When did you last make a decision that made you say, now that was a decision based on the data?
This week on the If You Market podcast we talk with Nick Amabile about Data-Driven Decisions. Understanding the context so you can train your instincts to make decisions quickly when necessary and knowing how to make data-driven decisions. Office Space, Berenstain Bears, Scar Face, Moneyball, Princess Bride, Hitchhikers Guide to the Galaxy, AI, the ‘Data’ Pronoun, and more.
Nick Amabile (pronounced: AM-uh-beel) is the CEO of DAS42, a US data analytics consulting firm that helps companies make better decisions, faster. Founded in 2015, DAS42 is comprised of data analysts, scientists, business professionals, and engineers who provide end-to-end data services—including data strategy, tech stack integrations, application implementation, and enterprise analytics training.
Nick’s FullStack Philosophy is centered around the two components critical to achieving data-driven success: building an effective data analytics environment and building a data-centric company culture. He brings world-class analytics and big data technologies like those he used and built at leading internet
companies, including Omaze, Etsy, and Jet.com, to his clients. Nick is passionate about and skilled at building internal teams and transforming companies at various stages of growth into data-centric organizations. His technical abilities and product management experience allow him to understand eCommerce and technology companies on a deep level. This allows him to create value quickly and consistently in many different areas of business.
Contact : Nick Amabile
If you have questions about the If You Market podcast or would like to suggest a guest, please email us at info@IfYouMarket.com.
Listen wherever you get podcasts:
Transcript:
Sky
00:00:02 – 00:00:44
beautiful sound of recording in progress. Alright we all look great and here we go shit one second I got a fan in the background, I forgot to turn off this server and it’s uh it’s gonna make a little weird sound so I’m not gonna stop recording bucket.
Sky
00:00:44 – 00:01:13
There we go. Sound of silence. Mhm mhm mm hmm. Yeah. Which apparently Sting’s favorite song and that’s not even Simon and Garfunkel, he said the actual sound of silence. Oh come on. Really? Oh you’re on mute. I thought my headset wasn’t working. Oh okay. I
Nick Amabile
00:01:13 – 00:01:17
probably tried to clear my throat real quick or whatever and I forgot to turn it back on.
Sky
00:01:17 – 00:01:20
Alright yeah, last chance to fart off mike,
Nick Amabile
00:01:20 – 00:01:22
yep.
Sky
00:01:22 – 00:01:26
Here we go.
Sky
00:01:26 – 00:02:02
321 good morning marketers and welcome to the, if you market podcast brought to you by Mountaintop Data. We are the only podcast that markets the ship out of it. I’m sky Cassidy and today we’ll be talking with Nick Automobile Automobile should have checked with you before. Nick, can you help me with that Automobile Mika Mobile of Das 42 about whether you should really be making data based decisions or not. As I said, Nick, he’s the ceo of Das 42 that’s a data analytics consulting company and thanks for coming on today. Nick
Nick Amabile
00:02:02 – 00:02:04
Hey Sky, how are you? Good to see you. So first
Sky
00:02:04 – 00:02:14
thing that jumps out at me as I’m um looking at the topic here as I’m reading off the topic from my cheat sheet um
Sky
00:02:14 – 00:02:41
whether you should really be making data based decisions. I wish there was another word for that because the word. Data and bass are so close together. I feel like it’s, we’re gonna be doing some sort of techno geek version of who’s on first. Um, but anyway, that’s just just just me. Uh instead of data based decisions, um, somebody’s got to come up with another word for that. It’s too, it sounds like data based decisions.
Nick Amabile
00:02:41 – 00:02:49
It does. Yeah, we like to, we like to say data driven decisions sometimes, but that’s a little buzz worthy, you know what I mean? But but data based decisions, like I get what you’re saying.
Sky
00:02:49 – 00:03:19
No data driven decisions. There it is. We’re gonna change everybody’s uh I will, I’ll belief out anytime I say data based decisions in this episode now and we’ll, we’ll right over it with data driven decisions. Let me write that down real quick. Okay, so starting over, we’re going to talk to him about whether you should really be making data driven decisions or not. Um, so I guess right off the bat should you be making data driven decisions?
Nick Amabile
00:03:19 – 00:03:40
Absolutely. I definitely think so. I mean that’s uh, I’m on a mission to help companies make data driven decisions. I think, You know, a lot of times what we see though is that folks don’t trust the data and they end up arguing over whose data is right. Uh and and they don’t, they’re not able to actually talk about. What should we do about the data. So that’s really the issue that we see in companies and that’s really what we help people saw that dust. 42.
Sky
00:03:40 – 00:04:07
So yeah, I mean, coming into this, I didn’t know if you were going to say yes or no um talk about prep for the show, I didn’t know, looking at the question, I felt like this guy is going to say that we shouldn’t be doing it and I like that’s I’m sure there’s a compelling argument there. Um And I guess I’d say for some companies maybe they shouldn’t because they just don’t have enough data, like everybody is telling you you have to do this, but if you don’t have the data to make it on, then you kind of gotta go off your instinct.
Nick Amabile
00:04:07 – 00:04:43
Yeah, absolutely. I mean, I think there’s there’s, you know, I like to say that there are things that computers are good at and there are things like that humans are good at, right? So humans are great at synthesizing information, understanding context, understanding kind of a vision and broader goals. You know, computers are good at crunching numbers, right? So we absolutely need to have humans involved in the equation here. We absolutely need to have contexts and sort of instincts and and guts driven decisions as well. But what I kind of view is like, you know, data is context data is sort of um you know helps you understand the context of what’s going on. So it’s not that, you know, hey
Nick Amabile
00:04:43 – 00:05:06
the numbers are up today. So let’s do X, Y. Z. And I think a lot of people get caught up in that fact of of actionable data, right? Is data actionable? Well there’s a lot of parts of data that are just context. And so if you’re over time understanding a trend in your business using data that helps develop your gut and hone your instincts that you’re able to make decisions quickly based on data and and also you got at the same time it’s it’s a it’s a yes. And
Sky
00:05:06 – 00:05:26
so you’re honing your instincts with data and hopefully you get yourself calibrated so you can make good decisions when you don’t have new data or the data for that specific situation, but still based on data, if you’re doing a good job, otherwise you’re gonna be getting things wrong and that’ll be pretty obvious that you based it on well.
