We discussed harnessing data for impact and success in a business, how you can use data for your personal success, some tools and techniques Konrad and his team are using, and more! Also, if you find their work interesting, and would like to join the team, make sure to check their open positions (link below).
About Konrad Kopczynski
Konrad has created $160mm in impact through analytics implementations for executives at firms including TD Ameritrade, Bridgestone Firestone, Indigo AG, Fortune 500 Cpgs and high growth e-commerce startups like Public Goods. He is doing this by bridging the gap between the business and technical worlds. Konrad is highly skilled in understanding stakeholder needs and translating those into clear requirements for technical teams while ensuring the highest standards. He is able to solve those complex business needs through building and leading teams that consistently deliver automated, scalable solutions that perform beyond expectations.
Overview
Konrad runs a data analytics consulting firm focused on direct-to-consumer companies in need of new tools and new infrastructures. One of their clients is Public Goods, a startup out of New York that sells grocery items. Konrad’s company helped them built out their infrastructure and get them data they could trust and take action on. He has helped a similar company build a real-time PnL.
In terms of tech stack, they don’t use a traditional software development model. There’s a front end with a lot of Excel and Tableau and others. SQL and Python are used for the data modeling side of things, as well as Splink which runs on Apache Spark for matching needs, deploying matching in a cloud rather than locally to take advantage of the increased computing power. The modeling techniques they do are often fit for purpose. Where hiring is concerned, Impakt tends to either develop entry-level talent or work with high-level contractors. They are a remote company that relies on Slack and Asana and 7Geese to stay on task, communicate, and follow OKRs.
Konrad describes his path to having a bit of a “random walk.” He started in management consulting with a focus on data, without the implementation element. Over time, he became interested in doing the technical work himself, rather than theorizing solutions and contracting out the work. He freelanced his services and one client, in particular, asked for more of Konrad’s time exclusively and he began the process of building his own company around the work. Daily, Konrad employs a “daily evolution” of check-ins to hold himself accountable for work, which he repeats at the weekly and monthly level. He tracks his sleep, sleep quality, exercise stats, his use of the Pomodoro method throughout the day, all as a way to continually check in on his productivity and himself. As many listeners know, I track similar metrics on myself to improve my mood and work throughout the day.
From there we discussed our own methods of tracking our own professional and personal success. Konrad uses an Apple Watch, I’m partial to Whoop. We both track our sleep and see how our days are affected by our sleep levels and quality of sleep, we both employ the Pomodoro technique to track our work throughout the day and track our fitness. Like many other things in the professional world, improvement requires data to understand trends and make educated adjustments to your behavior. All this work has helped Konrad become an Iron Man, a competitor in an intense triathlon which involves a multi-mile swim, an over 100-mile bike ride, and a marathon over the course of 13 hours. Ultimately, this is the same process on the personal level as Konrad employs in his work to track a business’s progress.
We closed out with a discussion of Konrad’s multi-year project to read a biography on every single US president and his learnings from that project and his readings.
In this episode you will learn:
- What does Konrad do [3:40]
- Tools and techniques used in Impakt Advisors [10:35]
- Impakt’s unique hiring model [18:53]
- How does Impakt manage remote work [21:36]
- Konrad’s professional history and daily structure [28:42]
- Konrad’s Iron Man triathlon [44:11]
- Konrad’s years’ long project on presidential biographies [47:46]
Items mentioned in this podcast:
- Open positions at Impakt Advisors
- SuperDataScience
- Splink
- XGBoost
- Whoop
- 7Geese
- Harvest
- Scrum
- SDS 456: The Pomodoro Technique
- Ambition, Pragmatism, and Party: A Political Biography of Gerald R. Ford by Scott Kaufman
- The Years of Lyndon Johnson by Robert A. Caro
- Deep Learning Illustrated by Jon Krohn
- Machine Learning & Data Science Foundations Masterclass
Follow Konrad:
Follow Jon:
Episode Transcript
Podcast Transcript
Jon: 00:00
This is episode #465 with Konrad Kopczynski, managing partner of Impakt Advisors.
Jon: 00:12
Welcome to the SuperDataScience Podcast. My name is Jon Krohn, a chief data scientist and best-selling author on deep learning. Each week, we bring you inspiring people and ideas to help you build a successful career in data science. Thanks for being here today. And now, let’s make the complex simple.
Jon: 00:42
Welcome back to the SuperDataScience Podcast. I’m your host, Jon Krohn. And we are very fortunate to be joined today by Konrad Kopczynski. Konrad is an absolute master of using data and analytics in a feedback loop to iteratively achieve the goals of your dreams, be they commercial goals or personal ones.
Jon: 01:03
Konrad is the founder and managing partner of Impakt Advisors, a consultancy that specializes in harnessing data for impact. They structure the various data sources firms have available into thoughtfully constructed data warehouses, and then they layer on top analytics, data science models and visualizations to enable real-time reports, dashboards and predictions across all the key areas of a business, including digital marketing, customer retention, behavioral segmentation, and ultimately, profit margin.
Jon: 01:32
Not only do we cover tons in this episode on using data for commercial success, in the second half of the episode, we dive deep into how you can use data on yourself to iterate on yourself and evolve into whatever sort of person you’d like yourself to be. As a bonafide Iron Man and founder of a mobile app for habit tracking, Konrad sure knows what he’s talking about in that department too.
Jon: 01:57
We do briefly get into the technical details of specific software libraries, tools and statistical techniques for a few minutes here and there, but overall, this episode should be of great interest to anyone who’s keen to optimize themself or their business using data, regardless of whether you have a technical data science background or not. All right, let’s get into it.
Jon: 02:25
Konrad, welcome to the show.
Konrad: 02:28
Thank you. Good to be here.
Jon: 02:30
Where are you calling in from today?
Konrad: 02:34
I am near Kingston, New York, so just an hour and a half north of the city.
Jon: 02:40
Nice, yeah. You recently moved out of Manhattan. So, if I remember correctly, you were born and raised in Manhattan on the upper east side?
Konrad: 02:49
Yup.
