Data Science Hangout | Matt Dancho, Business Science | Why some companies fail at data science

Hey, everybody, welcome back to the Data Science Hangout. So nice to see everybody today. If this is your first time joining us today, I know there was a lot of love for Matt on LinkedIn. So we might have some people here joining for the very first time. Thank you for joining us. This is an open space for the whole data science community to connect and chat about data science leadership, questions you're facing and getting to learn about what's going on in the world of data science across different industries and companies. So the sessions are recorded and they're shared to the Posit YouTube, as well as our Data Science Hangout site. So you can always go back and rewatch or find helpful resources.

But with all that, thanks again for joining. I'm so excited to be joined by my co-host today, Matt Dancho, founder of Business Science. And we were just talking, Matt and I spoke at our studio conference back in July about the Hangouts. We were like, let's get something on the books. And here we finally are. So thanks for joining, Matt. I'd love to have you introduce yourself and maybe share a little bit about the work you do.

Yeah, yeah. Well, first off, Rachel, so thanks so much for having me. You know, we were chatting and we had talked about doing this back in July, which was when the RStudio conference was going on. And like you said, I mean, things have just like flown by and here we are in November and this is exciting. I'm excited to be here. So thank you so much for having me. Just to introduce myself, my name's Matt. If you guys haven't ever met me before, you know, it's great to meet you.

A little bit about myself. My name is Matt Danchu. I've been doing R for about, I want to say like 2013 was like officially my moment when I like, yeah, I'm like, I'm going, you know, full into this thing, which back then it was just kind of like statistics or it wasn't even, I don't think quite called data science. I got, I got bit by the R bug bad. Like I became obsessed and I just started like learning as much as I possibly could and trying to apply it to my day job at the time, which I was working for a small company called Bonnie Forge. And I was just helping them, you know, handle their customer data, work with the manufacturing team, understand like back then there was a big drop in oil prices and the products that we made went into oil. So I was doing like a lot of forecasting and things of those things, think things of that nature.

So fast forward to today, in 2018, I kind of got into 2017, 2018, I got into consulting kind of as a side hustle at first, and then it slowly became my full-time job. And then in 2018, I pivoted into education because that's where I really saw a huge demand for, you know, as I was doing a lot of these consulting projects for companies in like the marketing domain or finance domain, you know, it was, it was clear to me that the problem wasn't that they needed a consultant, they really needed education and they needed help. So in 2018, I founded Business Science, that's my company. And since then, I've been building a lot of courses, that's what the first four years were. And now we're getting to the point where we're starting to add additional programs and scale up the programs, too. So now we have about, I think, four or 5,000 students.

Building a personal brand

Yeah, absolutely. So Rachel, building your personal brand and you, and you know this, cause you've done a fantastic job as well, especially with being part of our studio on Posit. I think you're doing a fantastic job at this, but like everyone can really benefit from building a personal brand. It doesn't matter if you're, you know, trying to get a job, if you're trying to build your own company, if you're, you know, trying to do even just, you know, kind of grow yourself personally and kind of step out of your comfort zone. I think it's really important for everyone to build their own brand.

Don't, don't feel like you need to get it perfect. There's a guy that I follow on Twitter, his name is Justin Welsh, and he does a really good job at like, for people who are trying to like start their own business, how to, how to do your own personal brand. And some of the things, like some of the value that I've gotten from him, talking about your journey, like no matter where you are, just, you know, every day, just going on LinkedIn or maybe every, you know, make it, make it a point like two, three times a week, just put a post out there of something new that you've learned or something that you're interested in. And it doesn't have to be like a whole production because this is the mistake that I always made was I always felt like my copy has to be like super on point. And it needs to be like something I've thought about for like days or, you know, and while that stuff does work, I think what is more valuable is like being authentic and just kind of showing like, Hey, you know, I'm not perfect. I learned something. I'm excited about it. You know, I learned this new art package, people identify with that. And then you start to grow your followers.