Nick Amabile
00:05:26 – 00:05:56
And the other piece too is like you want to be able to, you know, tweak something or change something and then measure the impact or the results or the change in the trend, right? So if you’re unfamiliar with the trend or you know like you know, the weather in december here in new york is typically let’s say in the thirties, but today it’s 60 degrees out here, right? So it’s sort of like, okay, That’s an anomaly. I know in December it’s typically the average temperature is between 30 and 40° 60° is outside of the, you know, historical average. So this is this is this is a, you know, a different trend than normal
Sky
00:05:56 – 00:06:31
right? So that makes me think as a as an analogy using the weather, if you have somebody who’s following the data but um using the weather in the time of year, but there the life cycle they have to collect data to act on is let’s say um spring through summer, right? Then fall in winter hit and they’re making decisions on the summer and spring data. Like if this then this I mean they’re gonna die. Yeah,
Nick Amabile
00:06:31 – 00:07:09
they’re gonna gonna walk outside in new york without a jacket or a coat or whatever. Right? I totally agree. And I mean that’s that’s the other piece to it. It really does depend on the business context. There are a lot of new products, new businesses, new business models out there. Especially if you think about kind of technology and understanding product market fit or coming to market with a disruptive type of product, Uh you’re not gonna have um you know, the data to actually be able to make those decisions. And so listening to customers, listening to feedback from the market is um is important. But again, there’s there’s part of that that’s qualitative and part of that, that’s quantitative, if you know, we were trying to sell our product at $100 a unit
Nick Amabile
00:07:09 – 00:07:37
and nobody’s buying it. Well that tells us you know hey we have zero sales and our prices ex you know we need to go back and think think of something and that’s why you know a lot of um more mature data and analytics practices include a pretty heavy user research component as well. So user research is very important. You know they have much smaller sample sizes but it’s also very qualitative data. Uh And then on the flip side we have quantitative data and sometimes the qualitative helps us better understand the quantitative.
Sky
00:07:37 – 00:08:27
So I guess the it seems like the person analyzing the data, you can do a lot of things with it. He said we’re not getting sales at this price point. So they could come in and say great let’s just start driving the price down until we start getting trails. Now maybe the buy button was broken on their site and they’re not getting sales because so it’s kind of also like what parts of data do you even have to make these decisions? It seems very dangerous. And this is where like the whole ai going wrong and killing everybody or ruining your business thing because the data sets, its program to use are incomplete or are you know like based on historically a period of time when you have cycles and it’s only taking part of the cycle or the industry is shifting and purely data
Sky
00:08:27 – 00:08:57
driven decisions only looking at hey what happened before. So this it’s like well when things change, what happened before doesn’t apply in the same and that’s I guess where the human element has to come in and you have to have a manager a an executive, a somebody with the intelligence and the information to analyze this and understand it. So you don’t make these catastrophic data driven decisions.
Nick Amabile
00:08:57 – 00:09:25
Yeah. And and you you kind of bring up two really important issues that I want to quickly talk about which is one first off data quality trust and data making sure that you um you know, you understand that your data is complete and and the other pieces of data silos we have and a lot of businesses you think about like a marketing team for example right? You might have some of your ad spend and Adwords some of your ad spend and facebook some of your ad spend on whatever you know out of home or whatever it is, radio tv.
Nick Amabile
00:09:25 – 00:09:47
And so having all these disparate data sets, you have to have a way to create a holistic view across these different channels in the in the marketing example. And the other thing you want to do is enable folks with the most domain expertise. In other words, the marketers in that example you want them to be able to ask and answer their own questions to be able to synthesize the data, they know the most about the campaigns. They’re running the channels that they’ve been working in
Nick Amabile
00:09:47 – 00:10:06
as opposed to a centralized I. T. Or analytics team. You know we don’t know as much about marketing as the marketers out there so we can’t anticipate all the things that they want to ask that we just don’t have the context. So that’s really the important pieces, creating a holistic view of your data across all different types of datasets and enabling those domain experts to ask and answer their own questions because they have the most context.
Sky
00:10:06 – 00:10:46
Alright so an analyst might look at the revenues from google adwords and say wow we’re only getting half the revenues we used to Um we should shift this or shift that but the marketing person might say excuse me we actually are only spending 10% of the budget. We used to so we’re getting multiple but we know we’ve saturated this and this and this so we don’t need to and we’re so good at it. Now we can actually get you know full results with with a fraction of the budget and this area over here was much more so again the partial data machine not being um the expert in the area and trying to look at the stuff very dangerous. So can you give some,
Sky
00:10:46 – 00:11:25
I mean so the topic is why you should be making data driven decisions. I keep seeing the word based type in front of me and what was that other word? Data driven decisions but I really want to dig into and we have been here like the when not to the Berenstein bears version of data driven decisions. Can you give some examples on um where include the ones we’ve already covered kind of but where data driven decisions go wrong? Like what should people be careful with stay away from when making these, trying to use data for decision making.
Nick Amabile
00:11:25 – 00:12:05
Yeah, so we talked a little about data quality, but the other piece is almost what we call data literacy. So understanding what the data means, what it’s telling us and and how it’s defined because for example, if I say, hey, uh you know sky, what was revenue yesterday for our company? You might have a slightly different viewpoint on what revenue means than I do. You might be in the accounting team and that’s way different than what I would call for revenue. Of course revenue does have a specific accounting definition but being able to say, Okay, here’s what exactly we mean by revenue. Here’s what we mean by customer and order. Like doesn’t include tax doesn’t include shipping. Is it somebody whoever purchased 20 years ago, is it somebody who purchased yesterday? So all these kind of nuances when you actually look at data
Nick Amabile
00:12:05 – 00:12:26
create a lot of confusion. And so I think what I see a lot of times is folks talking past each other when they’re working with data and they’re looking at the same report with same numbers and let’s let’s assume we agree on those numbers for a second. We may not actually understand what what the numbers are actually defined as and what they mean. Therefore. So we’re not able to then agree on a common framework to actually make decisions from.