Jon: 02:51
But under COVID, you took the opportunity to get some space, some greenery, move up north.
Konrad: 02:58
Yeah.
Jon: 02:58
How’s that working?
Konrad: 03:02
It’s great. For now, we’re just here full time, but the plan is to be back and forth once everything starts returning. Well, it’s already starting to return to normal, but once it really starts to return in full swing.
Jon: 03:13
Nice. It must be beautiful. I look forward to checking it out. We have actually known each other for quite some time. I don’t even know exactly. Probably coming on 10 years that I’ve known you.
Konrad: 03:26
Close to that. Yeah, definitely.
Jon: 03:29
Met in New York through friends, and have always been fascinated by the work you do. I am delighted to be able to have you on the show today, so that you can share the work that you do with our audience. So, tell us a bit about that. What is it that you do, Konrad?
Konrad: 03:43
Sure. So, I run a data analytics consulting firm. We primarily focus on CPG and direct-to-consumer companies that are growing past their existing vendors and tools. A lot of times, they’re exporting things to Excel, and we come in and help them get past that, get a real data warehouse set up, get real reporting, and then evolve from there to do more complicated and interesting things like predictive analytics, and different machine learning analyses, things like that.
Jon: 04:21
Nice. And so “CPG” was an acronym that you used there, an abbreviation. Consumer-packaged goods. I only know that, because I talked to you about it right before we started recording. But so, that is a particular niche. But I guess anything related to that is your client base, so any kind of company that’s selling some kind of product to consumers, you work with them as they begin to scale. Although, you’ve also worked with some pretty big clients.
Konrad: 04:49
We have. So, another segment and niche that we’ve serviced and kind of do work with is, innovation teams within large companies, or teams that are getting spun off of product launches, things like that. So, we’ve worked with some Fortune 500 companies, generally when they have a new product they’re launching, a new team that they’re launching, something along those lines, and they need to jumpstart and kind of kickstart out of the gate. So, it’s a similar type of work where we’re starting from scratch and helping them build something out that is going to run on a larger scale.
Jon: 05:25
Super cool. So, tell us a bit about that. Do you have a couple of case studies?
Konrad: 05:31
Sure. So, for example, one of our clients I can actually name is a company called Public Goods. They’re a e-commerce, CPG startup out of New York. They sell anything from shampoo to pasta, and we came in and helped them. They were doing everything manually, and exporting from Excel, and we helped them build that out, and do some reporting on top of that, and really get a better understanding of their data, so really get numbers that they could trust for revenue, for customer lifetime value, for ROI and marketing analytics, that sort of thing.
Konrad: 06:16
And then, kind of help them hand that over and take that on, and build their own data team out further. Another example is, it’s a similar type of company. Products they’re selling to different consumers, and we helped them build a real-time P&L. So they were trying to really optimize what marketing campaigns they’re doing, and they run probably 10, 15, 20 new campaigns a week on a relatively large scale. And they want to be able to know, within a day or two, “Is this the right campaign for us? Is this the right audience? Are all the parameters set up correctly?” And then, once you get past the typical things, which are just kind of like, conversions that Facebook gives you. One, there’s a problem that the conversions that Facebook tracks aren’t always accurate.
Konrad: 07:10
There’s cross-attribution from Facebook and Google. And then, second, for this company in specific, it was really important to understand how high quality the customers they were getting from a given campaign are. And they could tell that, based on what packages people were purchasing. And so, really understanding the P&L behind kind of like, the shopping cart mix, was really important to them so that they could see on a profitability level, “Is this a profitable campaign, because the people are buying the right products and coming back to buy the right things?”
Konrad: 07:45
And so, to do that, we kind of had to do a predictive, I guess, P&L for them, so that they could know right away both, “What was the actual profit on that initial transaction?” And then also, what to expect from that type of a customer going forward.
Jon: 07:59
Nice. So, if I try to summarize at a high level, I think both of those use cases may be kind of a generalization of what your company does in general, and it’s called Impakt Labs, which I failed to mention so far, although I probably did-
Konrad: 08:14
Well, Impakt Advisors.
Jon: 08:16
Impakt Advisors. Right. And “Impakt” with a “K?”
Konrad: 08:20
Yes.
Jon: 08:21
So, if I were to summarize what Impakt Advisors does, is that you take a company that … or a company or a team, that has a relatively small size. They’re probably working on local Excel spreadsheets, or maybe they have Google Drive and they’re sharing a few key spreadsheets, but data are in all different kinds of places, all different kinds of formats. There’s not really the consistency, and so you come in. You organize the data, bring it out of Excel, put it into data warehouses, and I guess allow SQL queries, dashboarding, real time analytics, now to allow of a sudden happen, so that instead of people … maybe not working in the dark, because they probably have some idea. They can probably look things up, but it’s very effortful.
Jon: 09:15
You allow them to much more easily get a sense of what’s happening in their business, where their valuable customers are, who their valuable customers are, how to better target them with marketing, and what kinds of actions they can take are more likely to be profitable or less profitable. Something like that?
Konrad: 09:33
Yeah. Yeah, definitely. I think, to build on the last piece of that a little bit, it’s two areas. One is just making things that they already do available in real time, or instantly, whereas in the past, maybe a report would take a few days to get, or therefore they can only do it once a month or once a quarter and now they can do it every day, every hour. And then, once that is done, then okay. Well, now that we have these basics, what can we do that’s more complex? And questions that we can answer or ask that we hadn’t even thought about before? Or had thought about, but didn’t have time to address.
Jon: 10:12
Nice.
Konrad: 10:12
Like for example, behavioral segmentation of audience, and trying to really understand, what are the segments of our customers? Are our customers the type of customers that we thought we were getting? And if they are the customers that we thought we were getting, are they buying what we were expecting them to buy?
Jon: 10:32
Nice. That is very cool. So, let’s talk a bit about the kinds of tools that you use on a day-to-day basis, or that your team uses. So, I know that you’re the managing partner of Impakt Advisors, and so not necessarily always in the weeds, technically. But I know that you do have some experience with that too, so you and I have talked in the past about you using Python and using R. It sounds like people that you hire would also need to have some awareness of Excel if they’re going to be porting things over into things like SQL. So, tell us about the kind of tech stack that you and your team are using.