Keeping remote audiences engaged

So, all right. There's, when you, when you do a presentation, you know, people are on Zoom calls all day, you know, with COVID and everything, the last thing you want to do is you want to, you want to just kind of like get onto a Zoom call and just talk at people for 60 minutes. And that's the last thing that they want to do too, because, you know, they're going to be, you know, after the first three minutes, they're going to tune you out, right?

So what we started doing on our marketing webinars and on our technical webinars is really just trying to keep the engagement up. So like the, like what I did here with like the threes and the fives, that's a, that's a way to engage. So you just ask a question to the audience and oftentimes, you know, that, that keeps them kind of in there that much longer.

I have a note on my, on my keyboard, don't be boring. And so I will do like weird stuff, like I'll kind of, I've started to like play around with the way that I talk. So I used to kind of like talk just like, like I would conversationally, but in a, in a webinar, like you want to talk a little bit differently. You want to actually accentuate, you know, your speech patterns. And also starting everything with a big promise, like, you know, Hey, what, why are you here today? Like, what are you going to get out of it?

Breaking down data silos

So I'm a big fan of including as many stakeholders throughout the project as possible. So a typical example might be like if I was doing a consulting project with like one of the companies I worked with back in the day, like S&P Global, MRM McCann, they're a big marketing firm, S&P Global, they're on, they're in the finance. So what I'll do, like I normally have a point of contact there that I'm working with, but what I'll try and do is as we come up with the project, you know, discovery phase, I'll ask a lot of questions like, what are your pain points? What, which, you know, how does this affect your organization? Like really trying to like go a couple layers deep with them. And what ends up happening is they'll be like, oh yeah, well this, you know, this affects this part of the company. I was like, well, you know, should we involve them in the conversation? And they're like, yeah, it's probably not a bad idea. So you start like kind of building a group of people, and then once you get them all kind of, you know, filled in on what you guys are trying to do, that often just helps break down the silos and they can kind of understand and talk in an open, you know, forum where everyone's equal.

Why companies fail at data science

I really view that data scientists, and like data analysts, I think that there's going to be a new role here in the next couple of years that companies are going to start adopting the role of the business scientist, because you already have a business analyst. And I haven't really talked a whole lot about this yet. But like, the more I think about it, the problem, the problem and the challenge with data science, and why companies often like fail with data science, it's not because they have, they don't have good people, or they don't have, you know, problem solvers, or this or that.

But they lose sight of the business first mentality, where you need to be able to start with a problem in an organization, be able to work with many different levels of the organization, from executives, to line managers, to people who are doing the work. And you have to work cross functionally and cross sectionally throughout the company. And that, to me, requires like a heavy business acumen, like not just a like, hey, I know data. And I know a few algorithms that I can throw at the data. Like, no, that's, that's not how it goes. Like, the most successful people that I've seen are the ones that are able to communicate effectively, that know what the target is of the organization, the objectives, and they can translate that into like all the different, you know, they can kind of take that down and, and translate it across the organization.

But they lose sight of the business first mentality, where you need to be able to start with a problem in an organization, be able to work with many different levels of the organization, from executives, to line managers, to people who are doing the work.

Overcoming impostor syndrome

So, Juan, this is a great one. So, I would venture to say you're not alone. I would say probably 90% of data scientists experience some level of impostor syndrome. It's very common. I experienced it.

My problem was, so I'm a data scientist, like, self-trained. I never went to school for data science. I was a mechanical engineer. It took me probably at least four or five years to think of myself as an actual data scientist, and there were some days when I was going through it, and I remember, like, trying to make reports and just, like, nothing's coming together. I'm drawing blanks. I'm just staring at the screen, like, the cursor's blinking at me.

Don't give up. Like, every day is a new day, so try and write at least one line of code a day. That was something that I implemented in my own personal journey, and I did that for, like, I don't know, seven or eight years straight, and that single activity of, like, saying, like, hey, I got to pull up RStudio. I got to start, you know, I just got to put one line of code in there, and you know what's going to happen? It's a mental thing, because you're going to write more than one line of code. Like, who's going to stop at one line of code, right?