Sky
00:12:26 – 00:12:35
And I suppose it could go as far as saying, oh, you scored, You know, a 92 and this guy only scored a 50. That’s great. You’re like, well those are golf scores. So actually
Nick Amabile
00:12:35 – 00:13:10
Yeah. Right. Exactly. Yeah, totally. Like you have to have the context, You have to have the context and you also have to understand, you know how each of those metrics and dimensions are defined. And this is actually a big problem that we see in companies that we work with at least 42 is that there’s a gap between the business team and the technology team. The technology team. You know, they understand the true databases, you know, the data flow, like how it all gets collected, how it gets stored and transformed, but they don’t actually end up using the data for most analytics. That’s typically the business team and the business team of course they understand a lot about the business context, but they don’t understand about that
Nick Amabile
00:13:10 – 00:13:21
Technology. And so that’s really what we do at just 42 is try to bridge that gap to understand first and foremost what are the business problems. And then we start working backwards to start actually using technology to solve those problems.
Sky
00:13:21 – 00:13:30
You guys are a version of the guy in office space that takes the specs from the customer and brings them to that. But why can’t the connecting these people?
Nick Amabile
00:13:30 – 00:13:36
Exactly right. Take the specs right to the engineering because they don’t have people skills. It’s
Sky
00:13:36 – 00:13:46
somehow that’s come up multiple times on this podcast. I love that particular scene because it’s so frequently misunderstood as well
Nick Amabile
00:13:46 – 00:13:47
by the viewers.
Sky
00:13:47 – 00:14:37
Um And I found out on an episode here that that’s a thing within I. T. Circles knowing that that the scene is commonly misunderstood by non I. T. People right? I was like oh wow I thought I’d recognize that. I didn’t know that it was a common thing. They’re like yeah that’s like a meme for I. T. People everybody else doesn’t understand what’s really happening in this scene. And uh he thinks it’s the opposite point. It’s kind of like the people who who love scarface and think that he’s a role model. Yeah I think you guys stop very deep about the movie kind of like kills everybody he loves gets everybody killed. And it turns out he’s not a good guy. He’s not a role model. He’s not you know
Sky
00:14:37 – 00:15:31
um it’s kind of like an explosion. It’s cool to watch but you don’t want to be in it. Uh Anyway so things to look out for when making data based decisions. Uh First things we covered that that you mentioned here were um just not having the experience I guess to analyze it, making sure everybody’s on the same page knows what the number even stands for and whether it’s better if it’s higher or lower. And uh the definition of the terms you’re using that could be going all over the place, which gets me to the I mean just the word data, I regularly rent on this, it’s it’s a pronoun and people throw it around like it’s all the same thing. And to me, if you go into a conversation with somebody, you don’t first establish what type of data you’re talking about, then
Sky
00:15:31 – 00:15:58
you’re just talking gobbledygook like you have No it’s it’s like having a whole conversation about this guy, but never saying who this guy is, like I need the name in the beginning before you can start pronouncing the person. Um so I know who we’re talking about at all. Otherwise the whole conversation is a bunch of he’s and she’s and days, but you don’t know who any of the characters are. You don’t know when you’re talking about the same person or a different person and I think there’s a lot of that in data.
Nick Amabile
00:15:58 – 00:16:26
Yeah. And it’s like I said, you know there there is like, you know kind of experience is data, right? You know, if I if I walk outside and I figure out what the, you know, the sun rises in the morning kind of situation you know I I can start to you know piece that together and say every morning I walk out at six a.m. And I can see the sunrise right? Like okay that that becomes becomes my experience and also you know some data over time right? So I understand the trend and that’s really you know the human element and I always view data and then you know even with the context of um
Nick Amabile
00:16:26 – 00:16:42
you know Ai you kind of mentioned that you know in the future robots taking over and stuff like that but I kind of view it more as like a I will still augment human decisions right and people will be augmented and be able to do more with Ai Ai won’t do everything will still need humans for things. So
Sky
00:16:42 – 00:16:43
that’s that’s kind of
Nick Amabile
00:16:43 – 00:16:44
my take on that generally.
Sky
00:16:44 – 00:17:18
Yeah and that seems like one of the main areas when the data driven decisions goes wrong is when you treat it so robotically with the impartial universe of what you know um you end up coming up with wildly inaccurate conclusions about stuff. Um So which leads me to an interesting like um tell me if this is a thing within data analytics and stuff but this whole issue of when you have the number 150 versus the number 100.
Nick Amabile
00:17:18 – 00:17:19
Yeah
Sky
00:17:19 – 00:18:07
um is one is 50% more But the other 30% or 33% or whatever less. Um So when you’re looking at the data things can be interpreted in wildly different ways and if people processing it don’t, you know how to you could also kind of pervert the data in a way over time by Analyzing it up and then down and then you don’t even end up at the same like oh let me add 50%, let me take 50% away. We’re not back at the same number. And it seems like kind of where a firm like yours is important because you guys coming in would know that kind of a number or data crunching problem. Is there a word for that in your industry? Am I making this up?
Nick Amabile
00:18:07 – 00:18:26
No, no. I mean like what we this is kind of all what I’ve been talking about is like we put this under the umbrella of data governance right? Which is you know, there has some data cataloging, so I know what the data means, where to find it. Uh data quality. So I know that it’s you know trusted and complete and other things like that. It’s you know timely.
Nick Amabile
00:18:26 – 00:18:59
Uh and then there’s this change management aspect where I know that if some change happens to the data um you know somebody makes a change to a definition that that it gets sort of trusted and transparent and throughout the organization. So that’s that’s kind of the key and I and again like this is why I kind of started a consulting firm was too because I think the human element was was missing like the sort of experience and the human side of it was missing this great technology out there uh these days with data and analytics but it still takes somebody to put it all together to make it uh make business value out of it and to make it make sense to to folks. So
Sky
00:18:59 – 00:19:07
So 101 50. Is it one third? Are we having a one third difference or a 50% difference here?
Nick Amabile
00:19:07 – 00:19:35
Well you know it depends on what your numerator is and what your denominator is. Right? That’s that’s the key. And so you know again like thinking about it and saying okay well where we at 1 50 then we went to 100 or did we start at 100 go to 1 50 then you think about you know kind of the denominator, the numerator in that. And and again it it We have to know the business context for those things if you just tell me 101 50 like that. It’s meaningless to your point earlier. Right? We’re not gonna have a productive discussion about what to do next.