Konrad: 11:12
Sure. So, we think about it kind of across front end, back end, although it’s not the traditional software development model, but more like the way that businesses use data. On the front end side, it’s a lot of Excel, especially in the prototyping stages. And then, once we start getting more automated, we get into using tools like Power BI and Tableau, for really showcasing what is coming out. And then, kind of the data modeling sides, it’s a lot of SQL. And then, when SQL gets too complicated, going into Python. And then we use Python for its predictive and library capabilities. And within that, there are different packages that we use. So sklearn, obviously we use a lot of. And Splink is one we’ve actually started using more lately, for-
Jon: 12:06
Splink? I’ve never heard of Splink.
Konrad: 12:07
Yeah, for … It runs on Apache Spark, so generally, it’s something we’re deploying into the cloud, but we use it for matching purposes, so that’s another big thing that we see clients have issues with, is matching data, whether it’s matching company names from two different sources, so that helps with that. Or another example is matching leads to data that’s been purchased about those leads, or business data. And so, it can really work at large scale, like one of these projects we’ve been working on is like, we’re matching across, I think, like 30 million records on either side. And it lets us scale that up and really get that good data, which then is given to the Salesforce to give them insights as they go in to have conversations.
Jon: 13:03
That’s super cool.
Konrad: 13:05
Yeah, so those are some the tools. I mean, just yeah. I would say.
Jon: 13:11
Also, you Apache Spark there quickly. You want to tell us a bit about when you would use that?
Konrad: 13:18
So, we use it specifically in this use case for matching. So basically, instead of doing the matching locally, we are deploying it on AWS service which lets us use much higher computing power. When we’re working through like 30 million records, it wouldn’t really work so well on our computers.
Jon: 13:44
Yeah, because it’s like, if you had two tables of 30 million, doing the comparison of every name in one 30 million table to another 30 million table, it would take a long time, yeah, on a local machine.
Konrad: 13:55
Yeah. It takes … Yeah. Luckily, there’s ways to segment it down, right? Like across states and things like that, but it still ends up being a lot.
Jon: 14:03
Nice. And so, what kinds of techniques do you use? Do you end up using any particular kinds of models very often, like statistical techniques? So, when you’re using R or sklearn, and you’re doing things like behavioral segmentation, are there particular kinds of models that you rely on?
Konrad: 14:24
It sort of depends what we’re doing. A lot of times, it’ll be kind of fit for purpose. Honestly, a lot of our work doesn’t rely on complex modeling, but rather relies on more of the traditional data modeling, and putting together, joining datas across different data sources, so getting a really robust customer record from different sources. And we don’t necessarily need any complicated library to do that. We just need to have an understanding of what the different sources are and what we’re looking for and how to do it. But when we do get to more predictive or machine learning applications, one project recently, we’ve been using XGBoost for predictive modeling, right? Or logistic regressions as well, is something we … one of the places that we see ourselves coming back to often.
Jon: 15:21
Yeah, I’m not surprised. Logistic regression is a workhorse. Not a huge number of predictors in the model. Maybe a dozen or 20. And you’ve got one outcome, a binary outcome that you’re predicting. “Is this prospective customer going to buy or not?” I imagine is something that you use logistic for a lot. And XGBoost, let’s expand on that for a moment for our listeners. So, XGBoost uses a technique called “decision trees” at its core. So, a decision tree is … It’s pretty hard to describe without a diagram, but basically, if you have a bunch of predictors that you’re using to predict some outcome, a good way of doing it, one that I use in some of my teaching, is a data set from the Titanic on passengers.
Jon: 16:14
And so, who survived and who didn’t survive? And so, you can use a decision tree to predict whether an individual would survive or not. And so, with the tree, the first branch of the tree would typically be the branch that has the most impact. So for the Titanic, it was something like, “Were you a male or a female?” So females had a much higher survival rate, so that can be a first branch of your tree.
Jon: 16:39
And then, down the female leg, you could then split on age or number of family members. And down the male leg, the same thing. And in the end, you’ll discover that class is very important in the model. Anyway, so our random forest takes a whole bunch of these decision trees that you create on sub-samples of your data, and XGBoost is a technique that builds on top of random forests where … Shoot, how can you describe exactly … Basically, you’re figuring out where your model makes its errors, and refitting where those errors happen, to have the errors happen less in the future.
Konrad: 17:23
Right. Yup. Yeah. Yeah, and it’s super useful for us also, because a lot of times, people want to understand what a model is predicting and why, and it gives you at least some insight into that. It’s not like a complete black box. So for example, there was a project we were doing for an insurance company, and they’re very regulated, so they can only make pricing decisions in certain ways, and they have to be able to explain their pricing decisions to the regulator. So, that was a good fit for them in that case, because they could really start to understand, what were the underlying factors that were going in?
Jon: 18:03
Nice. That’s a really good use case. This episode is brought to you by SuperDataScience, our online membership platform learning data science at any level. Yes, the platform is called SuperDataScience. It’s the namesake of this very podcast. In the platform, you’ll discover all of our 50+ courses, which together provide over 300 hours of content, with new courses being added on average once per month. All of that and more, you get as part of your membership at SuperDataScience, so don’t hold off. Sign up today at www.www.superdatascience.com. Secure your membership and take your data science skills to the next level.
Jon: 18:46
So, thank you for filling us in a bit on the tools that you use regularly, the kinds of techniques that you apply. When you’re looking to hire someone, what skills do you look for in them? What kinds of people do you hire?
Konrad: 19:01
Yeah, so we have a little bit of a different approach than most companies do, in that we basically hire in two categories. We hire either entry level, or we work with highly experienced contractors. So, that works for us, because we have the ability to tap really knowledgeable resources while developing talent in our way of client service and product management, or product development, or service development. And then, on the entry level side, what we typically look for is, we’re looking for people that already have a deep understanding of Excel, some understanding of SQL, and then are able to write a memo, is the way that we test, right? So, we have a three-part test, which is an Excel challenge, a SQL challenge, and a business memo challenge.
