and Steel, there’s no racial, there’s nothing about race that dictates the peculiarness or whatever.
Keegan O’Shea: 32:26 Irony there is that it’s the sense that well geography and interaction and the Eurasian continent they were next to each other at the right temperature and they shared stuff so they were able to get advanced.
Keegan O’Shea: 32:37 The thing I really liked of those books, one of the concepts I liked that was relevant to my work was Homo Deus, Yuval Harari’s book with that usually a history of going from Homo sapiens to creating things and we’ve evolved beyond Homo sapiens. He talks about dying at the end of it in a really fascinating way. He talks about the idea that all of society is information processing systems. So churches …
Cindy Tonkin: 33:06 We are a metaphor for data.
Keegan O’Shea: 33:07 Yeah. So a church is a centralized data processing system, whereas a democracy is a decentralized because you’re sharing information and all the critic later is centralized and the idea is that as time goes on, all that is is sharing information.
Keegan O’Shea: 33:27 It’s I’m telling you a story, my story is that this $20 note will give you four coffees. It’s just a story we tell each other and we all agree in this social contract that it’s valuable and when you survive is that as time goes on, we’re getting more and more connected and all that data is getting more and more valuable but the barriers to that are people not accepting information.
Keegan O’Shea: 33:47 So I thought that was really interesting and it’s saying that as time goes on the value of the thing in future, speaking of the history of utility and service, the value of something is how much data it outputs, which I don’t necessarily agree with.
Cindy Tonkin: 34:01 Does that mean I’m more valuable if I tell more stories?
Keegan O’Shea: 34:04 I think you just stand outside and yell a bit it would be great. It would be really great.
Cindy Tonkin: 34:08 Great. I’ve got a soapbox, it should be fun. I’ll go down to the Domain.
Keegan O’Shea: 34:10 But if science puts out information it’s useful. Wikipedia is really useful whereas a rock may not be.
Cindy Tonkin: 34:19 Rock and roll is.
Keegan O’Shea: 34:20 Rock and roll, isn’t information distributed in the same way.
Cindy Tonkin: 34:28 So you’re also a keen music listener. You can’t be without music you tell me. Tell me more about that.
Keegan O’Shea: 34:30 Yeah. I realize it’s not necessarily music, it’s I don’t like silence. Silence is deafening and I think I do love music, and I think all of the music, my family, my mum is all the time is very into it, my dad is into music.
Cindy Tonkin: 34:49 Do you play music? Are you a musician who plays an instrument?
Keegan O’Shea: 34:53 Yeah. When I was younger I was in bands around town.
Cindy Tonkin: 34:58 How cool.
Keegan O’Shea: 34:58 Yeah, cooler than it sounds and yeah I played for electro sound.
Cindy Tonkin: 35:04 Unless you were the drummer. Were you the drummer?
Keegan O’Shea: 35:06 No. I was the control freak, singer and guitarist.
Cindy Tonkin: 35:09 Nice.
Keegan O’Shea: 35:09 I danced in the band. Actually my now partner, which is how I met her. I still make music now under the name Street Hawkers. Check us out on soundcloud.
Cindy Tonkin: 35:21 Street Hawkers.
Keegan O’Shea: 35:24 Yeah. It’s really just me on my laptop when my kids are asleep and I don’t want to look at work anymore. It’s just …
Cindy Tonkin: 35:29 So you write?
Keegan O’Shea: 35:34 It’s less structured than that but yeah. Yeah, it’s more depth.
Cindy Tonkin: 35:38 You create.
Keegan O’Shea: 35:39 I create.
Cindy Tonkin: 35:39 Music.
Keegan O’Shea: 35:39 Soundscapes.
Cindy Tonkin: 35:40 Right.
Keegan O’Shea: 35:41 Yeah, they are songs and recording. Recording is more my preference. I never enjoyed live performance really outside of when I was at a cool venue or whatever it might be. But yeah, the music production and audio editing and engineering was the biggest thing.
Cindy Tonkin: 35:54 What’s the thing you didn’t enjoy about live performance?
Keegan O’Shea: 35:56 Crowd sizes usually.
Cindy Tonkin: 35:58 Right. Not big enough or too big?
Keegan O’Shea: 36:01 I wish I had the luxury of getting an audience.
Cindy Tonkin: 36:04 Getting an audience is the hardest …
Keegan O’Shea: 36:04 Getting an audience was hard and I think that you rehearse for weeks and then you play a show and there’s 12 people in the audience, it’s somewhat disheartening. But it’s still fun. We actually called ourselves The Russian Brides because we wanted people to have trouble finding us online. When they searched our name they ended up with all sorts of porn sites. We were purposefully sabotaging ourselves. And things only started to go south when we started to get a little bit of success and we started taking it seriously and that’s when we stopped having fun. Yeah, succeeding in music is like winning the lottery. You can’t really do it.
Cindy Tonkin: 36:43 If you don’t enjoy the process.
Keegan O’Shea: 36:47 If you don’t enjoy the process it’s a really weird selection for a career. But no, I love listening to music, I love creating it. A lot of people when they know the technical of how something works they get jaded about it. It’s only made me appreciate it more. So it’s never going to leave me or anything.
Cindy Tonkin: 37:20 This is Cindy Tonkin on the Consultants’ Consultant and you have been listening to smarter data people. This is part of what I do to understand how it is that data scientists can be more effective in the workplace smarter, faster and nicer. I help you build analytics capability in your teams.
Cindy Tonkin: 37:37 If you have a team and you’re finding them harder to manage than they could be, if you’re constantly trying to squeeze more out of your budget and out of their time, and if they have got stakeholders who are less than happy sometimes, maybe a lot more than sometimes, it can be really annoying and it can make you feel incompetent, I can help you help them get to the important problems faster, target the wasted time and save you time and money. Ultimately delight stakeholders so that you can feel competent again. It’s such a good feeling. Talk to me.
You’ll find the actual podcast and the links to all of the things Keegan refers to here.