Sky
00:19:35 – 00:19:45
Right. So are we up 50% or were we 30% lower before? Pretty much. Right. Exactly. It’s hard to you have to have a directionality I
Nick Amabile
00:19:45 – 00:20:08
guess you have like a time dimension right? So then you start to have this dimensionality where you have like you know today it’s 100 yesterday it was 1 50. So we you know, you look at it on a line chart or you know you look at it on a table and it would be ordered by date and you can say, okay, we’re moving from, you know the first to the second or 2nd and 3rd and not the other way around because we all know that time advances kind of thing. So yeah,
Sky
00:20:08 – 00:20:17
you can’t go back in time. We’re always going forward. Don’t try to reverse crunch the numbers and screw everything up if use the crap out of us. Right? Exactly.
Nick Amabile
00:20:17 – 00:20:37
Which is which is which comes back to that data literacy aspect of understanding how things are defined if if you give me a growth rate, a growth rate. I needed to define that to say it’s today versus yesterday or today versus last year or this Tuesday versus last Tuesday or whatever it is right? There. There has to be this context and a definition specifically for it.
Sky
00:20:37 – 00:21:25
I feel like I’m going to lean into statistics a little bit here, but that’s I think part of what you guys do, it sounds like it frustrates me when you also see things and I see it in the economy and stuff like that these days. But we get this this data these numbers and you see things like um the increase in, in whether you’re talking about crime, whether you’re talking about unemployment, whether you’re talking about uh any of those kind of things and we have this anomaly over the last couple of years with Covid and so it seems this is more of a like sociopolitical thing. But I think maybe the same kind of issues come up in making data driven decisions um
Sky
00:21:25 – 00:22:13
where it’s really easy to say, you know, the job numbers increased. Um you can have the largest percentage ever or there’s a big increase because there was a decrease before it’s like, well that’s not really um it seems like it needs to be done against the long term average or something like that, not against yesterday’s outcome. Where you can say, well if I just really ship the bed yesterday, I can make myself look amazing today by doing a mediocre job or if I did amazing yesterday. Now I’m gonna look shitty today if I do an above average job still like that’s kind of ridiculous that it seems like is that just purely politically things being analyzed for those purposes or like what’s going on there man? Yeah, I
Nick Amabile
00:22:13 – 00:22:49
mean look, I’m not going to ascribe motives to sort of how people, the government reports statistics, but I mean to your point this is this is a very good point around how folks can conceptualize and contextualized data, right? And you know, if you actually do drill down into the government statistics, they have very detailed documentation around like hey we include this, we don’t include this. Like you know obviously recently inflation is kind of a hot topic in the news. And so I saw some discussions online about inflation and you actually start to look at what they include, what they don’t include. And you know, then you can have opinions on whether you think that that’s appropriate to include or exclude certain items from, you know, inflation for example, but
Nick Amabile
00:22:49 – 00:23:24
at least they haven’t documented and you know to your point before, you know, a lot of the data visualizations in the news. Uh you look at you know new york times watching post big big news outlets tip. Typically they do show you like the trend, right? They say, okay here was pre covid and we’re actually now comparing, you know, jobs to pre covid levels or whatever it is. And, and and like stocks to write, you look at, you know, you watch the NBC, they’re always like, you know, what’s the 100 day moving average 200 day moving average and like where is that in relation to where the stock is today because if I tell you, hey the stock is worth $100 today, you’re like, I don’t know if that’s good or
Sky
00:23:24 – 00:23:26
bad. Right?
Nick Amabile
00:23:26 – 00:24:00
Yeah right. I need the additional context and that’s really what we’re talking about here. And so I think there is there is there is this aspect of what we call storytelling with data which I think is kind of underrated but you know being able to provide enough context and to draw conclusions from it and to you know also allow folks to um make their own conclusions from data as well because because we’re all going to interpret things slightly differently but you know a lot of times and I think your point you know about some of the job metrics and things, you know they are politically sensitive folks do want it to be good. Right? And so they’re going to tell a certain story with the data right? Some people
Sky
00:24:00 – 00:24:28
want it good and to make this applicable to the listeners. I mean forget about about the the U. S. Politics and stuff. Office politics get people you’re trying to make data driven decisions but you get people who are messing with things. Something that meant something last you know people say hey lead conversions we’re going to look at. Well the marketers decided that something is not going to count as a lead until it’s been market verified.
Sky
00:24:28 – 00:25:05
So now all your statistics are blown away because before anything that got filled out the form was counted as a lead but now they’re cleaning the stuff out so it looks like they’re doing a way better job. They’re actually doing a little worse job with lead generation. Um but they’re pre filled throwing out certain stuff. Um So it looks like conversion rates are way higher lead quality is way higher and it’s it’s, you know, it’s like, that’s why you don’t change much in baseball ever because you don’t want to ruin the stats. You have to compare people in different eras and stuff like that. Um, It happens internally, but you have internal politics, people do things for a reason that can screw up the data you’re trying to make decisions on as well.
Nick Amabile
00:25:05 – 00:25:40
Yeah, well, in this, you know, let’s, let’s for a second give give folks the benefit of the doubt and say, you know what, maybe that was the right decision to make, was like whatever, like we’re gonna give them the benefit of the doubt. This comes back to my idea around trust and transparency and train change management, right? You have to be able to say, hey guys, we’re gonna change the lead, uh definition of a lead to be not just somebody who feels out of form anymore. We’re gonna say it has to be whatever qualified to a certain level now form plus qualification. We’re gonna change that. Here’s why we’re going to change it because it’s important for our business and, you know, we have to now convince folks to do it. Um and now we need to make sure that
Nick Amabile
00:25:40 – 00:26:08
that lead number is updated in all the other reports so that we can compare it and we can say, you know, if we actually go back and sort of adjust our lead score for last year. Uh You know, we’re going to get a different number and so now we can actually kind of compare where we’re at this year versus last year. Uh So anyway that that trust that transparency, the change management aspect and and you know, it’s really an organizational problem has nothing to do with technology. Right. A lot of folks think that, oh yeah, I can solve that with technology really easily now. It’s it’s a process and organizational,
Sky
00:26:08 – 00:26:49
I would say they’re a employee um uh skill problem or office politics problem because either the person doing that, it wasn’t good enough at their job to know they need to everybody, everybody knows something foundational that these decisions are being made on has changed or office politics, they don’t want people to know because they’re trying to make something they’re doing look better or something somebody else is doing look worse than it actually is. And there’s a ton of that, especially when you get into the size of companies that want to do. I mean most companies are small and they don’t do a lot of data analysis. It’s all like I’ve gathered
Sky
00:26:49 – 00:27:04
by accident, little bits of data over time and I just make decisions on the fly based on that. If they’re really good at that, they’ll do well and if they’re not they fail. Um So we’re kind of talking about larger companies most of the time with the state analysis, would that be accurate?