Konrad: 20:07
And we find that if somebody can do really well on all three, they’ll do really well in our business. But you’ll have people that are really good at Excel but can’t write, and that just doesn’t work when you’re working with clients, communicating with people and all that. So, that’s kind of the way that we think about it. And that requires a little bit of extra work, because we do a lot of training. But we find that that kind of pays off for us, because we have people that are really dialed in into the way that we’re working.
Jon: 20:38
Nice. That was a really clear answer. And that sounds like a really interesting business model to me, so training people in-house that can become experts in dealing directly with your clients, but also then interfacing with these highly experienced contractors. And if memory serves, you’ve been working with highly experienced remote contractors for a long time. Pre-pandemic, you’d been for years, working in that kind of way. And so that allows you to find the most talented people for a given problem anywhere in the world. So, I remember years ago, you were working with contractors in Africa, for example?
Konrad: 21:20
Yeah. So, our contractors are spread all over the world. We have contractors in the US, and in South America, in Europe, in Africa, in India, Pakistan. All over.
Jon: 21:33
Wow. That’s cool. And so, how do you manage that? A lot of us now have been remote for the last year, but how have you been managing this for years? How have you been making it work, and having a successful business, working remote like this for so long?
Konrad: 21:56
A lot of it is about structure. So, we use a bunch of different tools, but none of the tools themselves are going to be that different than what everybody else uses, so we have Slack, we have Asana. We use GitLab. But then, on top of that, we also … So, we really believe in … We use a tool called 7Geese, which is a performance management tool, is the way that we use it. And we’re really big on making sure that we’re using it, so we do quarterly performance reviews. And within that, we also do weekly one-on-ones. So, I do a weekly one-on-one with each of my direct reports, and other managers do the same.
Konrad: 22:45
And I think that sort of structure lets us make sure that we are getting feedback and having those conversations frequently, on the human capital development side. On the work itself, we have our own version of, probably bastardized to some extent, Scrum, combined with other product management methodologies to kind of pull everything together.
Konrad: 23:13
And so, that all kind of lives in Asana, and we make sure that everything is being managed around that. And we have a framework that then … so each task gets tracked against the tool, Harvest, that we use for time tracking. And we use that for forecasting as well, so we can see where people are going to be overbooked or under, and how can be allocating different resources and all that? And the biggest piece, I think, that we use from Scrum is the retrospectives. So, we do retrospectives on a biweekly basis, which really helps us. Because typically, what we’ll do is, we’ll have a retrospective, what went well, what didn’t go well.
Konrad: 23:53
And we’ll celebrate what went well, and then a lot of times, like at least 50% of the “didn’t go well” turn into after-action reviews, which help us really … Then we go and do that, and take that separately and go and spend half an hour or an hour with the team that was involved, and really think about, “How can we do this better next time? What changes can we make?” And I think the unifying thread throughout that is really documentation of … So, doing things, partially in a written way, but documenting what we’re doing, documenting what we’re planning on doing, and then reflecting and doing a feedback loop on that.
Konrad: 24:31
And that’s what really helped us, I think, most over the years, because I don’t think even now, that we do it in a perfect way. But the fact that we’re constantly iterating on it, constantly getting feedback, constantly changing it, means that even from a week-to-week basis, we’re making significant changes that lead to better results for our clients and for us, and for everybody involved.
Jon: 24:52
Nice, that’s cool. So, I imagine a lot of listeners know what a Scrum process is, particularly anyone that does software engineering or works on a software engineering team. But probably, quite a few of our listeners also don’t know Scrum very well, because you might not be doing software development. And not every software developer or software development firm uses Scrum, but I’ll try to summarize it in a couple of bullet points, is that you break down periods of work into cycles. So a common cycle, it sounds like the one you’re talking about, is a two-week cycle. That might be the most common kind of Scrum cycle.
Jon: 25:26
And at the beginning of the Scrum, you’d have an idea of what should be accomplished by the end of that two-week period, and then you typically have daily stand-ups or near-daily stand-ups to see where people are getting blocked on their particular tasks, to reach that two-week goal. And then, what you’ve been talking about specifically, the piece that seems to work really well for you, is those retrospectives on that two-week cycle. So, you get to the end of those two weeks. You say, “Okay. Two weeks ago, we planned on getting all this stuff done. Where are we on that? And if we didn’t get there, what’s wrong?” And you’re saying documentation is … Yeah, I can imagine is extremely helpful for having that work well.
Konrad: 26:07
Yeah. In a lightweight way, though. Because you can definitely get into this pattern of just doing too much documentation, too much process. All that, right? And so, it’s this constant balance, I feel, of going one direction where we don’t have enough, and then okay, we need a process for this. We need to think about this. We need to think about this. And then three months later, we’re like, “Okay, this is too much and we have to rip some stuff apart, and how do we do it differently?”
Konrad: 26:31
And so actually, Scrum, to some extent, pieces of Scrum were a little too much for what we’re doing, partially because we’re working across so many different projects. We might have one person working across a couple different projects. And so the biggest pieces that really work for us are like, the daily stand-ups, the retrospectives, and then those two really are … we find drive the most, and there’s other pieces of Scrum, like task pointing and more formal sprint planning, that we don’t do as much of. We have tried in the past, but it’s tough in this sort of an environment, where we’re trying to adapt and move quickly from different priorities.
Jon: 27:18
Right, right. Right, right, right. So, the task pointing would be assigning a certain number of story points, I guess you typically call them? So, how complex is a task? And so you can rate the complexity and say, “Okay. Well, this team with these people has a certain capacity to do these many points in the Scrum.” And yeah, I can see how, especially if you’re working with quite a few different clients, and you’re balancing that, that sometimes you just got to meet the deadline. It’s not like a multi-year software project, where … Yeah, you can kind of be like, “All right. We’ll just … I guess we didn’t get it in that two-week period, so let’s try again the next two-week period a little better.” You’re like, “No. That deadline’s come and gone. The client expected that last week, so …” Got it.