I had a problem. One of my biggest mistakes is I would get on LinkedIn, and honestly, I would see everyone talking about deep learning. This was back 2016, you know, they would be talking about TensorFlow, Keras. That, for me, was a huge black hole. I actually, true story, I built a regression model for my company that ended up, like, honestly, no joke, helping me take the part of the company that I was overseeing from about $3 million to $15 million in revenue. It was a, I implemented a lead scoring algorithm that literally was just a logistic regression.

I spent probably six months on that, and I got nothing out of it. My logistic regression model was better, just trying to, like, understand TensorFlow, and then realizing TensorFlow is too complicated, so then I'm going to switch to Keras, and then, oh, somebody's talking about PyTorch. You know, it's just the process of, like, seeing all these, like, I call them red herrings, like, they're, you know, these shiny objects out there that are trying to, like, steal your attention, and it's taking you away from, like, the thing that you're, that's actually getting you the value, so data wrangling, ggplot, you know, work on those, get the foundations, get the basics done, and get the quick wins, and then kind of work your way up as you go.

I call them red herrings, like, they're, you know, these shiny objects out there that are trying to, like, steal your attention, and it's taking you away from, like, the thing that you're, that's actually getting you the value.

What's next at Business Science

So I'm working on two big things right now. The first one is actually a program that combines. So for those of you who don't know how business science is structured, I have a core product, which is called the five-course R-Track. The time series course that Siavash is taking is the third course in that system, and it's basically designed to kind of take you from wherever you are now to, you know, being super productive, being able to out-compete all of your competition if you're trying to get a job, or if you already have a job and you're trying to apply data science, give you all the skills necessary.

The next big thing is to kind of take that system and add on an additional course to it, but also, so the course, the additional course is to help get you a data science job in 30 days, and then the whole package is designed to also add in a coach, a personalized coach that takes you through the entire program in under 180 days, along with the whole group that goes through the process with you.

She had no prior data science experience, like, working as a data scientist. She had no prior working experience other than she used to sell wine for her dad's business. Her dad owned a company that distributed wine. So she was able to, in seven months with this program, be able to get her first data science job. She just landed here about two weeks ago at the Federal Reserve in Minnesota, so Minneapolis, Minnesota. So she crushed it, and I had to hire her on because she had everything stacked against her. No working experience prior to that.

How learners learn best

I find learners are either 2x learners, 1.5x learners, or 1x learners. So, no matter what, once they get the hang of how to do 2x, it's like you never go back, right?

One of the biggest metrics that I started out with was understanding if my courses were better than the rest in the market. I think the Slack channel with the support, getting the support, as soon as they have a problem, they can hop in there, get in, you know, ask any question, no questions off the table, and really get feedback at the speed of learning. I think that's incredibly important. When I implemented that, that really helped kind of take my completion rates up.

Keeping videos small, I think that's super critical. So, there's this thing, I can't remember the name of the curve, but people learn kind of in a, it's like a U-shaped curve. So, if you do, like, an hour webinar, people only remember the first part of it and the end of it. Like, in the middle, they're only going to, their chances of remembering that, it's like three to five percent. But at the beginning and end, so if you can shrink that time down, and one of the things I did was, I tried to keep no video longer than five minutes. So, you're going to at least remember the beginning and end of each of those videos, and then it's also going to be like M&Ms, you're just going to pop one after the other, right?

A lot of other stuff that we do differently, starting off with a business problem first, I think that's super critical. Getting them a quick win, showing them like a jump start, like, not a full analysis, but a bite-sized analysis, so they can begin with the end in mind, they can kind of see and visualize how this course is going to play out, and how that value is going to stack, and then by the end of it, they're like, you know, beyond thrilled, because they saw just at the beginning, you know, what they're going to accomplish.

Yeah, it was so great being on here, guys. Honestly, Rachel, it's an honor for you to invite me on here. And whatever you guys are doing, you know, just know that you're on the right path here with our data science. It's really going to help you. It helped my career. I know it's going to help you guys, too.