Nick Amabile
00:27:04 – 00:27:27
Yeah. I mean I I’d say that to your point uh spreadsheets and you know, google analytics and some of the basics worked up into you know 100 couple 100 people company. And then basically beyond that, you know it becomes difficult for everyone to have the automatically shared context. Right? And so now you have to have programs in place to start to document definitions and you know, you know all that stuff that we’ve been talking about
Sky
00:27:27 – 00:27:42
And the majority of companies are 10 employees or less. So now we’re really talking like oh okay go with your gut is not the right word. I know we had a past guest that would crucify me if I I say go with your gut but go with your that which has been trained over years of
Nick Amabile
00:27:42 – 00:27:43
collecting data and
Sky
00:27:43 – 00:28:23
analyzing it. So those that majority of company, this kind of data analysis really just kind of not in there the scope of their work, they don’t need to worry about it. And I think it seems one of the things you brought up earlier if if all these very small companies um were to try to make data based decisions. They spend a lot of time collecting the small amounts of data that were available to them even trying to crunch it so they could feel like they were doing the best practices and come up with probably the wrong decisions or something that you have to ignore because there’s just not enough information there. Um, in the, in the size of the company.
Nick Amabile
00:28:23 – 00:28:46
Yeah. Well, you know, I’ve worked for 10 person companies and you know, all this kind of stuff. And like nine times out of 10 you probably have something more pressing to do like a fire to put out or you know, product to develop or something else. So you have to move very, very quickly. And at that point you’re, you’re really prioritizing like speed versus accuracy and those kind of things. And to your point, maybe you just don’t even have the data, maybe you have it, but you know, there are other important things to focus on
Sky
00:28:46 – 00:29:17
right and it’s a lot of the data is few and far between and so it really doesn’t give you any valuable context change things change so fast and companies outside, so I’d say probably a lot of our listeners are in small companies and don’t worry if you’re not doing massive amounts of data crunching to analyze this stuff. Maybe this is good to go about, but I don’t feel like you’re a marketing fraud and you need to change everything, spend a bunch of time on it and then get the wrong conclusions and have to throw them out anyway. This is kind of medium to large sized companies
Nick Amabile
00:29:17 – 00:29:53
and I’ll and I’ll make folks out there feel feel a little bit better. Nobody has this stuff figured out, right. Like no, like it’s very, very few and far between that we see a company. I mean it’s, it’s much harder even in larger companies to like manage these types of programs and to really, you know, institute, you know, data driven culture because of all the challenges that they have. And, and you know, you think about it as well. Most, most smaller companies, they’re just more agile, they’re much more, you know, tech forward or modern. A lot of, a lot of the companies are, I think newer and born in the cloud, whereas, you know, if you go to a big company that we work with, a lot of those folks have legacy systems and you know, just a lot of tech debt, entropy, etcetera.
Nick Amabile
00:29:53 – 00:30:01
And, and that’s, that’s very difficult to untangle all that. So I’ll say no one out there is doing it like super, super well and there’s very few, at least, let’s put it that way.
Sky
00:30:01 – 00:30:38
And I went out going back to the movie Moneyball and the analytics stuff there. I think there’s a reason that happened in baseball and not in a sport, another popular sport, even like, like football, there’s so many less variables and you have so much, so many stats because you have so many more games in a game like football where there’s so few games over a season, it’s much less likely that you have relevant enough stats where there aren’t these variables that can throw things off, there’s just so many other variables, so many moving pieces at any given time on the field,
Sky
00:30:38 – 00:31:19
the data is just not nearly as valuable. Um Baseball is just this fixed thing where people are literally fixed for most of the time sitting there and knowing exactly what to do on any place at any given time in any scenario. Um there’s just not a lot of moving pieces, but there’s a ton of stats and it can be really accurate. Um So you’re able to apply that kind of thing pretty effectively there. You don’t have a lot of game theory, you don’t have a lot of kind of um human adjustments happening. Whereas in something like football, you have a coach on the other side who’s, you know, it’s a scene from Princess Bride where you have
Sky
00:31:19 – 00:31:58
the guy sitting there saying, oh but I know you’re thinking this, so I’m thinking that, so you’re thinking I’m gonna do this and then that and this, it’s just like a this yes or no type type question, Some sports, there’s a lot of that, a sport like baseball, there’s not. So you can really go with the data driven decisions there. I see sometimes um you know, in a sport like football, they say, oh this is what the numbers are telling them to do in this case because percentage wise when this than this and I look at it and say, yeah, but something feels off about it and why we’ll probably the reason when teams
Sky
00:31:58 – 00:32:20
go for it in this situation and get it is because it’s a different scenario. Like you guys suck. That’s why you’re in the situation, you’re not going to complete it as much as the average team. So for you, the numbers are very different. You’re just having to take everybody’s numbers and put them together because you don’t have enough of your own stats and you’re on a backup quarterback and he’s got, you know, one leg, he’s not gonna sneak it for three yards and
Nick Amabile
00:32:20 – 00:32:41
a well it comes back to context and, and you know, like you’ve, you’ve seen all the, you’ve seen all the stats now that they are trying to put it into football and it’s it’s interesting. But to your point, I think there’s so much randomness. There’s so much athleticism and other things like that that come into play that, you know, if you put time tom brady, you know, in at quarterback versus some, you know, third string quarterback, you know, the stats are just different, right?