Konrad: 28:08
Yeah. Definitely.
Jon: 28:11
Cool. All right, so that is very helpful, hearing about the processes that are helpful, as well as the particular tools that you’re using. So, we’ve learned a ton about Impakt Advisors, and the impact that a consultancy firm like yours can have on a company, by bringing data together, by making it easier to run queries or have automated reports against those data, and therefore become a more efficient, more profitable company. How did you find yourself in this position? You’re pretty young, and you’re managing a highly international, distributed consultancy firm with tons of big name clients. How did you end up here?
Konrad: 29:02
Yeah, so like I think most people’s paths, it was a little bit of a random walk, I guess. I started in the management consulting space, but with a focus on data. So, my first job out of college was for a company that did management consulting, often on data, but we wouldn’t really do the implementation, where we would basically contract out the implementation. And that got me interested in the space. But over time, I got more and more interested in the actual implementations, and doing the technical work. And from there, I went and I worked at that company for a while, and then I went and started freelancing on my own, while working on a variety of different projects.
Konrad: 29:57
And while freelancing, basically, the origin story is that there was a client that wanted more of my time, and I threw out a big number, for me at the time, of like, “Well, this is what it’s going to really take to have me more focused on your work on a day-to-day basis.” And they went for it, and then that was kind of the, “Okay. Well, now I can think about this as something to build, and really build on that.” And that’s kind of what the journey was to where I am now.
Jon: 30:30
Nice. Are there particular … and I have a feeling that there are. So, from conversations that I’ve had with you in the past, I know that you spend a time thinking yourself about performance and productivity. We haven’t talked about that yet in this episode. I think we should probably open that up. Tell us about how you structure your day, your life. I know that there’s a lot of different cycles, so in the same kind of way that a Scrum cycle could be two weeks, I know that you have daily kind of retrospectives, as well as you used to have them sort of monthly or quarterly, about what you’re doing with your life, and how you can be more on task with reaching your end goals.
Konrad: 31:17
Yeah, definitely. So, it’s very similar to the way that we run the business. I think the core cycle for myself, personally, is a daily one as you alluded to. I have a daily process that I go through, that gets me ready for the day, that has … It’s basically a checklist of things, that includes clearing out email, checking the calendar for the day, setting up tasks, organizing tasks, and then putting those on the calendar so that I can actually tell that I’m actually going to be able to complete the things I am going to be able to complete.
Konrad: 31:53
And that’s kind of the core of it, is this daily … I call it kind of like a daily evolution for myself. And that happens on a daily basis. And then from there, there’s a weekly cycle. So I have a weekly checklist. There’s a monthly checklist. There’s a quarterly checklist, and then it goes to a yearly checklist that I do every year around the New Year. And that all kind of flows together with, at each level, there’s different steps that cause different levels of reflection and thought process, that help me keep moving forward. And as part of that, I’m also tracking a bunch of different metrics on a daily basis, that help me see whether I’m in a good place or not, and that’s varied widely over the years.
Konrad: 32:42
I think now, I’ve been tracking things on a daily basis, probably for like six or seven years, and I actually took a year or two break a few years ago. And it’s gone from, at a low point of tracking, none or maybe just five or 10 things, to at a high point, tracking 90 or 100 things on a daily basis.
Jon: 33:00
That’s what I remember most recently. The last time I asked you how many things you were tracking, I think you were at about 90.
Konrad: 33:07
Yeah.
Jon: 33:07
Do you have any specific examples of things from those 90 that you’ve realized since were … You were tracking too much?
Konrad: 33:17
I can’t remember really, what are the things that I was tracking too much of. I do know that the things I keep coming back to tracking are sleep, sleep quality, exercise. Recently, I’ve started doing like, Pomodoro method for task management, and I’ve been tracking the number of those on daily basis, and I find that’s really good. And I also track focus and things like that, so yeah. It kind of depends on what I’m … I think I got to 90 because I was trying to change all these different things, and every time I would change something, it would take a few months to do it, and then I’d get in the habit of it, and then I’d keep it on the list, and then I’d go to the next thing.
Konrad: 33:57
And so, then it slowly accumulated up to 90. But it did get to be a lot, and that’s when I decided to stop for a year, because I was like, “Okay. This is too much.” And I stopped for a year, but then a year later, I did my review and I realized that I actually did need to track some, but not everything. So, yeah.
Jon: 34:14
Yeah. So, I guess it’s the classic thing of getting that balance right, just like with the Scrums, of what’s the right amount of keeping track of information? And so, it was years ago … Man, I don’t know. Five, six years ago? That you showed me your process. At that time, it was a spreadsheet for keeping track of habits on a daily basis. And to me, it was so intuitively the thing that I needed in my life, to be more productive. I latched on to it immediately. I have been doing it every single day since you first showed that to me, whatever, six years ago now. So now, I have a series. I have a spreadsheet for each year, and the things that you’re outlining are the most important, I think, as well.
Jon: 35:03
So, leading indicators of my productivity and my happiness. Sleep, for sure. Getting hours of sleep is huge. If you’re a listener and you don’t track your sleep, or you’re not conscious about how much sleep you’re getting, it is crazy. So, that’s an example of a strong leading metric that I track. But I also track things like, “How focused was I today? How happy was I today?” And when I get a full night’s sleep, when I get eight, nine hours in bed, I enjoy my whole day, and I have patience with myself and with other people.
Jon: 35:44
But as soon as that gets just a little bit lower, six hours, seven hours, I’m a markedly different person. I don’t love my work. And if you don’t love your work, you’re not going to be as productive. You don’t enjoy it, and you’re easily distracted. So, yeah. That’s huge. Exercise is something that I track in there. Anyway, I greatly appreciate you showing me that. And something that you haven’t talked about is, at one point, that even developed into an app that was in the iOS store, in the Apple store?
Konrad: 36:16
Yeah, it was. I think it’s technically still there, but I’m not sure that it’s functional, but it was an app called “Build Habits.” Honestly, it didn’t encompass everything that we wanted to do with it, in the way that we wanted it to work, and so I personally have also gone back to using a spreadsheet. But that’s probably something that I will return to at a later date.