Sky
00:32:41 – 00:33:10
But I think that’s well, and the analytics say you run it up the middle on fourth and one here. But the problem is the other team’s coach and that’s the game theory part knows the analytics say that and they know you’re going for. So now they’re gonna stack everybody in the middle and you’re gonna say, oh, so maybe if we get Princess Bride again, so maybe we should actually do a sweep or a pass play and let’s throw it deep because they’re just going to leave the wide receiver wide open. But what if he knows you’re going to do that? So now they’re going to do this and not cover the middle and then they say, okay yeah, the analytics don’t really work because you have this human
Sky
00:33:10 – 00:33:53
element here adjusting to what they think you’re going to do based on those things. So when you get the coaches who just go by the numbers, like congratulations, you made yourself so predictable. You know exactly what you’re going to do. Um, you know, in baseball, if it just said, oh, the guy strikes out on this pitch, this percentage of the time. So you knew with two strikes, the picture was always going to give you that pitch. It doesn’t work anymore. You know, the same thing happens in marketing if you do what worked yesterday and everybody else doesn’t work yesterday, it doesn’t work because everybody’s doing it now. Um, so I think getting back to the, uh, try to go off on a tangent there. But the, some of the problems with purely data driven decisions, you have to account in the human element
Nick Amabile
00:33:53 – 00:34:30
Well. And I think the other thing too is like you kind of mentioned, um, you know, just the ability to separate the signal from the noise with data to write because especially if you talk about large companies that we work with at least 42, there’s so many different data points that they have available to them. Um, but focusing on what matters is actually very, very difficult. Right? So it’s like you can throw everything into a report or excuse me, a dashboard, uh you know, or a data science model or whatever you want to have and it doesn’t necessarily give you any more signal. It’s, it’s all kind of noise. And so you have to be understanding like again the business context to say,
Nick Amabile
00:34:30 – 00:34:51
hey, our goal is to whatever increase lead gen this quarter or to convert more leads or whatever it is, right? And then now we have to identify, hey, what are the metrics that matter here? And let’s identify those and focus on those as opposed to, you know, putting every number under the sun on a dashboard on a screen. That’s that’s not gonna be helpful and it’s not going to help people make decisions better
Sky
00:34:51 – 00:35:41
when it seems to go full data driven is really to say I analyzes it. Um, and that’s where the human part is great because the human part can no, when you know, if aI goes full data driven, when you first start inserting the data, it’s going to be massively wrong if they had every bit of relevant data in the universe. Then a I could rant like a machine learning could get come to the right conclusion, the right decision on stuff but your data set is always limited. So the human element, it seems incredibly important for all companies regardless of how much data you even have of saying, hey, I know this is limited and then I like I know that some there’s a change coming here in the future. Like we can see the future as humans. It’s kind of awesome.
Nick Amabile
00:35:41 – 00:36:08
Yeah. Well I think that I mean well you said it right there was like, you know, humans are still putting the data into the model and deciding what data goes into it. So it’s like the weather example before it’s like if you just, you know, put a machine learning model for only you know january or june, you know, you’re gonna get a much different results right? And and if you put, you know, for example this is talking about ethics and Ai and machine learning. It was a hot topic these days. You know, if you only put data into a machine learning model from a certain demographic
Nick Amabile
00:36:08 – 00:36:38
or if you don’t understand that hey, actually certain demographics are underrepresented in the dataset that I do have for whatever reason, which are, you know, have nothing to do with the outcome or more societal or political type type things then the ai model is gonna be gonna be biased and you know, that’s where people get nervous because we’re removing the human element and and in fact we’re actually, you know, you could almost reframing the other way your urine putting too much of the current, you know, human biases into the model. Of course the model’s gonna be biased at the data. We put it in as biased,
Sky
00:36:38 – 00:36:48
right? If it’s based all off of faces that look like mine and no faces that look like yours, the ai is gonna not think you’re human because you have a beard and I don’t. Yeah.
Nick Amabile
00:36:48 – 00:37:07
Right. Exactly. It’s something silly like that, right? But it’s like you can see kind of, you know, it’s sort of a reductive kind of piece of logic. But of course it makes total sense. Right? That that that that you know, the models are just going to represent the data that we put in there and that’s and deciding how to put the data into the model and what to put into the model is going to affect the outcome.
Sky
00:37:07 – 00:37:40
Excellent. Well if nobody, if we haven’t caught on yet, um we’re not gonna take a break in this episode. It happens. We just get into the flow of talking and there’s not time we get way past the breaking point. Maybe we’ll have one shoe horned in back there somewhere and you’ll say that was a weird break. Well that’s why um So I want to move into the post break discussions here and nick, I’d like to um I haven’t said your name and a guest to hear it nick probably cause I’m afraid of your last name. Amable. Automobile. Yeah.
Sky
00:37:40 – 00:37:59
Nick nick and mobile. Um Let’s jump over to you and I want to talk a little about you where you came from, how you got to where you’re at uh you know ceo of a data analytics company. Um So it’ll start wherever you want to in your story but let us know who you are a little bit.
Nick Amabile
00:37:59 – 00:38:33
Yeah. So you know I think we’re going back to college. I went to school for economics and uh you know it’s one of those things that I I think uh now working in data analytics for a very long time. A lot of folks come from social science backgrounds, traditional engineering backgrounds actually don’t see a lot of people coming with computer science backgrounds into the data analytics world because you know you do have to have that that sort of dual sides of the coin, the business and the technology aspect. But I went to school for economics and you know I got my first job at an economics consulting firm doing econometric analysis for antitrust litigation. Now correct me if I’m
Sky
00:38:33 – 00:38:45
wrong. But economics I always think money I think economics of course but it’s really kind of the analysis of data that doesn’t apply to non monetary things as well. It’s just money is where there’s a lot of numbers.
Nick Amabile
00:38:45 – 00:39:17
Yeah, absolutely. And it also has to do with you know we’re talking about game theory. You know it took some game theory classes and it’s all about like understanding like statistics, probability. Um You know you think about regressions right? You think about machine learning these days like most of it’s like you know based on you know pretty simple regressions and things like that. So all the statistics that I learned really did prepare me well for for a career in data analytics. Um And you know actually one of the cool things that I learned was statistical programming. So you know writing like sass and data for those folks that that kind of remember that stuff
Nick Amabile
00:39:17 – 00:39:58
um That really you know interests me a lot because I was able to actually kind of apply what I had learned in statistics and probability and all my economics courses to actual data problems and you know at the time we were mostly working with like you know census data and stuff like that to to run different models on on census data. But I really took to the technical aspect. And so that was really kind of where I got my career started. And you know after leaving the consulting firm uh the economics consulting firm that I was doing antitrust litigation and um went to work at a number of different startups. Uh you know always in a quantitative type of role. Um Supporting marketing, supporting finance. All these different product analytics right? All these different types of things
Nick Amabile
00:39:58 – 00:40:20
like A B. Testing for example. Right? You apply your statistics background to to a. B. Testing and product analytics. Um And so that was really how I got my start. Uh and you know it’s funny because one of the companies I worked at, I didn’t know sequel which for a lot of folks that don’t know that that’s that’s a programming language that you used to interact with databases which is really kind of the lingua franca of data analytics.
Nick Amabile
00:40:20 – 00:40:43
And I didn’t know how to do this but I would have to go to a diva and ask them for a report and you know he would be kind of grumpy and be like I got a million things to do so I can’t give this report and then I would go take the report eventually when he gave it to me I’ll take it to the Ceo and say here’s the report. And the Ceo would start asking me all kinds of questions like well what does this mean? What does this number like how do you define this? Doesn’t include this or does it exclude that?
Nick Amabile
00:40:43 – 00:41:00
And I was like I don’t know I didn’t write the thing and I didn’t write the report. All I did was ask the other guy for the report. And so I realized that you know again there’s this gap that I was in and I was kind of like in no man’s land because I wasn’t the person with the technical shops to actually create the reports and I wasn’t necessarily the ceo coming up with all these questions. You were literally taking the
Sky
00:41:00 – 00:41:11
specs from the customer and giving developers without, you weren’t like taking them and and making them legible to you were just handing out physically.
Nick Amabile
00:41:11 – 00:41:13
Yeah, I was like physically taking
Sky
00:41:13 – 00:41:15
the job kind of working. Yeah,
Nick Amabile
00:41:15 – 00:41:37
exactly. And so I was like, I was like, this is no good, right? Like I can’t, I can’t have a career doing this. And so um, I started teaching myself sequel and different programming languages. I was like, you know, this D. B. A. Has got a million other things to do for his job, right? His job is not to give me reports and do analytics. That’s my job. So you know, I really started teaching myself different programming languages and and just, you know, basically never looked back. So
Sky
00:41:37 – 00:41:56
it seems another good analogy maybe as a translator that only speaks one language and really just plays a game of telephone just says the same thing to the person so they don’t speak to each other. And it’s like, well you’re not really a translator, you don’t have much value then, do you? You need to actually translate the information so it can be received on the other side.
Nick Amabile
00:41:56 – 00:42:23
Yeah. And so you know, so that that’s exactly what I, what I started to do is teach myself a lot of the technical skills and you know, worked in financial planning and analysis for awhile, worked in the kind of operations, got my feet wet in a lot of different areas of business. Uh and then most recently I was the head of business intelligence at Jet dot com and prior to that was that etc. In new york and senior analytics roles and learned a ton of both of those jobs. And as I was kind of nearing the end of my time at Jet dot com,
Nick Amabile
00:42:23 – 00:42:49
um you know, I went to interview with a bunch of companies here in the new york city area and they all wanted me to set up their data warehouse and to set up their reporting so that it would all be standardized and I was like, well I’ve been doing this for a long time and I just did this at Jet, I did this and that he had this other companies prior to that. I was like, there must be something here, if everyone has this problem that they’re asking me to solve and I’ve already been solving it for a long time. You know, there must be something that I can do to, to really focus on that problem and get good at it. And um
Nick Amabile
00:42:49 – 00:42:57
that’s what I did. I hung up my own shingle and uh our first customer was some small start up here in new york are second customers amazon and you know, we were off to the races from there.
Sky
00:42:57 – 00:43:43
That’s a decent second customer I guess. So people get thrown off when talking to you or when you would show up for a job interview, let’s say early on one of these jobs and you look more like a foreman on a construction site than a data analytics type of guy. I think that’s a that’s a, that’s a generous interpretation of how I look, but um but it seems like, I mean you do some profiling people come in for an interview and you’re like this guy just doesn’t look like a data analyst. I’m gonna hire the guy that looks like Milton from the Simpsons Milton the character from the Simpsons, not the guy who looks like, you know, he’s the foreman somewhere working on a steel mill.
Nick Amabile
00:43:43 – 00:44:10
Yeah, no, it’s it’s it’s funny, I mean like, you know, I think when, when you start to uh you know, kind of get really deep into the field, you realize that the problems are so common across these different companies and that was really the insight that I had, I happened to be at the right place at the right time as as sort of, I started out um with a number of our technology partners um like Looker and then Snowflake and then google cloud and others that you know, they were really supportive of our of our business initially and gave us a lot of our initial customers
Nick Amabile
00:44:10 – 00:44:37
and the relationships that I had prior to to start in the consulting firm. When I was at for example, I was a customer of lookers, got to know a lot of folks there and then when I started out on my own, of course the you know, as many folks who have done this before, the sort of it’s a chicken and the egg problem. Right. You know, if you don’t have any customers, it’s really difficult to get customers and once you get a couple of customers it starts to snowball a little bit and we were able to kind of break that chicken and egg cycle through our through our technology partners, which we still work with very closely today.
Sky
00:44:37 – 00:45:10
Excellent. So um you have very small company is your first client. Very big companies. Your second client. It seems to me when you’re crunching data um a is a big company easier because they have more data, like you have all MLB data or you have the data from one intramural basketball game, who you’re going to give better results too. Um Is is it easier to work and solve the problems for a larger company?