Jon: 36:42
Have you seen … So, I use a fitness tracker, or a heart rate monitor called “Whoop,” which I know you’ve also used. So, for people who aren’t aware, it’s a relatively new fitness tracker. It’s something that, their marketing is around being always on. So, you don’t even need to take it off to charge it. So, the battery pack charges separately. You slide it on. If people are watching the YouTube version, I’m showing it on my arm right now.
Jon: 37:12
And so, I just have this thing on my arm all the time. I take it off to shower, but I don’t have to. I could wear it in the shower. I just don’t like having a wet band on my arm for an hour after I get out of the shower. And so, this thing, it tracks your heart rate. It tracks how much you’re moving around, and so you can look at relationships between sleep and activity, and it gives recommendations.
Jon: 37:35
It says, “Based on how much activity you had today, you had a really big workout, so you should get an extra hour of sleep tonight. If you want to get that extra hour of sleep tonight, you’re going to need to spend an extra 90 minutes in bed tonight.” And it recommends when you go to sleep and when you wake up. So anyway, Whoop. I think it’s really cool. I know you’ve used it in the past, and you’ve actually … You’ve found that the Apple Watch was better for your use case, if I remember correctly?
Konrad: 38:01
I just found that it was not a large enough difference, in terms of the data that I was getting, that it was worth keeping and charging and maintaining two different devices.
Jon: 38:17
Yeah, yeah, yeah. And having something on both wrists.
Konrad: 38:18
There are apps on the Apple Watch like … Yeah. Right, yeah. So, there are some apps on the Apple Watch that give you like a readiness score, which is similar to the Whoop does. And I found that that is … That’s one of the things I track on a daily basis is my readiness score. And I found that it’s not as 100% good as the Whoop, but given that I am using the Apple Watch for other things anyway, I sort of consolidated it in that one thing, although I have been tempted to redo the two wrist thing lately, so we’ll see. I don’t know. That might happen at some point.
Jon: 38:57
We’ll see. So, where I was going with that, so I ended up needing to describe Whoop and then I kind of wanted to explain it to listeners anyway. But I don’t know if you know this or not. You probably don’t, because it’s something that’s happened relatively recent. I think it’s happened since you’ve had Whoop. I don’t know why you would know about it. But it has the “swipe left, swipe right” on daily habit tracking that you implemented in your app. It’s identical.
Konrad: 39:22
Oh, that’s cool. Okay.
Jon: 39:22
So, there’s this huge list of daily habits that you can choose to have or not. So, because I’m doing mine in a spreadsheet separately, I don’t keep very many on the Whoop. But when you wake up in the morning, and it gives you your readiness score, so it says, “Okay, your resting heart rate overnight, your heart rate variability overnight, the quality of sleep that you had, you’re ready for a really big run today. And also, please log your habits.”
Jon: 39:49
And in fact, actually, it prompts you to do the habits before it gives you the readiness for your day, so you kind of … Me, I’m working out in the morning, so I’m kind of waiting to see how ready I am. What kind of workout am I going to have today? And while I’m waiting for that, I fill in the habits, and it’s exactly … So it’s like, “Swipe left for no. Swipe right for yes,” on whichever habits you selected. So, I think the only ones that I track in there are caffeine, so how much caffeine did I have the day before? So you can also, you can optionally put in how many cups of caffeine you had. And that’s the really important one, that potentially varies for me and has a big impact on my sleep quality.
Konrad: 40:30
Right, right, right.
Jon: 40:31
But they also then do predictive analytics. Or actually, it’s not really predictive. I guess you could use it for predictive, but it then tells you, so you get a monthly report, and you can go back and look. I do drink coffee every day, so it’s not a really good example. But let’s say on half of days, I drink coffee-
Konrad: 40:51
The rest of the days, you didn’t. Yeah.
Jon: 40:51
Yeah, and then the other half, I didn’t. It would then show me the impact of that on my sleep quality, on my resting heart rate, on my readiness. And anyway, it’s kind of a cool thing.
Konrad: 41:03
That’s cool. That’s definitely making me want to reconsider trying it out again, at least to check that out. So I might do that.
Jon: 41:09
Yeah. Yeah, it bakes in analytics that you might be doing manually. I also want to highlight, so something else that you talked about there was the Pomodoro technique, which is a technique for doing 25 minutes of work, and then five-minute breaks. And by doing those kinds of sets, you can be productive for long periods of time, because you can pretty much, in almost any job, you can switch your phone off, your email off for 25 minutes at a time.
Jon: 41:34
And so, it’s a great way to stay on-task with some deeper work, and if you’re interested in hearing about that and you missed the episode, I did a FiveMinuteFriday, episode 456, on the Pomodoro technique specifically so you can check that out. And the thing that inspired me doing that episode was a phone call with you shortly before that. I was like, “Ah, I should do an episode on that,” after talking to you about Pomodoro technique.
Konrad: 42:00
Nice. Yeah. Yeah, I’ve found that to be really helpful, because I, for a long time, tracked a focus score, like just “one, two, three.” Three, highly focused. Two, mediocre. One, just completely off. And I find that the number of Pomodoros I completed the day before is a much better indicator, and it also … I know we’ve talked about that you use it also sometimes for meetings, but I don’t use it for meetings, and so it also gives me an idea of when I actually have time for focused work.
Konrad: 42:37
And it’s really interesting, because it’s basically only one or two days a week that I get a real stretch of focus time, and the rest of the time is spent with interactions and all that. And so, trying to optimize that is interesting.
Jon: 42:52
I completely understand. I flip-flop all the time on whether I should be putting meetings in or not. I’m like, “Well, I was really focused in that meeting, continuously for those two hours. I wasn’t checking email or doing anything else. I was completely in that meeting.” So I was like, “Is that a Pomodoro?”
Jon: 43:08
But it probably shouldn’t really be. I do different … I fill in a red circle for the 25-minute block if I was actually running the Pomodoro clock and focused on my own, and I just circle a circle in red if I’m marking down meetings.