Nick Amabile
00:45:10 – 00:45:35
No, I don’t I don’t I don’t know if it’s easier per se they’re very different challenges and I think I alluded to this before because bigger companies, they have, you know, again it’s it’s it’s not a technical challenge, it’s more of a sort of people process, organizational type challenge where you have such a diverse set of stakeholders and larger companies and um you know, you have to really work cross functionally and understand how to get things done, which can be difficult in large companies, whereas smaller companies, you know,
Nick Amabile
00:45:35 – 00:46:07
it’s great to work for small clients a lot of times because you know you come in, you’re like you’re like, hey uh you know, can you fix this? And we’re like yeah we can fix this and we go away, we fix it and knock, knock it out of the park right? Like and then bigger companies, the stakes are higher, the complexity is higher. Again, not from a technical perspective, so um you know, and and really that that was that was how we were able to make that transition from smaller companies to to larger companies was because the fact that it is so similar, right? Like I had worked in e commerce for a long time and coming to work at at amazon as a consultant was was not
Nick Amabile
00:46:07 – 00:46:22
way outside of my realm of of what I understood, it’s still, you know, kind of orders, shipments, inventory customers, it’s like all the basics, right? So um you know, we, I had a really good sort of background so then foundation to go in and execute successfully there. So that was that was good,
Sky
00:46:22 – 00:46:36
So you end up at at the last 42 you end up solving different problems in a way because the size of the company dictates what the problem is. Yeah it seems like it’s not just a matter of how much how much data they have for you to provide. Its
Nick Amabile
00:46:36 – 00:47:14
right. And I’d say that the thing the lesson that I’ve really learned over the years is how to be a better consultant right Because I I I’m a data and analytics practitioner. I’ve been doing that for a long time and that’s what I’m passionate about and I love helping customers solve problems with data and analytics but you know becoming the consultant where you can manage expectations and timelines and budgets and communication and stakeholders and all this stuff that that was a learning curve for us. Um you know we were able to kind of get it up you know get up that learning curve relatively quickly and now you know I’d say we’re quite good at it but you know again nobody ever said hey you know that’s 42 nick. You guys don’t know how to
Nick Amabile
00:47:14 – 00:47:20
program or you don’t know how to do data analytics. It’s all just like you just worked on the wrong thing for the last three months and
Sky
00:47:20 – 00:47:24
so like understanding
Nick Amabile
00:47:24 – 00:47:46
absolutely. And I think that’s that’s why at the end of the day we consider ourselves a business consulting firm not an I. T. Consulting firm because we’re trying to solve business problems. And so like literally the first question I asked when I come into a company and a new client is tell me how you make money. You know, it’s like that’s that’s just so simple because so many people skip that part and so many people assume that hey, if it’s amazon right,
Nick Amabile
00:47:46 – 00:48:06
you might ask, hey, oh I obviously know how they make money, right? I must know how they make money, but maybe you probably don’t, there’s a lot of nuances to it. There’s a lot of different ways that they think about their business that probably are new to you. Uh and you know, you can’t really assume anything. And so you come in with the blank slate and then you start with the business questions like tell me what’s important to you, how do you make money?
Sky
00:48:06 – 00:48:32
So have you ever Das, to have you guys ever been called in to to a job or to consider a job and looked at it and said, oh you don’t need us. Like in fact you don’t need to be analyzing data. You, you need to be just, I mean it seems like it would be a year too small for this to even be practical because someone tried to bring you in before and and and there wasn’t really something that you were the solution for.
Nick Amabile
00:48:32 – 00:49:03
Yeah, absolutely. I mean you can, we kind of talked about smaller companies and when we first got started out uh you know probably spoke to every small start up here in new york like, you know, two person companies, five person companies, 10 person companies, even 50 60 person companies they all know that they need data analytics at some point. And you know, especially given kind of the media like hype and stuff like that or just the buzzwords that are out there, folks are familiar with the concepts and they’re like, yeah we want we want to work with, you seem like smart, the person that really can really help help us do this.
Nick Amabile
00:49:03 – 00:49:33
But then as I said before it’s like well you know, you guys should probably focus on other things. You should be focusing on developing your product and going to market and finding product market fit and all that kind of startup stuff that you want to go do data analytics is important but it’s not the end all be all so as you get up the maturity curve then you might need, you know, our services um and certainly were not the right fit for everybody. I mean that’s the other piece too, is like we focus pretty specifically on a certain set of technologies and a certain methodology were quite opinionated around how we do these things. So you know,
Sky
00:49:33 – 00:49:38
I like that as a company description were quite opinionated by Yes,
Nick Amabile
00:49:38 – 00:50:11
I mean we’re always you know trying to develop and challenge our thinking on things but like at the same time, you know we want to make sure that if we come come into work with a customer, we’re gonna knock it out of the park for them? They’re gonna be super happy and successful, um you know, in our prices reflect that to be fair, but like at the same time, you know, we’re not the best fit for everybody. And so even in a couple of cases, there have been bigger companies that um have said, hey, we have this problem and I’m like, you don’t need us, you can actually just uh you know, you can handle this internally and I’ll give them some advice and some free consulting and you know, that’ll that’ll kind of open the gates for them.
Sky
00:50:11 – 00:50:19
But when you say you’re opinionated, is it like, do you have a little litmus test, you give people like GIF or Jif? Sorry,
Nick Amabile
00:50:19 – 00:50:22
Right? You know, it’s like we have to go into a company and they have to be using
Sky
00:50:22 – 00:50:25
Macs and Pcs and I’m just kidding.
Nick Amabile
00:50:25 – 00:50:50
Yeah. No, no, no. But you know, it, but it’s it’s, you know, I think it’s more around kind of the methodology around how we do things with, with kind of modern cloud based technology, you know, like I talked about data governance and that’s a big piece of it as well, so having a much more holistic approach. So a lot of times companies will call us up and they’ll say, hey, I need you to set up our data warehouse. Um and the first thing I’ll ask is why, right? Like what are you trying to do? Right? So if people want purely
Nick Amabile
00:50:50 – 00:51:05
technical I. T. Services right? Like that’s not us. We’re gonna come in and we’re gonna ask a lot more questions. We’re gonna want to understand the context and really you know shape the solution the way that we think it needs to be shaped rather than saying hey here’s the specs, go implement
Sky
00:51:05 – 00:51:06
the specs right?
Sky
00:51:07 – 00:51:10
It’s kind of like a steakhouse is not going to cater a vegan wedding.
Nick Amabile
00:51:10 – 00:51:13
Yeah that’s right. Exactly. Certain situations
Sky
00:51:13 – 00:51:14
where you’re like that’s not our
Nick Amabile
00:51:14 – 00:51:15
yeah
Sky
00:51:15 – 00:51:19
not what we do that’s the kind of problem we solved but not the problem we solved. Yeah.
Nick Amabile
00:51:19 – 00:51:22
Right. Exactly. Well cater a wedding just not a vegan one.
Sky
00:51:22 – 00:51:28
Okay excellent. So that’s 42. Yeah what does that mean? Where does it come from?
Nick Amabile
00:51:28 – 00:51:35
Yeah I get that question a lot. So D. A. S. Stands for the stands stands for data and analytics services which is the original corporate name of the company.
Sky
00:51:35 – 00:51:37
You had 41 failures first……