Jon: 43:26
And I should break it out in my daily habit tracking. I think I should break it out into two separate rows, but I haven’t done that yet. Maybe this conversation will spur me to actually do that, because I think there’s different … There’s value in knowing, generally speaking, how productive I was over the whole day across all kinds of tasks. But there is, you’re absolutely right, there’s something special about that uninterrupted time alone where you can get that deep work done.
Konrad: 43:53
Yeah, yeah.
Jon: 43:54
Cool. And it seems to be working for you. You’re managing this international company. Something we haven’t even had a chance to talk about on the episode is that you are a bonafide Iron Man, so you’ve done a full-length Iron Man. And one of the things that blows my mind about that is that it was your first triathlon. So, break down for listeners, what’s involved in a full-length Iron Man triathlon? Tell us about the stages and how long they are.
Konrad: 44:21
So, I might mess up the distances a little bit, but it’s something like 2.6 miles swim at the beginning, and then it’s 112, I think, mile bike ride. And then you do a marathon at the end, so it’s 24-and-a-half miles run.
Jon: 44:38
Yeah. 26.2 miles, I think, is a marathon. Yeah.
Konrad: 44:41
26. Yeah, yeah. And I guess that was a few years ago now, and I think I’m actually kind of getting to the point where I’m like, “Oh, maybe I’ll do it again.” But it’s taken a few years since I did the first one. Definitely not one of those people … You have people that get into the Iron Mans and they do like, three a year. That has not been my experience. But I think I’m getting close to doing another one.
Jon: 45:02
Nice. And so, all of these techniques, I think tracking is key to that success, because you were able to, through planning, through these kinds of cycles that you have, the weekly, monthly, yearly cycles and review on your process, you were able to train for that crazy distance.
Jon: 45:21
And so, if you can’t imagine in your mind … Well, actually, I mean, even think about how long it takes to drive 112 miles in a car. So, if you’re on a highway going the typical speed limit in most countries, it takes you two hours to go 112 miles in a car.
Jon: 45:40
And so, that’s the bike leg in between a really long, typically open water swim, and then a literal marathon, a 26-mile run, which is the longest distance that pretty much anybody ever trains to run. So, what did it take you? 12 hours to do that?
Konrad: 45:59
I think my time was around 13 hours, if I’m remembering correctly.
Jon: 46:07
It’s amazing that it can be done in a day. And anyway, you were able to do that without ever having done a shorter triathlon distance before. And I think, so that’s a testament to this kind of planning, keeping data on yourself, reviewing the data, and iterating and improving from there. So, a philosophy that transcends through your life, as well as your business philosophy, and now something that you’ve, for years, been bringing to other businesses, and allowing them to capitalize on data and tracking more and more real time. It’s cool, man. It’s a really cool story.
Konrad: 46:47
Yeah. I mean, I would say it’s like, it sounds very different in terms of the applications and all that, but it’s basically the same principles. It’s just how you figure out how to track something in a way that is painless, and then how you iterate on it, right? And it’s the same thing that I’m doing personally. It’s the same thing that we do for clients. It’s just like, for a business, the metrics are a bit harder to track than they are for an individual. So, that’s why there’s more work into figuring out what they are and how to calculate them and all that, but it’s the same process.
Jon: 47:20
Yeah. You can’t, business-wise, across a business, wake up in the morning and say, “How great did we do yesterday?” And put it on a scale of one, two, three.
Konrad: 47:28
Right, yeah. Yeah. Exactly. It takes a little more digging than that.
Jon: 47:34
Cool. All right. So, tell us one question that we always ask, and I have a feeling you’ll have a really interesting answer, is what are you reading right now? Do you have any book recommendations for us?
Konrad: 47:46
Sure. So, I’m in the process of probably like a year-long project, I think, to read a biography of every US president. And I’m getting pretty far along, so I just am in the finishing-
Jon: 48:01
You did it chronologically, right?
Konrad: 48:03
Yeah, chronologically. So, I’m in the finishing chapters of a biography of Gerald Ford.
Jon: 48:10
Oh, yeah. You’re getting there. 21st century.
Konrad: 48:13
Yeah. And so, I can’t say that I necessarily recommend it that much as a standalone book. But in the context of all the different presidential biographies, it’s interesting. The one I’m reading is called “Ambition, Pragmatism, and Party.” But I think that I would have two notes on this project, and takeaways. One is just that it’s so interesting seeing presidents, and just individuals from different perspectives, and especially early on, starting with George Washington and John Adams, you read these biographies and they overlap so much. And you could sort of think that, “Oh, it’s kind of boring hearing the same story over and over.”
Konrad: 49:04
But you get a different perspective on the same story, which I think is really interesting, because it really points to how each of us is living a different story, how each business challenge that happens has different sides to it, whether you’re the customer, whether you’re the company, whether you’re a competitor. And just thinking about those things really helps think about even my day-to-day, like the business and data and different analyses from different perspectives, to really think about like, “How can you better advance those that you’re doing by taking a different look at it?”
Konrad: 49:44
But in terms of a specific book, the one that impressed me the most, honestly, probably, out of all of them, has been Robert Caro’s five-part series on Lyndon Johnson, which I think is just called like, “The Life of Lyndon Johnson” or something like that. Each volume has a different title. But he goes into so much depth about the rise and the career of Lyndon Johnson, to the extent where he talks about how Lyndon Johnson was powerful in the Senate. But instead of just saying, “Hey, he did a really good job in the Senate. He got this and this passed,” he spends a couple chapters on, “This is how the Senate works. This is why it was dysfunctional before Lyndon Johnson came into it. And this is why it’s so important and impactful that he was able to get this legislation passed that he did, even though he got criticized so much during the time for that.”
Konrad: 50:49
And it really gives you this context. And again, makes you realize how the power of storytelling, and really understanding and putting things in context like tying it back to data, right? Really the power of putting the things that you’re presenting in context to your readers, because you might have some amazing insight that you’ve seen, because you’ve been down in this data for maybe months, even years, and you say, “Look. It’s right there. This is the best thing ever.” Right? But your audience is going to say, “Well, that’s great, but either I don’t know why that matters, I don’t believe it.” And so that really gives this understanding of why it’s so important to tell his whole story of how and why what you’ve discovered matters.
Jon: 51:36
Amazing.
Konrad: 51:38
And I think that series really just puts the rest of the biographies to shame. Now, I’m reading these … I reach the Richard Nixon and now I’m on Gerald Ford, and I’m like, “Yeah, these are great, but they’re very high-level and almost surface level compared to … I think it’s thousands of pages, on one person.”
Jon: 51:58
Right, right, right. You’re like, “Who do you think I am? I mean, I’m the kind of person who reads every biography of every president. You think that this high-level bull crap is going to be good enough for me? What are you thinking?” And it’s really interesting. That name, Robert Caro, is a name that I know, and I don’t know why. Does he do a lot of biographies?
Konrad: 52:17
So, he’s done a couple biographies. So, another one that he’s really well-known for is on Robert Moses.
Jon: 52:27
Oh. Yeah, so … Yeah.
Konrad: 52:30
It’s called “The Power Broker.” And if anything, he might be just as famous, or if not more famous for having written that one. For those of you who don’t know, Robert Moses was the head of … I think eventually the Triborough Authority. But basically in New York City, he was, from the 1920s until the 1960s, one of the largest influencers on city infrastructure. So, he got parks built. He got highways built. And he was very controversial for a couple reasons. One, he was very car-focused, so he knocked down a lot of neighborhoods to build big highways through them, and he was the one who wanted to knock down the West Village and build a highway through it.
Konrad: 53:17
And he also had some classist and racist tendencies. He felt like his parks should be most accessible to people of the middle class, often white. But he did get a lot done, and he did increase overall quality of life. So it’s, again, a very interesting story. Same with Lyndon Johnson. You read about Lyndon Johnson and his path to power is always very … He was a very corrupt politician, very much doing things for favors, using dark money, getting cash from donors, all sorts of tricks.
Konrad: 54:00
But he was the first person to be able to pass any sort of civil rights legislation in the Senate for, I think it was like 50+ years. And the first time it was passed was under him, and the first thing he passed was also not that great. There was all sorts of problems with it, but he got something passed. And so, I think that Robert Caro really is fascinated with this kind of gray area of these people that can be really painted as villains, and in certain situations are, but their story is often more complex than just that.
Jon: 54:35
Nice. That was an amazing set of stories. I am excited to be able to read either of those … Well, I guess I was going to say “books,” but that’s actually four books. Three LJB books, and one Robert Moses book. And tying it into the overarching narrative of data and telling a story, you’re doing my job for me, so thank you very much.
Konrad: 54:57
Yeah, yeah.
Jon: 54:59
Nice. All right, Konrad. So, thank you so much for being on the show. Is there anything … So, I imagine, given your line of work, if we have listeners that are looking for help with getting their data structured, having analytics be more real time, being able to make better predictions from their data, visualize their data, any of that kind of reporting, they should probably be reaching out to you, right? I guess especially if they’re working in consumer-packaged goods.
Konrad: 55:27
Yes. Yeah. In general, our greatest level of expertise is around CPG, growing firms in that. In general, anything that’s being sold online, directly to consumers, often. And then, yeah. I’ll kind of leave it there.
Jon: 55:47
Nice. So, how should people get in touch with you, either for further biography recommendations or for business inquiries?
Konrad: 55:56
Sure. So, LinkedIn is easy, or you can go on our website, which is just impaktadvisors.com. “Impakt” is spelled with a “K.” And either place is the best place to go.
Jon: 56:11
Nice. All right. Thank you so much, Konrad, for being on the show. Super fascinating to have you here, and hopefully we’ll have you again soon.
Konrad: 56:18
Thank you, and likewise. Very much enjoyed it.
Jon: 56:26
Wow. What an impressive individual Konrad is. He blew me away at the end there with his knowledge of the biographical literature, after all of the earlier technical and commercial insights. In today’s episode, we covered leveraging data and analytics for iteratively improving a business across all aspects, from optimizing marketing campaigns through to predictive, real time profit margin calculations, specific data science tools and approaches for making these commercial improvements possible, including Splink, logistic regression, and XGBoost.
Jon: 56:57
Konrad also provided tricks for successfully running a completed distributed company, including incorporating some Scrum techniques, engaging highly experienced contractors, and hiring adaptable, entry level full time employees that meet your core needs.
Jon: 57:13
We also had guidance for your own personal evolution driven by data and analytics, including habit tracking, methodical reflection periods, and physiology monitoring tools like Whoop and the Apple Watch. As always, you can get all the show notes, including the transcript for this episode, the video recording, any materials mentioned on the show, and the URLs for Konrad’s LinkedIn profile, at www.superdatascience.com/465. That’s www.superdatascience.com/465.
Jon: 57:43
If you enjoyed this episode, I’d of course greatly appreciate it if you left a review on your favorite podcasting app or on YouTube, where we have a video version of this episode. And you’re welcome to add me on LinkedIn, but it might be a good idea to mention you were listening to the SuperDataScience podcast, so that I know you’re not a random salesperson. As this is a free podcast, if you happen to be looking for a way to help me out, I’d be very grateful if you left a rating of my book, Deep Learning Illustrated, on Amazon or Goodreads, gave some videos on my YouTube channel a thumbs up, or subscribe to my free, content-rich newsletter on jonkrohn.com.
Jon: 58:19
To support the SuperDataScience company that kindly funds the management, editing and production of this podcast without any annoying third party ads, you could create a free login to their learning platform at www.superdatascience.com, or consider buying a usually pretty darn cheap Udemy course published by SuperDataScience, such as my Machine Learning and Data Science Foundations Masterclass.
Jon: 58:40
All right. Thanks to Ivana, Jaime, Mario and JP on the SuperDataScience team for managing and producing another amazing episode today. Keep on rocking it out there, folks, and I’m looking forward to enjoying another round of the SuperDataScience podcast with you very soon.
